You are on page 1of 53

Journal of the Association for Information Systems

Volume 24 Issue 2 Article 7

2023

Theorizing about the Early-Stage Diffusion of Codependent IT


Innovations
Srikanth Parameswaran
Binghamton University, sparames@binghamton.edu

Rajiv Kishore
University of Nevada, Las Vegas, rajiv.kishore@unlv.edu

Xuanhui Yang
Xiamen University, xxhhyang@gmail.com

Zhenyu Liu
Xiamen University, zhenyliu@xmu.edu.cn

Follow this and additional works at: https://aisel.aisnet.org/jais

Recommended Citation
Parameswaran, Srikanth; Kishore, Rajiv; Yang, Xuanhui; and Liu, Zhenyu (2023) "Theorizing about the
Early-Stage Diffusion of Codependent IT Innovations," Journal of the Association for Information Systems,
24(2), 379-429.
DOI: 10.17705/1jais.00789
Available at: https://aisel.aisnet.org/jais/vol24/iss2/7

This material is brought to you by the AIS Journals at AIS Electronic Library (AISeL). It has been accepted for
inclusion in Journal of the Association for Information Systems by an authorized administrator of AIS Electronic
Library (AISeL). For more information, please contact elibrary@aisnet.org.
Journal of the Association for Information Systems (2023) 24(2), 379-429
doi: 10.17705/1jais.00789

RESEARCH ARTICLE

ISSN 1536-9323

Theorizing about the Early-Stage Diffusion


of Codependent IT Innovations

Srikanth Parameswaran,1 Rajiv Kishore,2 Xuanhui Yang,3 Zhenyu Liu4


1
Binghamton University, State University of New York, USA, sparames@binghamton.edu
2
University of Nevada, Las Vegas, USA, rajiv.kishore@unlv.edu
3
Xiamen University, China, xxhhyang@gmail.com
4
Xiamen University, China, zhenyliu@xmu.edu.cn

Abstract

In this paper, we focus on the early-stage diffusion of codependent IT innovations, which are a type
of innovation in which the overall innovation consists of two complementary parts that are adopted
by two different adopter communities but where both parts need to be jointly adopted by the two
coadopter communities for successful diffusion of the overall innovation. Using innovation
diffusion, organizing vision (OV), and institutional entrepreneurship theories as the key theoretical
lenses, and an in-depth case study reconstructed using 20 years of discourse surrounding Walmart’s
campaign in the early stage of diffusion of the RFID-in-retailing technology, we develop a four-
phase process model for the early-stage diffusion of codependent IT innovations. We make three
specific contributions to the IS discipline, specifically to the literature on IS innovation adoption and
diffusion. First, we add the notion of coadopter relative advantage and posit that the organization in
the coadopter community with a higher coadopter relative advantage that perceives the highest
degree of coadopter relative advantage will emerge as an institutional entrepreneur (IE) and will
influence the early-stage diffusion of the codependent IT innovation. Second, we add the notion of
an internal-external influencer and posit that the IE may be an actor who is internal to the overall
adoption phenomenon, which involves two different coadopter communities, but external to the
coadopter community with a lower coadopter relative advantage that adopts the innovation
component. Third, we divide the early-stage diffusion process into four phases—emergence,
structuralization, evolution, and chasm—and identify the institutional entrepreneurship strategies
used and the OV functions enacted by the IE during each phase. We propose that the IE for a
codependent innovation will: (1) use the rationale development strategy and enact the interpretation
OV function during the emergence phase, (2) use the resource mobilization strategy and enact the
mobilization OV function during the structuralization phase, (3) use the relationship development
strategy and enact the legitimation OV function during the evolution phase, and (4) use all the three
institutional entrepreneurship strategies and enact all the three OV functions during the chasm phase.

Keywords: Codependent IT Innovation, Early-Stage Innovation Diffusion, Coadopter Relative


Advantage, Internal-External Influencer, Institutional Entrepreneurship, Organizing Vision,
Innovation Diffusion Theory, RFID Technology

Suprateek Sarker was the accepting senior editor. This research article was submitted on August 22, 2015 and
underwent five revisions.

379
The Early-Stage Diffusion of Codependent IT Innovations

technology vendors, and consulting firms,


1 Introduction communicate the innovation to potential adopters,
Compared to the extensive research on the diffusion of shape their perceptions, and fuel the early adoption of
single-whole innovations in the IS field, scant attention the innovation (Brancheau & Wetherbe, 1990; Jeyaraj
has been paid to the diffusion of a class of innovations & Sabherwal, 2014; Rogers, 2003). Later on, these
termed multiproduct innovations in the prior literature early adopters become internal influencers within the
(Bucklin & Sengupta, 1993), which we call adopter community and fuel more widespread
codependent innovations. Distinct from single-whole adoption (Jeyaraj & Sabherwal, 2014; Rogers, 2003).
innovations such as personal computers, spreadsheets, While the notions of relative advantage and
ERP systems, and database technologies, codependent external/internal influencers are the primary
innovations are a type of information technology (IT)- explanations for the diffusion of single-whole
based innovation in which the overall innovation innovations in the literature, scholarship on the
consists of two complementary parts that are adopted diffusion of codependent innovations has been rather
by two different adopter communities, where both the scarce.
parts need to be jointly adopted by members, or Codependent innovations challenge both of the key
coadopters, of the two adopter communities for IDT notions discussed above. Consider the notion of
successful diffusion of the overall innovation (Bucklin relative advantage: Regarding codependent
& Sengupta, 1993). The radio frequency identification innovations, the costs and benefits that the two adopter
and detection (RFID) technology used in retailing for communities will gain from the adoption of their
supply chain management, the focus of the current unique part of the codependent innovation are likely to
study, is an example of a codependent innovation. be different and unique to that community. Further,
There are two main parts to this RFID-in-retailing due to these differences, the relative advantage gained
innovation1—the RFID chip (also known as an RFID from the codependent innovation for the two coadopter
tag) and the RFID reader—which are adopted by communities is likely to be asymmetric, with one
different adopter communities. Suppliers of consumer community enjoying a higher relative advantage from
packaged goods attach RFID tags to their products the codependent innovation, especially during the
before shipping them to retailers. Retailers install early stages of the innovation diffusion process
RFID readers in their supply chain, including (Gaukler et al., 2007). Therefore, the overall interest of
warehouses, trucks, retail stores, etc., to track these coadopter communities in adopting their respective
packages. RFID technology promises to benefit the parts of the innovation is also likely to be asymmetric
overall supply chain of retailing only if both retailers in the early stages of the diffusion phenomenon.
and suppliers jointly adopt their respective parts.
In the RFID-in-retailing example, retailers can expect
Innovation diffusion theory (IDT) posits two key immediate benefits from enhanced product visibility,
notions that explain the diffusion of single-whole traceability, availability, and lower ongoing costs due
innovations. The first notion is the innovation’s to the one-time fixed cost of installing RFID readers
relative advantage, as perceived by the adopter and associated systems. On the other hand, in the early
community. Potential adopters perceive a relative stages, the supplier community realizes no immediate
advantage in adopting an innovation if they assess that benefits but incurs higher ongoing costs due to the
the innovation’s benefits outweigh its costs, given their recurring cost of affixing RFID chips onto products
existing practices (Akhlaghpour & Lapointe, 2018; sold to retailers, a cost that is not incurred by retailers.
Rogers, 2003; Sun et al., 2018). The earlier that As such, in the early stages of diffusion, the retailer
adopters perceive a relative advantage, the earlier they community is likely to have more favorable opinions
adopt the innovation. The second notion is the role of about the RFID-in-retailing innovation than the
influencers, both external and internal to an adopter supplier community and is likely to adopt its part of the
community, in influencing the adoption of the innovation sooner than the supplier community. In
innovation at various stages of diffusion (Akhlaghpour contrast, in single-whole innovations, there is a single
& Lapointe, 2018; Bass, 2004; Strang & Soule, 1998). adopter community with a single cost-benefit calculus
In the early stages of the diffusion process, external (Agarwal & Prasad, 1997; Junglas et al., 2019; Moore
influencers, such as mass media, research institutes, & Benbasat, 1991). For example, organizations

1 Theoretically there can be more than two complementary paper, or in the case of other innovations, such as EMV credit
parts to a codependent innovation, with more than two cards (i.e., cards with a computer chip). For the sake of
different adopter communities adopting each of the different clarity and tractability in our exposition, in this paper, we
part of the codependent innovation. However, typically there consider only two complementary parts to a codependent
are two complementary parts to a codependent innovation, as innovation.
in the case of RFID technology, which is the focus in our

380
Journal of the Association for Information Systems

adopting ERP systems calculate relative advantage coadopter community. In the context of the RFID-in-
from this technology over current technology based on retailing innovation, external influencers might find it
similar benefits (e.g., seamless business process difficult to influence the adoption of RFID tags by the
integration) and costs (e.g., software licensing, supplier community due to their higher costs and lower
workforce training) as those incurred by other relative advantage vis-à-vis the supplier community.
members of the adopter community (Su & Yang, On the other hand, members in the coadopter
2010). In essence, while the traditional notion of community adopting the innovation component with a
relative advantage considers heterogeneity in higher relative advantage might have the incentive to
perceptions of relative advantage, such heterogeneity serve as influencers for the coadopter community
is in the perceptions of members within a single adopting the innovation component with the lower
adopter community based on a single set of costs and relative advantage, as the former community will
benefits; however, heterogeneity in the perceptions benefit from the codependent innovation only when
between two different coadopter communities, which the latter community also adopts the Innovation. In the
is based on two different and unique sets of costs and context of the RFID-in-retailing innovation, retailers
benefits, is not considered. might have an incentive to serve as influencers for the
supplier community that is adopting RFID tags
In addition, the successful diffusion of a codependent
because retailers will benefit from this overall
innovation requires the joint adoption of two different
innovation only when suppliers adopt their part of the
parts of the innovation by two different coadopter
innovation. So, are these influencers internal or
communities. This implies that the realization of a
external influencers? Traditional IDT would suggest
coadopter community’s relative advantage from
that retailers as influencers for the supplier community
adopting their part of the codependent innovation is
are internal influencers because they belong to one of
contingent on the other coadopter community adopting
the two adopter communities. However, we argue that
their part of the codependent innovation. Therefore,
retailers are external to the supplier adopter
even if a coadopter community perceives a high
community that is being influenced to adopt the
relative advantage, its members will not be able to
innovation component with a lower relative advantage,
realize this relative advantage by adopting their part of
i.e., the RFID tag.
the codependent innovation if members of the other
coadopter community resist adopting their part of the Therefore, we ask: How does the diffusion process
codependent innovation due to their lower relative unfold during the early stages of codependent
advantage, especially during the early stages of the innovations? Specifically, which actors play the role of
diffusion process. In other words, a high relative influencers during the early-stage diffusion process for
advantage for one community, which is a strong codependent innovations? And what actions do these
predictor of diffusion in the traditional innovation influencers engage in to influence the early-stage
diffusion literature, may be insufficient to kickstart the diffusion process for these types of innovations? These
diffusion of codependent innovations overall. In the questions are important to address because new
RFID-in-retailing example, retailers’ relative codependent IT innovations continue to be developed
advantage in adopting RFID readers will bear fruit and implemented in various industries, as IT continues
only if suppliers also adopt RFID tags and affix them moving toward more open and decentralized systems
on product packaging or palettes. As discussed above, and as interdisciplinary advancements in fields
members of the supplier community might resist the including nanotechnology, computer science, and
adoption of this innovation due to a lower or negative biomaterials continue pushing the horizons of these
relative advantage, especially during the early phases types of IT innovations. Further, many of these
of the diffusion phenomenon, in turn reducing the innovations are targeted for adoption by very large
likelihood of realization of the retailers’ relative adopter communities, such as the recent Europay,
advantage from the RFID codependent innovation Mastercard, and Visa (EMV) chip-embedded credit
(Dutta et al., 2007; Gaukler et al., 2007). For this card payment system, which is expected to be adopted
reason, we also need to extend the traditional notion of by hundreds of millions of consumers and millions of
relative advantage to the notion of a coadopter relative retailers, banks, credit card companies, and other
advantage to explain how the early-stage diffusion organizations. Finally, these innovations will likely
process for codependent innovations unfolds. transform industries, organizations, and the way that
individuals live and work in fundamental ways, as can
Next, we consider the role of influencers. External
already be seen with the RFID and EMV-credit card
influencers who typically play an influential role in the
technologies in the supply chain and retailing arenas.
early stages of the diffusion process in the context of
However, the successful diffusion of such innovations
single-whole innovations might find it hard to
necessitates that all parties (e.g., individuals, retailers,
influence the adoption of the part of the innovation that
and processing networks for credit cards) jointly adopt
yields a lower relative advantage to the relevant

381
The Early-Stage Diffusion of Codependent IT Innovations

these innovations in order for the benefits to flow to Based on the case of the adoption of RFID-in-retailing,
different parties in the value chain. As with the RFID- we develop a process model of the early-stage
in-retailing diffusion discussed above, the diffusion of diffusion process for codependent innovations and
the EMV-chip credit card payment technology has also examine the different roles of the institutional
been slow and complicated. A CNBC report showed entrepreneur (IE) during the different phases of the
that in its early stages, only 37% of US businesses had early-stage diffusion process. There are three specific
adopted EMV credit card payment technology by ways in which we make this contribution. First, we add
March 2016, despite the deadline of October 1, 2015 the notion of coadopter relative advantage and posit
that was set for adoption of this payment technology that the member of the coadopter community that
by all parties concerned and despite significant perceives the highest degree of coadopter relative
penalties for the failure to adopt (Dimyan, 2016). advantage between the two different adopter
communities will emerge as an IE and will influence
This paper is an attempt to fill the current research gap
the early-stage diffusion of the codependent
through an in-depth case study of the early stages of the
innovation. Second, we add a nuanced understanding
diffusion of RFID technology in retailing. We use
that the key influencers in the early stages of the
innovation diffusion theory (IDT) (Akhlaghpour &
diffusion process need not be external to the adopter
Lapointe, 2018; Rogers, 2003; Strang & Soule, 1998),
community, as suggested by the traditional innovation
organizing vision theory (OVT) (Swanson & Ramiller,
diffusion literature. We add the notion of an internal-
1997), and institutional entrepreneurship theory (IET)
external influencer and posit that the IE as an
(Hardy & Maguire, 2017) as the key lenses for this
influencer may be an actor internal to the overall
study. First, IDT deals with the notions of relative
adoption phenomenon involving two different adopter
advantage and influencers (i.e., external vs. internal) in
communities but external in terms of the actor’s
explaining innovation diffusion (Akhlaghpour &
influence on the focal coadopter community. Third, we
Lapointe, 2018; Rogers, 2003; Strang & Soule, 1998).
divide the early-stage diffusion process into four
The notion of the coadopter relative advantage helped
phases—emergence, structuralization, evolution, and
us conceptualize a potential coadopter’s assessment of a
chasm—and identify the institutional entrepreneurship
codependent innovation. The notion of internal-external
strategies used and the OV functions enacted by the IE
influencers allowed us to conceptualize the source of
during each phase. Our contributions also have
influence, which is internal to the codependent
important practical implications for all the entities
innovation’s adopter community but external to the
involved in the institutional field surrounding a
coadopter community that has a lower coadopter
codependent IT innovation.
relative advantage. Second, IET studies why and how
proactive actors who may benefit from new business
practices “create new institutions or transform existing 2 Theoretical Background
ones” (Maguire et al., 2004, p. 657). Thus, this theory is
aligned with our goal of theorizing the role of proactive 2.1 Codependent Innovations and
actors who may benefit from a codependent innovation Innovation Diffusion Theory
to act as influencers and the strategies they use in the
Codependent innovations differ from single-whole
early-stage diffusion of codependent innovations.
Finally, OVT deals with the public discourse innovations according to two aspects. First,
surrounding IT innovations created by actor codependent innovations consist of two
communities with a vested interest in the diffusion of the complementary parts that are adopted by two different
IT innovation. At the institutional-field level, an adopter communities (e.g., RFID in retailing). On the
organizing vision (OV) interprets and legitimizes the IT other hand, single-whole innovations are adopted as a
innovation and mobilizes appropriate resources for the whole by a particular community of individual or
diffusion of that IT innovation, functions that may lead organizational adopters. The adopters stand to benefit
to the development of an institutional field surrounding from the adoption of “single-whole” innovations both
the focal IT innovation (Swanson & Ramiller, 1997; directly from their use of the innovation as well as
Wang & Ramiller, 2009). OVT provides us with a indirectly from any network externalities that may
theoretical lens for analyzing the key OV functions that result from the diffusion of the focal innovation in the
influencers enact through their actions during the early- adopter network (Katz & Shapiro, 1985, 1986; Rogers,
stage diffusion process in the context of codependent 2003). Enterprise resource planning (ERP) is an
innovations in the institutional field. example of a single-whole innovation. While there
may be several parts of ERP, including software,
We contribute to the IS innovation diffusion literature hardware, database, services, etc., all of these parts are
by studying codependent innovations with two adopted by a single adopter community (Wang &
complementary parts, and examine how they diffuse Ramiller, 2009).
during the early stages of the innovation process.

382
Journal of the Association for Information Systems

Second, there is a need for joint adoption of the two adopters, also called innovators, see the relative
different but complementary parts of a codependent advantage from adopting the innovation (Libai et al.,
innovation by two different adopter communities in order 2008; Moore, 2014; Rogers, 2003). Next, the innovation
for the benefits to flow in the value chain. Joint adoption is embraced by early adopters, who are visionaries and
occurs when an entity in one adopter community adopts do not require much evidence of the innovation’s
one part of the codependent innovation, and another entity relative advantage. As the diffusion process continues to
in the other adopter community adopts the other evolve, the innovation is adopted by the “early
complementary part of the codependent innovation. For majority,” followed by the “late majority,” and finally
example, the RFID-in-retailing innovation will be by “laggards” according to the adopters’ understanding
beneficial to the overall supply chain, i.e., both retailers of the relative advantage to be gained from the
and their suppliers, only when there is joint adoption of innovation based on both external and internal
this codependent innovation, with suppliers adopting communication influencing the adoption phenomenon.
RFID tags and the retailers adopting RFID readers (Dutta The percentage of adopters from the overall adopter
et al., 2007). Retailers can better manage their supply community in each of these adopter categories may vary
chains using RFID technology only when their suppliers for different innovations, but the generally accepted
collectively tag all their products, because only then can percentages based on the assumption of a normally
retailers read the RFID tags to identify each product as it distributed diffusion process are 2.5% innovators,
flows through their supply chains in order to accurately 13.5% early adopters, 34% early majority, 34% late
estimate product arrival times, shortages, stock-outs, etc. majority, and 16% laggards. Once a certain percentage
Suppliers can benefit by directly tracking shipments, of the adopter community, somewhere between 10%
which can obviate the need to investigate chargebacks and 20%, has adopted the innovation, the peer influence
from cargo theft. Further, if retailers share their data on process accelerates the S-curve until all other adopter
sales, inventory, and shipments collected from RFID categories—early majority, late majority, and
readers at the point-of-sale, store shelves, and laggards—adopt the innovation. The diffusion is said to
warehouses, etc., suppliers can replenish out-of-stock be in an early stage until a majority of adopters embrace
items faster, based on sales data, and verify the timeliness the innovation. In the S-curve, the early stage ends when
of promotions schedules, based on information about the early majority starts adopting, after which the growth
when products reach the store shelf (Wailgum, 2006; stage begins.
Whitaker et al., 2007). These supply chain benefits will
Scholars have argued that there is a chasm between the
not be realized if either the retailers or their suppliers do
not adopt their complementary parts. early and growth stages of the diffusion (Clarysse et al.,
2014; Moore, 2014). The validation provided by early-
These two aspects of codependent innovations are stage adopters does not serve as a good reference for the
intriguing and have important implications for adopters in the early majority (i.e., growth-stage) due to
advancing the two key notions in the IDT literature: (1) the psychographic profile differences between the
relative advantage and (2) the role of influencers. The adopters in these groups (Moore, 2014; Xu et al., 2017).
first notion of relative advantage is well established as Until the innovation fits the pragmatist orientation of a
one of the most important innovation attributes in the typical early majority adopter, diffusion reaches a
broader innovation diffusion literature (Agarwal & standstill, creating what scholars call a “chasm” between
Prasad, 1997; Rogers, 2003). This construct has also
the early and the growth stages of diffusion (Moore, 2014;
emerged as one of the most dominant predictors of the
Xu et al., 2017). Indeed, the innovations that fail to appeal
diffusion of IT innovations (Akhlaghpour & Lapointe,
to the early majority adopters do not cross the chasm
2018). Relative advantage “is a ratio of the expected
benefits and the costs of adoption of an innovation” (Davidson et al., 2015; Moore, 2014; Purdy & Gray,
(Rogers, 2003, 233). However, relative advantage is not 2009). Only those innovations with a confirmed relative
an objective, invariant innovation attribute, and different advantage and applicability to the early majority adopter
adopters in a community of adopters perceive an category will cross this chasm to achieve mainstream
innovation’s relative advantage at different levels given adoption in the growth stage (Moore, 2014).
their current situation and future expectations (Moore & In the context of codependent innovations, there are two
Benbasat, 1991). Because different adopters have different coadopter relative advantages for the two
different perceptions of the relative advantage of the different adopter communities adopting two different but
innovation, they adopt it at different times. This
complementary parts, which confer different sets of
phenomenon is the essence and basis of the innovation
benefits and costs to their respective coadopter
diffusion process and the associated S-shaped diffusion
communities. For example, in the case of the diffusion of
curve (see Figure 1), which is essentially the cumulative
normal distribution curve. The S-shaped curve shows the RFID-in-retailing innovation, suppliers and retailers
the cumulative number of adopters as a function of time are the two different adopter communities (i.e., the
(Rogers, 2003). During the early stages, the diffusion coadopter communities) adopting RFID tags and RFID
curve rises slowly because only a few venturesome readers, respectively.

383
The Early-Stage Diffusion of Codependent IT Innovations

Note: yellow line—S-shaped diffusion curve showing the cumulative number of adopters; blue line—bell curve showing the number of adopters;
double-headed brown arrow—distinguishes early and growth stages of diffusion

Figure 1. Diffusion of Innovations Model for Single-Whole Innovations

Hence, we add nuance to the notion of relative case of the RFID-in-retailing innovation, the retailer
advantage and introduce the notion of coadopter community is overall more likely to have a higher
relative advantage in this paper. We define coadopter coadopter relative advantage as compared to the
relative advantage as the extent to which one part of a supplier community, especially during the early stages.
codependent innovation (e.g., RFID reader) offers a Suppliers bear RFID tagging costs from the start, which
relative advantage (over prior technology) to its are immediate and recurring, whereas the benefits from
coadopter community (e.g., the retailer community) RFID-in-retailing are long-term and not immediate. In
contingent upon the other coadopter community (e.g., contrast, retailers incur an initial fixed cost of installing
the supplier community) adopting the other part of the RFID readers but they realize supply-chain benefits
codependent innovation (e.g., RFID tags). We explore right from the earliest stages. Therefore, the retailer
whether and how a member of the coadopter community community perceives a higher coadopter relative
that perceives a higher degree of coadopter relative advantage than the supplier community in the early
advantage influences the early-stage diffusion of a stages (Gaukler et al., 2007). However, this does not
codependent innovation. mean that there is no heterogeneity in how different
members of the supplier community view their
We clarify and emphasize that there will be
coadopter relative advantage. Indeed, some suppliers
heterogeneity in perceptions of coadopter relative
might perceive a higher coadopter relative advantage as
advantage among members of both coadopter
they foresee a long-term reduction in their supply-chain
communities, including the community with a lower
costs by investing in RFID tags in the early stages,
coadopter relative advantage, such that some members
despite knowing that their adoption of RFID tags will
of each community are likely to perceive their respective
also benefit retailers and likely to a higher extent. The
coadopter relative advantage to be fairly high while
same is true of the retailer community. Essentially, it is
others perceive it to be quite low. In fact, the diffusion
these innovators and visionaries in the two coadopter
of a codependent innovation will occur only if some
communities, i.e., retailers and suppliers, who initiate
members in each of the two coadopter communities
and spur the early-stage diffusion of the RFID-in-
perceive their respective coadopter relative advantage to
retailing codependent innovation.
be high and thus adopt their respective complementary
parts of the codependent innovation. For example, in the

384
Journal of the Association for Information Systems

The second notion is the role of influencers, both of actors formed by the same regulatory processes or
external and internal to an adopter community, in sharing the same meaning systems (Nielsen,
influencing the adoption of the innovation at various Mathiassen, & Newell, 2014). These actors include
stages of diffusion by serving as communication adopters, vendors, consultants, governments, investors,
channels (Bass, 2004; Strang & Soule, 1998). The journalists, research institutions, and professional
classical innovation diffusion process unfolds for single- associations. Using an institutional perspective, we
whole innovations fueled by some innovators and early borrow from two streams of research. In the first stream,
adopters who adopt based on information obtained from we focus on the institutional entrepreneurship literature.
external influencers, e.g., a technology vendor, a In the second stream, we focus on the OV, which is the
consultant, etc. External influencers, such as mass media IT discourse created and employed by an
channels (e.g., television, radio, brochures, and interorganizational community of actors with varying
magazines), rapidly disseminate generic information material interests (Swanson & Ramiller, 1997).
about the innovation to a large number of potential
adopters (Jeyaraj & Sabherwal, 2014). However, the 2.2 Codependent Innovations and
early majority does not adopt the innovation based Institutional Entrepreneurship Theory
solely on information from mass media. Early adopters
become internal influencers within the adopter We borrow from IET and conceptualize influencers as
community by sharing more specific, experiential institutional entrepreneurs (IEs). IET explores why and
information about the innovation with adopters in the how proactive organizations with strategic interests try
early majority category with whom they share an to initiate changes that bring about new institutions or
interpersonal relationship (Jeyaraj & Sabherwal, 2014). transform existing ones (Leca et al., 2008). This theory
In the case of codependent innovations, the coadopter views IEs as actors trying to bring about an institutional
community that sees the highest degree of coadopter change while being institutionally embedded. If actors
relative advantage may act as an influencer for the other are institutionally embedded, their norms and collective
coadopter community with a lower coadopter relative beliefs will be institutionally determined (Battilana et
advantage. In the case of RFID-in-retailing, a retailer al., 2009). How, then, do IEs initiate institutional change
may influence the supplier adopter community to foster in spite of the institutional field’s pressure toward stasis
joint adoption of this codependent innovation at a rapid (Battilana et al., 2009)? This central problem in IET is
clip. This finding is in consonance with the extant called the “paradox of embedded agency” (Seo & Creed
literature on diffusion. However, a nuance we add here 2002). IEs can bring about institutional change by
is that of an internal-external influencer. In the case of certain conditions that trigger them to sense an
RFID-in-retailing, a retailer is an internal influencer opportunity for institutional change. These conditions
when we consider the overall innovation diffusion include institutional field conditions like the extent of
phenomenon, which involves two different coadopter institutionalization of existing practices, regulatory
communities coadopting two different but forces, and states of crisis (Hardy & Maguire, 2017).
complementary parts of the codependent innovation. These conditions can increase uncertainty or create
However, a retailer is an external influencer in terms of problems for an actor and consequently persuade actors
the influence on the coadopter community of suppliers, to seek ways to reduce problems (Hardy & Maguire,
who are not likely to be highly motivated to adopt RFID 2017). Moreover, an actor’s relative power among the
chips due to a lower coadopter relative advantage for actors in the institutional field and/or the availability of
them as compared to that for the retailer community. financial, political, or social resources at their disposal
Therefore, we add a nuance to the notion of the could grant them the power to influence the field.
influencer, and define an internal-external influencer as To study the role of influencers in the diffusion of
a member of one coadopter community (e.g., the retailer codependent innovations, we focus on the influencing
community) who adopts one part of a codependent strategies used by an IE to affect the diffusion of an IT
innovation (e.g., RFID readers) but influences members innovation within an adopter community in the
of the other coadopter community (e.g., the supplier institutional field. Based on their synthesis of the IET
community) who are adopting the other part of the literature, Hardy and Maguire (2017) proposed that IEs
codependent innovation (e.g., the RFID tags). utilize three broad strategies to promote change: (1)
When organizations explore innovative IT solutions, rationale development: IEs construct and communicate
scholars recommend an institutional perspective to learn rationales to persuade other actors to support the
how institutions develop, produce, and diffuse new IT institutional change; (2) resource mobilization: IEs
solutions (Hinings et al., 2018; Nielsen et al., 2014; mobilize resources to reward or punish other actors to
Wang & Ramiller, 2009). Organizations seldom exist as persuade them to support the institutional change; and
silos. They are embedded in an institutional field, its (3) relationship development: IEs establish relationships
interorganizational relationships, and the material and to garner support for the proposed change (Hardy &
linguistic objects that both constrain and enable their Maguire, 2017). Rationale development activities may
efforts to innovate with IT. An institutional field is a set include but are not limited to telling stories about their

385
The Early-Stage Diffusion of Codependent IT Innovations

(un)successful implementation of an innovation, using and actors using discourse to outline how to diffuse the
frames to get the community to focus on the innovation, innovation (Amadoru et al., 2021; Swanson & Ramiller,
and producing, consuming, and distributing knowledge 1997; Wang & Swanson, 2007). Legitimation in the
pertaining to the innovation (Battilana et al., 2009; institutional field will be evident in actors using
Phillips et al., 2004). Resource mobilization activities authority and power to attach legitimacy to an
may include but are not limited to actors funding innovation, using success stories to convey why other
research activities for developing an innovation, actors need to embrace the innovation, setting standards
implementing negative consequences like penalties on for the innovation, and using discourse to show how the
dependent actors, materially rewarding supporters, and innovation fit into norms (Kaganer et al., 2010; Marsan
using formal authority in legal and professional spheres et al., 2020; Paré et al., 2020; Swanson & Ramiller,
to authorize new practices or create new rules (Hardy & 1997; Wang & Swanson, 2007).
Maguire, 2017). Relationship development activities
may include but are not limited to conducting research 3 Theoretical Integration
studies for developing an innovation in collaboration
with another actor, joining a standards body as a
3.1 Innovation Diffusion Theory and
member, and forming alliances and partnerships (Hardy
& Maguire, 2017). Organizing Vision Theory
OVT and IDT share similarities and differences. Both
2.3 Codependent Innovations and theories discuss the role of adopters, vendors,
Organizing Vision Theory consultants, research institutes, standards organizations,
and mass media. Actors in IDT, including adopters, act
We borrow from OVT to conceptualize how
as communication channels interacting with potential
institutional change happens around a codependent IT
adopters to transmit the innovation (Akhlaghpour &
innovation. OVT concepts of the public discourse
Lapointe, 2018). Potential adopters are also placed in the
around an IT innovation, and OV functions that foster
social system category to conceptualize the adopter
diffusion provide a framework to analyze the diffusion
community (Hu et al., 1997). OVT focuses on an
of codependent innovations. Analyzing the public
interorganizational community consisting of actors with
discourse surrounding a codependent innovation would
material interests in an innovation. However, the
help us understand how different adopter communities
interactions represent this community’s public
communicated and perceived the coadopter relative
contemplation of new technology to make sense of the
advantage. OV is a collective (to some degree),
innovation at the community level, in contrast to a
convergent, institutional field-level idea of an IT
relatively private sensemaking effort discussed in IDT
innovation that is based on actors’ knowledge about and
(Swanson & Ramiller, 1997).
experience using the IT innovation (Wang & Ramiller,
2009). Actors in the innovation community such as The relative advantage concept is also applicable in both
adopters, vendors, research analysts, consultants, and theories. According to IDT, potential adopters assess the
journalists develop a set of ideas about the application relative advantage of an innovation based on
of an IT innovation in organizations. information provided by external and internal
influencers. In OVT, potential adopters make an
An OV diffuses an IT innovation through three
adoption decision based on the business problematic
functions: (1) interpretation by conveying the uses and
linked to the OV and discussed by way of public
usefulness of the innovation to the community
discourse around the innovation, which in turn helps
(Davidson et al., 2015), (2) legitimation by backing the
them assess the costs and benefits (i.e., the relative
innovation and demonstrating its applicability to
advantage) associated with the innovation (Marsan et
accepted business problems (Davidson et al., 2015;
al., 2012; Miranda et al., 2015; Ramiller & Swanson,
Kaganer et al., 2010), and (3) mobilization by
2003; Swanson & Ramiller, 1997).
motivating actors to participate in the diffusion process
(Amadoru et al., 2021; Davidson et al., 2015). The mechanisms in IDT are the influencers’ actions and
Interpretation in the institutional field will be evident the relative advantage perceived based on adopter
through research and development activities to identify profiles (Akhlaghpour & Lapointe, 2018). OVT
the usefulness of the innovation, activities funding mechanisms are mobilization, interpretation, and
research institutes to better understand an innovation, legitimation. In contrast to IDT mechanisms, OVT
and using public discourse to generate buzzwords counterparts have broader outcomes because they foster
(Kaganer et al., 2010; Liao, 2016; Swanson & Ramiller, diffusion by shaping the institutional field around an IT
1997). Mobilization in the institutional field will be innovation. Thus, while both theories focus on the
evident through vendors displaying new products in adoption and diffusion of technology as outcomes, OVT
forums, powerful actors coercing other actors to adopt also focuses on commercial opportunities, community
the innovation, actors deploying or calling other actors structure, and technology invention and adaptation as
to deploy resources for implementing the innovation, outcome variables (Nielsen et al., 2014; Swanson &

386
Journal of the Association for Information Systems

Ramiller, 1997). Moreover, the mechanisms posited by campaign for an innovation and to author special
the two theories are related. For instance, an OV’s sections in business magazines discussing the future of
mobilization function forges a network of adopters to an industry (Wang & Swanson, 2008). Such actions
bring about strong ties needed for IDT’s internal create a market for the innovation at an institutional
influence mechanism (Strang & Soule, 1998). level. For potential adopters, these actions convey the
relative advantage to be gained from adopting an
Both theories discuss the influential roles different
innovation, thereby acting as the IDT’s external
actors play in different phases based on their perceived
influence mechanism.
benefits from the diffusion of an innovation. While OVT
discusses which actors dominate the discourse at each Regarding outcome variables, both theories also focus
phase (Wang & Ramiller, 2009), IDT posits that on the adoption and diffusion of a new technology. In
external influencers initially influence potential addition, IET also focuses on institutional field change.
adopters, which is followed by influence based on The two theories taken together can better illuminate
firsthand experience by internal influencers (Jeyaraj & the diffusion process by conceptualizing the role of
Sabherwal, 2014). Taken together, these two theories actors in (1) directly influencing potential adopters (as
can better illuminate the diffusion process by (1) in IDT) and (2) developing an institutional field when
identifying the sources that shape the relative current institutional arrangements do not support an
advantage—influencers and the business problematic innovation (as in IET).
linked to the OV, (2) highlighting the role of
influencers—directly influencing potential adopters and 3.3 Institutional Entrepreneurship Theory
shaping the OV, (3) conceptualizing the mechanisms and Organizing Vision Theory
fostering diffusion, and (4) explaining multiple
outcomes at the adopter and the institutional levels. OVT and IET also share some similarities and
differences. First, both theories highlight the role of the
3.2 Innovation Diffusion Theory and actor community in influencing institutional change
even though they do so differently. On the one hand,
Institutional Entrepreneurship Theory
OVT designates roles to each actor in the community
IDT and IET share both similarities and differences as that collectively creates the OV, namely vendors,
well. Both theories focus on the interorganizational adopters, research institutes, consultants, etc.
community surrounding an innovation, with IET (Davidson et al., 2015; Swanson & Ramiller, 1997).
focusing more on an IE that engages proactively with On the other hand, IET discusses one proactive actor
the adopter community for an innovation’s widespread embedded in the adopter community of the
adoption. Both theories discuss different actors playing institutional field who occupies a position of power
an influential role at different phases based on their within that community and serves as an IE for
perceived benefits from an innovation’s diffusion. institutional change (Hardy & Maguire, 2017; Marsan
et al., 2020).
Both theories also focus on the role of influencers and
their strategies as mechanisms that drive the Both theories also focus on institutional change
widespread adoption of an innovation. The influencers mechanisms. The OV in OVT interprets, legitimizes,
in IDT act as communication channels transmitting the and mobilizes to diffuse the IT innovation and make it
relative advantage to potential adopters. Despite the the new paradigm (Amadoru et al., 2021; Kaganer et
influencer’s efforts, some innovations do not diffuse al., 2010). The IE in IET interprets the field, develops
due to the existing unsupportive institutional rationale, mobilizes resources, and develops relations
arrangements, which reduce the benefits and increase to effect institutional change (Hardy & Maguire,
the cost of adoption (i.e., lower the relative advantage). 2017). While in OVT, diffusion mechanisms are
To address this, IET discusses the role of influencers, performed by the institutional field-level construct OV
also called as IEs, who bring out the new institutional on the actors in the field, in IET, individual actors
arrangements needed to foster innovation diffusion (organizations or individuals) perform their actions to
(Wang & Swanson, 2007). The IE’s actions go far influence other actors in the institutional field toward
beyond just communicating the relative advantage institutional change.
with potential adopters. They build relationships,
The mechanisms in both theories are related. When an
mobilize resources, and develop rationale (Hardy &
IE’s actions build rationale and develop relationships to
Maguire, 2017) to develop, test, fund, regulate,
understand an innovation, it represents the OV function
legitimize, and use the technology (Wang & Swanson,
of interpretation (Marsan et al., 2020). An IE using their
2007), essentially increasing the relative advantage for
resources to back an innovation and building rationale
potential adopters.
by publishing case studies represents the OV function of
The mechanisms for innovation diffusion are also legitimation (Kaganer et al., 2010). When an IE builds
related in the two theories. For instance, IET suggests relations to structure the community and mobilizes
that we can expect vendors to fund media houses to resources to activate the market, it represents the OV

387
The Early-Stage Diffusion of Codependent IT Innovations

function of mobilization (Wang & Swanson, 2007). In a According to IDT, actors who see an advantage in a
similar vein, OVT argues that, like an IE, certain new technology serve as communication channels and
opportunistic actors try to command the OV to suit their transmit the innovation directly to the potential
interests (Swanson & Ramiller, 1997). adopters to enhance the adopter’s relative advantage
and foster adoption (Jeyaraj & Sabherwal, 2014). The
Both theories also focus on the dynamics of the
diffusion is initially confined to venturesome and
diffusion process. While OVT identifies the actors who
visionary adopters based on the innovation’s potential
are likely to dominate the OV at each phase of the IT
communicated by external influencers (Rogers, 2003).
innovation diffusion (Amadoru et al., 2021; Liao, 2016;
Subsequently, the innovation diffuses in the
Wang & Ramiller, 2009), IET discusses the actions
mainstream market once benefits are confirmed by
performed by an IE at each stage of an institutional
members of the adopter community, i.e., the internal
change (Morrill, 2006; Purdy & Gray, 2009).
influencers (Rogers, 2003). Thus, the new technology
Scholars have also integrated OVT and IET by arguing becomes the dominant technology in the industry,
that IEs can enact an OV function when they perform replacing the incumbent one.
actions that interpret, legitimize, and mobilize an IT
IET posits that actors with vested interests in the
innovation (Kaganer et al., 2010; Marsan et al., 2020;
innovation strive to bring about institutional-level
Paré et al., 2020; Wang & Swanson, 2007). Through
arrangements through rationale development, resource
these actions, ambitious players in an
mobilization, and relationship development (Hardy &
interorganizational community can carve out a
Maguire, 2017). Such actions promote diffusion
preferred institutional environment where they can
through legitimacy, authorized practices, regulations,
benefit from innovation diffusion and
rules, interactor relationships, resource pools,
institutionalization. In line with this thought, Wang
community attention, norms, mandates, and rewards
and Swanson (2007) studied the diffusion of
(Battilana et al., 2009; Hardy & Maguire, 2017; Purdy
professional services automation and proposed that IEs
& Gray, 2009). Consequently, a new institutional field
promote IT diffusion by (1) perceiving the opportunity
emerges surrounding an innovation whose adoption is
for institutional transformation based on the likelihood
taken for granted and advantageous.
of IT diffusion, (2) using discourse to incorporate
compelling success stories from adopters and vendors According to OVT, opportunistic actors shape the
to legitimize the innovation, and (3) mobilizing the discourse surrounding an innovation (Munir &
community by urging the members to focus their Phillips, 2005) and enact the OV functions of
attention on the innovation. interpretation (Amadoru et al., 2021; Kaganer et al.,
2010), legitimation (Kaganer et al., 2010; Marsan et
3.4 Moving from One Institution to the al., 2020; Paré et al., 2020; Wang & Swanson, 2007),
Next and mobilization (Wang & Swanson, 2007) to promote
the diffusion of the innovation. Such actions promote
We now integrate the three streams of literature from the material realization of the innovation through
IDT, OVT, and IET and discuss how they collectively market offerings and support infrastructure, interactor
inform how an institutional field around an incumbent relationships, standards, adoption manuals, technology
technology moves to a new state to foster the diffusion development, and research. These actions motivate the
of a different new technology. The incumbent community by promoting a common understanding of
technology is the well-accepted and widely adopted the technology’s uses, usefulness, legitimacy, and
means for solving a business problem. However, this future (Davidson et al., 2015). Hence, potential
technology is never perfect. New technologies are adopters obtain institutional support for their adoption
launched to address the incumbent technology’s and draw upon the common understanding to shape
problems and, in many cases, provide additional their perception of the innovation’s business use cases
applications for the user (Fayard et al., 2016). Some (Marsan et al., 2012; Miranda et al., 2015; Ramiller &
actors may see a tremendous advantage in launching a Swanson, 2003).
new technology. But launching it requires massive and Table 1 summarizes the theoretical foundations we
enduring new institutional arrangements necessitating discuss above. This table presents the key theoretical
an institutional field change, which faces major concepts and their definitions from the three theories—
resistance from other actors in an institutional field IDT, IET, and OVT. These concepts form the building
(Marsan et al., 2012). The three theories discussed blocks of our processual model of early-stage diffusion
above inform this change process by which an of codependent innovations that we develop in this
institutional field moves from a prior state to a new paper. This table also presents the coding examples for
state, as well as the role the key influencers play during the concepts of interest in our conceptual model, which
this process. Below, we discuss how each perspective is discussed later in Section 5. We describe our coding
informs the institutional change process. process in Section 4.4.

388
Journal of the Association for Information Systems

Table 1. Key Theoretical Concepts


Concept Definition Coding examples Sources

Key notions in the early-stage diffusion of codependent innovations—Rooted in innovation diffusion theory

Extent of joint Adoption and diffusion have been “Wal-Mart has installed RFID systems in 104 Akhlaghpour &
adoption/ joint established concepts in the innovation Wal-Mart stores, 35 Sam’s Clubs and three Lapointe (2018);
diffusion literature since the 1970s. Adoption occurs distribution centers. In total 14,000 pieces of Lai et al. (2006);
when an entity in a particular community hardware has been installed connected with 230 Rogers (2003)
adopts (e.g., accepts, buys, starts using, miles of cable” (IDTechEx, 2005b).
etc.) an innovation, which is a new idea, “In a related development, Gillette announced
technology, product, etc. Diffusion is the Monday that it agreed to purchase 500 million
process by which information about an so-called radio frequency identification (RFID)
innovation is communicated through tags, which contain the special microchips that
certain channels over time among the can communicate wirelessly with computers
members of a community and results in when in the presence of a scanning device—all
the adoption of the innovation by an without any human intervention” (Gilbert,
increasingly larger number of entities 2003)
within the adopter community. The extent
of diffusion of an innovation over time has
been found to typically follow an S-curve.
We extend these concepts in the context of
codependent innovations. Joint adoption,
therefore, occurs when an entity in one
adopter community (e.g., the supplier
community) adopts one part of a
codependent innovation (e.g., RFID tags),
and another entity in another adopter
community (e.g., the retailer community)
adopts the other part of the codependent
innovation (e.g., RFID readers). The
extent of joint diffusion of a codependent
innovation captures the extent to which
the different parts of the codependent
innovation have been adopted by the
different adopter communities.

Perceived This is an enhancement of the established “The retailers involved have the same interests. Akhlaghpour &
coadopter concept of relative advantage, which has ‘In-stock is a key focus for us, but so is loss Lapointe (2018);
relative been studied in the innovation literature prevention,’ says Wal-Mart spokesman Bill Lai et al. (2006);
advantage since the 1970s. Relative advantage is “a Wertz. ‘This technology offers us some Rogers (2003)
ratio of the expected benefits and the costs potential to improve in both
of adoption of an innovation” (Rogers, areas’” (RFIDUpdate, 2003).
2003, p. 233). “Gillette says that retailers and consumer-goods
As an extension, coadopter relative firms in America lose around $30 billion a year
advantage is the extent to which one part in sales because shop shelves run out of
of a codependent innovation (e.g., RFID products and stand empty….The benefits to
readers) offers a relative advantage to its manufacturers include far fewer wasteful
coadopter community (e.g., the retailer inventory glitches and, for retailers, lower
community) contingent on the other shoplifting losses. Procter & Gamble's goal is to
coadopter community (e.g., the supplier use the intelligence provided by the tags to cut
community) adopting the other part of the its inventory by 40%, or $1.5 billion. P&G’s
codependent innovation (e.g., RFID tags). preliminary analysis is that they could lop 4
cents off the dollar in every transaction”
(Schoenberger, 2002).

External This has been an established concept in “The Auto-ID Center continues to add to the Bass (2004);
influencers the innovation literature since the 1970s. impressive lineup of vendors backing its Jeyaraj &
External influencers are actors who do not standards efforts. Intermec Technologies of Sabherwal
adopt the innovation but disseminate Everett, Wash., and STMicroelectronics N.V. (2014); Rogers
generic information about the innovation (NYSE: STM) of Geneva, Switzerland, join the (2003); Strang &
to a large number of potential adopters. likes of Intel, Philips Semiconductor and Tagsys Soule (1998)
For example, research institutes, as sponsors” (RFIDJournal, 2002).
technology vendors, consultants

389
The Early-Stage Diffusion of Codependent IT Innovations

Internal This has been an established concept in “In addition to the investment, CPG Bass
influencers the innovation literature since the 1970s. manufacturers are also worried that RFID will (2004);Jeyaraj &
Internal influencers are actors within the bring little benefit and perhaps disruption to Sabherwal
adopter community who share specific, their current operations” (Violino, 2003a). (2014);Rogers
experiential information about the (2003);Strang
innovation to potential adopters. and Soule (1998)

Internal- We extend the ideas of external and “Why is RFID Hot Now?” Bass
external internal influencers (above) in the context (2004);Jeyaraj &
“In a word, Wal-Mart. Like so many other
influencers of codependent innovations and develop Sabherwal
aspects of the business world, the enormous
the notion of an “internal-external” (2014);Rogers
impact of Wal-Mart’s buying power will dictate
influencer. We define an internal-external (2003);Strang &
how and when consumer goods companies
influencer as a member of one coadopter Soule (1998)
adopt RFID technology. Wal-Mart had a similar
community (e.g., the retailer community)
impact when they began using barcodes in the
who adopts one part of a codependent
1980s. But Wal-Mart could not have made their
innovation (e.g., RFID readers), but who
decree if many other factors were not already in
influences the members of the other
place …The impact of Wal-Mart’s directive on
coadopter community (e.g., the supplier
RFID will speed the adoption of RFID
community) who adopt the other part of
technology not only for their suppliers, but also
the codependent innovation (e.g., the
for the supply chain industry as a whole. To
RFID tags). We call this influencer as an
avoid being left behind in an unfavorable
internal-external influencer because they
competitive position, all supply chain
are internal to the overall adoption
participants must begin planning how RFID
phenomenon, but are external in terms of
should best be deployed for their companies”
their influence for the focal coadopter
(RedPrairie, 2003).
community.

Influencing strategies used by an institutional entrepreneur to effect institutional change (i.e., to effect the diffusion of an IT
innovation within an adopter community in the institutional field)—Rooted in institutional entrepreneurship theory
Institutional entrepreneurs (IEs): Actors who have an interest in particular institutional arrangements and who leverage resources to
create new institutions or to transform existing ones (DiMaggio, 1988; Hardy & Maguire, 2017; Maguire et al., 2004).

Rationale IEs use discourse to construct and “‘I think a lot of the [tag] studies that have been Hardy & Maguire
development communicate rationales to other actors done prior [to the introduction of near-field (2017)
about why they should support, or, at a tags] are invalid, and they need to be done
minimum, not resist the again,’ says Langford. ‘The whole [playing]
institutionalization project in question. field has been changed now’” (Simon Langford,
Wal-Mart's manager of RFID
strategies) (O’Connor, 2006c).

Resource IEs mobilize resources for endorsing and “If anyone still has any doubts that Wal-Mart is Leca et al.
mobilization supporting the implementation of serious about deploying RFID technology in its (2008); Hardy &
institutional change. supply chain, they should be dispelled by its Maguire (2017);
latest revelation. The world's largest retailer says Misangyi et al.
it will require all suppliers to put RFID tags (2008)
carrying Electronic Product Codes on pallets
and cases by the end of 2006. …‘We have
asked our 100 top suppliers to have product on
pallets employing RFID chips and in cases
with RFID chips,’ says Wal-Mart spokesman
Tom Williams. ‘By 2006, we will roll it out
with all suppliers.’” (Roberti, 2003a).

Relationship IEs establish new interactor relations and “Two major retailers and a consumer products Garud et al.
development perform collective actions to implement giant are teaming up to test a new ‘smart shelf’ (2002); Hardy
institutional change. technology that allows for real-time tracking of and Maguire
inventory levels….Gillette, Wal-Mart and the (2017)
U.K.-based supermarket chain Tesco plan to
install specially designed shelves that can read
radio frequency waves emitted by microchips
embedded in millions of shavers and related
products” (Gilbert, 2003).

390
Journal of the Association for Information Systems

Functions of organizing vision enacted by actors within an institutional field to effect institutional change (i.e., to effect the
diffusion of an IT innovation within an adopter community in the institutional field)—Rooted in organizing vision theory
Organizing vision: A focal community idea for the application of information technology in organizations (Davidson et al., 2015;
Swanson & Ramiller, 1997)

Enacting IEs’ actions that convey the innovation’s “‘This technology won’t just improve the Davidson et al.
interpretation essential features and its institutional supply chain—it will revolutionize it in ways (2015); Swanson
coherence to the innovation community. we are only just beginning to understand,’ and Ramiller
Ashton said. ‘Computers are basically blind (1997); Kaganer
today. The technology we are developing will et al. (2010);
enable them to see, for the first time ever’” Amadoru et al.
(Mayfield, 2002). (2021)

Enacting IEs’ actions that activate the entire “The group’s next important step will be to Davidson et al.
mobilization community of actors surrounding the contract an independent, third-party consultant (2015); Swanson
innovation and get them involved in and issue an open call for patent submissions for and Ramiller
realizing the innovation. all parties believing they hold patents essential (1997); Amadoru
to the deployment of UHF RFID technology. et al. (2021)
The consultant will review each patent to
determine whether it is essential to standards
for UHF air-interface RFID protocols”
(O’Connor, 2006b).

Enacting IEs’ actions that convey to the community “This study provides conclusive evidence that Davidson et al.
legitimation the rationale, backed by evidence, for the EPCs increase how often we put products in the (2015); Swanson
innovation adoption. hands of customers who want to buy them, and Ramiller
making it a win for shoppers, suppliers and (1997); Kaganer
retailers” (IDTechEx, 2005b). et al. (2010);
Marsan et al.
(2020)

interests exemplify the role of institutional


4 Research Design entrepreneurship in the early-stage diffusion process of
RFID in retailing. Fourth, RFID diffusion was not just
4.1 Research Site Selection about upgrading technology; it was also about changing
the institutional environment (Riggins & Slaughter,
We chose Walmart’s promotion of RFID in retailing as
2006) as well as transforming existing supply chains
our research site because it is well-suited to this study
(Soon & Gutiérrez, 2008) and business models of
for several reasons. First, RFID technology is an
retailing (DiMaggio, 1988, p. 15), because it would
example of a codependent innovation. There are two
become an essential IT infrastructure in retailing after it
main parts to this innovation, the RFID tag and the RFID
was adopted. These four reasons make the diffusion of
reader. These two parts were adopted by two different
RFID-in-retailing directly pertinent and quite valuable
adopter communities, suppliers and retailers,
to enhancing the understanding of the early-stage
respectively. Second, because of the uncertainties in
diffusion of codependent innovations.
RFID adoption, including the retailer mandate,
technology issues, and cost-benefit absorption, the
criteria used to justify RFID investment in retailing were
4.2 Research Method
based on cost-benefit calculations and a mental image of We used the case study research method, which is an
the diffusion trajectory (Rogers, 2003). Therefore, both empirical inquiry that analyzes the phenomenon under
the individual calculus as well as collective beliefs about question in a real-life context (Myers, 1997). The case
diffusion helped organizations consider their options study method is best suited when studying processes
when choosing whether or not to adopt RFID. This (i.e., “why” and “how” questions), when the focus is on
exemplifies the roles of the coadopter relative advantage contemporary events and when the boundaries between
and OV in the diffusion of RFID. Third, powerful the context and the phenomenon under study are not
initiators and vested interests characterized the process apparent (Benbasat et al., 1987; Yin, 2008). Since we
of early-stage RFID diffusion in retailing. The mandate were studying the contemporary events and the process
by Walmart and other retailing giants were the most through which IEs influence the early-stage diffusion of
remarkable events that shaped the OV (Swanson & a codependent innovation in an institutional field, we
Ramiller, 1997) and the diffusion of RFID-in-retailing. found that the case study methodology best suited our
The strategies of these influential actors with vested purpose. In our case study, we employed the interpretive

391
The Early-Stage Diffusion of Codependent IT Innovations

philosophical perspective (Walsham, 1993). This monitored several sources to understand industry
philosophical perspective posits that reality can be events. Table 2 summarizes our data sources and the
understood only through social constructions such as purpose of each particular source of data. We collected
language, consciousness, tools, documents, shared reports on RFID in retailing published in several online
meanings, and other artifacts (Klein & Myers, 1999; news services, including RFID Journal, The New York
Walsham, 1993). Further, this perspective gave us the Times, The Wall Street Journal, InformationWeek, and
freedom to use theory as an initial guide (IDT, OVT, and Computerworld, among other similar sources. We also
IET in our paper) while also being open to the data subscribed to several RFID-specific blogs including
collected from the case. Thus, the interpretive tradition RFID Update, RFID Weblog, and RFID Gazette. We
allowed us to iterate between our data and the theory acquired white papers and analysis reports from
and, in the process, helped us refine, expand, or abandon consulting companies including IBM Business
the initial theory we started with (Walsham, 1995). Consulting, A.T. Kearney, AMR Research, and others.
We also obtained information from the RFID tracks of
We used a case study reconstructed using 20 years of
leading IS conferences (e.g., ICIS, HICSS, and
discourse surrounding RFID diffusion and promotion by
INFORMS). Finally, we kept records of public
Walmart in retailing. Based on the data (described in
statements on RFID released by Walmart 2 and other
detail below) we collected from the discourse
major players in retailing, such as Best Buy, Target,
surrounding RFID technology, we used the narrative
P&G, and Intermec to triangulate the validity of news
strategy to explicate the process of RFID diffusion. The
reports. These sources formed a cross-reference network
narrative strategy involves the construction of a story
to ensure that our data not only represented the discourse
from raw data (Bartis & Mitev, 2008; Langley, 1999;
surrounding RFID in retailing but were also detailed and
Myers, 1997; Pentland, 1999), which in our case is the
comprehensive. All the data were stored in a qualitative
data from the RFID discourse. Time is a key anchor
research database for further review, retrieval, and
point in this strategy (Langley, 1999). In this strategy,
interpretation.
the researcher develops a chronology of events that
serves to organize data and present theory (Langley,
1999). Using this strategy, we developed a chronology
4.4 Analysis Strategy
to uncover the timing and sequence of events We analyzed our data in three steps. First, using data
championed by certain actors that led to the diffusion of from the above sources, we developed a chronology of
RFID. Then we focused on Walmart to present a key events in RFID diffusion in retailing as they
narrative of the early-stage diffusion of RFID in evolved. The time horizon of the list ranged from
retailing. December 1999 when the future of RFID in retailing
was framed for the first time to the end of 2019 when
4.3 Data Collection RFID diffusion crossed the chasm. Thus, this list
provides a comprehensive snapshot at different points in
For our analysis, we collected data about RFID-in-
time as the diffusion of RFID in retailing unfolded.
retailing from the internet, disseminated through the
Table 3 presents a chronology of key events in the early-
following channels: trade press, academia, online
stage diffusion of RFID in retailing. We narrate these
journals, white papers, and analysis reports. We chose
events when we present the process model in the next
internet-based sources for the following reasons: First,
section. In particular, we focus on the key events of our
internet-based data sources have been used to capture
case that relate to actions of the four key actors—RFID
the status of RFID diffusion in research (Fish, 2006;
research institutes, retailers, suppliers, and RFID
Quaadgras, 2005). These sources are timeless and have
vendors—that pertain to the various institutional
paid great attention to Walmart’s RFID strategy.
entrepreneurship strategies and functions of OV in the
Second, most of the internet reports are based on official
context of early-stage diffusion of codependent
interviews conducted by journalists from leading media
innovations along with the development of the RFID-in-
outlets. Although we did not conduct interviews in
retailing institutional field. In our second step of data
person, we used the interviews that are publicly
analysis, we used our theoretical concepts as sensitizing
available. Finally, text on the internet is an essential part
instruments to analyze our research questions (Nielsen
of the RFID discourse that RFID players use as a
et al., 2014; Patton, 2014). We pooled all the text
reference.
documents and analyzed our data using the recursive
We conducted a systematic effort to access publicly pattern of interpretive qualitative research (Klein &
available information about RFID diffusion on the Myers, 1999; Miles et al., 2018; Miles & Huberman,
internet from 1999 to the end of 2019, thereby amassing 1994; Nielsen et al., 2014; Walsham, 1993).
20 years of public discourse on RFID diffusion. We

2 Walmart publishes its public statements in Walmart Facts


(http://www.walmartfacts.com), the official website for
Walmart Inc.

392
Journal of the Association for Information Systems

Table 2. Sources of RFID Texts Used in the Study


Source Purpose Examples Number

Online news To get insights into the coadopter InformationWeek, RFID Journal, Inbound Logistics, 178
services and relative advantage of RFID for Computerworld, CNET, Business Wire, CFO
online press coadopter communities and the Magazine, News.com, PRNewswire, Capgemini,
statements from reception of RFID technology by Creative Weblogging, SecureIDNews, Business Week,
companies various actors in the innovation Guardian Unlimited, IDTechEx Ltd, ID TechEx., CIO,
community, their discursive and ABI Research, BaseLine, RFID consortium,
influential activities with respect to CNNMoney, The Wall Street Journal, and PCWorld, as
RFID, and the extent of joint well as public statements from Walmart* and other
adoption of RFID. major players in RFID in retailing such as Impinj,
Kimberly-Clark, SUN Microsystems, and RFID-
specific blogs such as RFID Update, RFID Weblog,
and RFID Gazette

White papers To get insights into the costs and IBM Business Consulting, A.T. Kearney, AMR 47
and analysis benefits of RFID and eventually Research
reports from shed light on the issue of coadopter
consulting relative advantage.
companies

Academic To get insights into the factors International Conference on Information Systems 33
conferences related to the adoption of RFID and (ICIS), Americas Conference on Information Systems
the issues of coadopter relative (AMCIS), Hawaii International Conference on System
advantage and joint adoption. Sciences (HICSS)

Magazine To get insights into the diffusion CIO, InformationWeek, Retail Merchandiser, Supply 36
articles pattern of RFID, the actors taking Chain Management Review, Logistics Management,
part in RFID activities, and the Computerworld, CIO Insight, SETLabs Briefings, DM
powerful positions of and Review, Manufacturing Business Technology Focus,
dependence between various actors Telephony, DC Velocity, Manufacturing Engineering,
in retailing. In addition, these Pharmaceutical Technology, EDN, Computerworld,
sources also provided insights into Digital ID World, Communications of the Association
each actor’s discursive and for Information Systems, Business Week
influential activities with respect to
RFID.

Books To get insights into the technical RFID Sourcebook; Revolution at the Checkout 8
details of RFID as a technology, Counter; RFID: Applications, Security, and Privacy;
the business case and the and Wal-Mart World: The World's Biggest
economics surrounding RFID, and Corporation in the Global Economy.
the history of RFID technology.

Note: Walmart publishes its public statement in Walmart Facts (http://www.walmartfacts.com), the official website for Walmart Inc.

Table 3. Key Events in the Institutional Field Development of RFID in Retailing


Date Event
10/99 The Auto-ID Center was set up (Quinn, 2003).
6/03 Walmart mandated top suppliers to use RFID on pallets and cases (RFIDJournal, 2003).
10/03 EPCglobal was formed to manage EPC network and RFID standards (Violino, 2003d).
11/03 EPCglobal released the UHF Class 1 RFID Standard Draft (EPCGlobalInc, 2003).
11/03 Walmart detailed RFID mandate requirements to its top 100 suppliers (Becker, 2004; Wysong, 2005).
2/04 Target announced its RFID mandate plan (RFIDJournal, 2004a).
3/04 Albertsons announced its RFID mandate plan (PRNewswire, 2004).
4/04 Walmart launched the first RFID-tagged cases and pallets (Whitcomb, 2004a).
5/04 Walmart lowered the tagging rate expectation for suppliers to meet its deadline (Sliwa, 2004 ).

393
The Early-Stage Diffusion of Codependent IT Innovations

6/04 Walmart outlined a three-step RFID expansion plan (Whitcomb, 2004b).


8/04 Best Buy announced its RFID mandate plan (Roberti, 2004).
12/04 EPCglobal announced the ratification of the UHF Gen 2 RFID standard (RFIDUpdate, 2004).
1/05 Walmart presented its RFID mandate achievement (Roberti, 2005c).
4/05 The first EPC Gen 2 UHF product, Impinj’s RFID tag, was announced (O’Connor & Roberti, 2005).
10/05 Target and Walmart launched a pilot to share EPC data with suppliers (Roberti, 2005d).
4/06 Walmart announced it would no longer accept the Gen 1 tag after 1/30/2006 (O’Connor, 2006c).
7/06 ISO approved the EPC Gen 2 Class 1 UHF standard as an amendment to the ISO18000-6C (O’Connor, 2006a).
4/07 EPCglobal ratified the EPC Information Services specification (Bacheldor, 2007).
5/07 Metro asked 650 top suppliers to place EPCglobal Gen 2 RFID tags on all pallets (Roberti, 2007).
9/07 Walmart, Best Buy, and the entertainment industry conducted an item-level tagging pilot (Swedberg, 2007b).
10/07 Walmart unveiled three new RFID initiatives (O’Connor, 2007).
1/08 Sam's Club asked suppliers to use EPC Gen 2 RFID tags on single items (RFIDUpdate, 2008).
1/08 Walmart announced its first penalties for noncompliance to its mandates (RFIDJournal, 2008b).
1/09 Sam’s club relaxed the mandate deadlines and sharply reduced noncompliance fees (RFIDJournal, 2009b).
2/09 P&G pulls out of its landmark promotional campaign trials with Walmart (RFIDJournal, 2009a).
5/09 A research study by Dillard’s and the University of Arkansas establishes item-level RFID’s superiority to
manual inventory counts (Studio98Test, 2009).
8/09 Bloomingdale’s research study finds that item-level RFID improved accuracy by 27 percent and reduced labor
costs (O’Connor, 2009).
10/09 GS1 EPCglobal released RFID-based Electronic Article Surveillance (EAS) guides (GS1, 2009a).
7/10 Wal-Mart announced the adoption of item-level RFID to track clothing (Bustillo, 2010).
10/10 Major retailers and industry groups launch item-level RFID Guidelines Initiative (Swedberg, 2010).
07/11 Voluntary Interindustry Commerce Solutions Association’s (VICS) item-level RFID initiative enters phase II
(Swedberg, 2011c).
09/11 Macy’s announced the commencement of item-level tagging in 850 Stores (Swedberg, 2011a).
03/12 The representatives from MIT and Johnson & Johnson were elected to lead GS1 EPCglobal Board (GS1,
2012).
06/12 The University of Arkansas opens a new RFID research center (Studio98Test, 2012).
09/12 GS1 US and Voluntary Interindustry Commerce Solutions Association (VICS) merged their operations
(Edwards, 2012)
11/13 GS1 upgraded and ratified the Gen 2 RFID standard (GS1, 2013).
04/14 The RAIN RFID Alliance was launched by Google, Impinj, Intel, Smartrac, and AIM Global to promote RFID
adoption (RAINRFID, 2014).
09/14 Macy’s expands RFID deployments by tagging fashion items at its Macy’s and Bloomingdale’s stores
(Swedberg, 2014a).
01/16 Macy’s introduces the Pick to the Last Unit Program for omnichannel sales (Swedberg, 2016c).
03/16 American Apparel and Postmates offer one-hour delivery service using RFID visibility (Swedberg, 2016a).
10/16 Macy's announced that it would RFID-tag 100% of its inventory by 2017 (McKevitt, 2016a).
12/16 Target’s RFID investments lead to a massive increase in digital sales (Gillo, 2016).
04/17 Macy’s RFID effort initiative boosts fulfillment and sales (McKevitt, 2017).
10/18 Auburn University and GS1 US studied retailer-supplier data exchange and showed RFID’s superiority over
bar code (GS1US, 2018; Lopez, 2018).
03/19 Nike shared that RFID improves its product visibility (Cosgrove, 2019).
Note: We include only critical events above. We shaded rows that pertain to the institutional entrepreneur at each stage. Walmart played the IE
role until the chasm. During the chasm phase, Walmart and other big retailers played the IE role.

394
Journal of the Association for Information Systems

We identified four distinct phases heuristically based on performed when actors used financial, political, social,
the substance of the discourse and inspired by the and discursive resources to either gain support for the IT
distinct issues that needed to be addressed during those diffusion or to lever against the detractors (Hardy &
phases for the successful diffusion of RFID in retailing. Maguire, 2017). To code relationship development we
We identified the cutoff points for the four phases of coded all actions performed by an actor when they
early-stage diffusion based on the coadopter relative developed new relationships, such as collaborations,
advantage of RFID-in-retailing, specific OV functions coalitions, and alliances, to get other actors to help diffuse
enacted and IE strategies used by the IE at different the IT innovation (Hardy & Maguire, 2017). To code4
points in time, and the extent of joint adoption of the enacting interpretation we coded all the actions the IE
RFID technology by retailers and their suppliers. The performed to convey the essential features of the
four periods range from 1999, when the RFID idea innovations and their usefulness (Swanson & Ramiller,
surfaced, to 2019, when RFID diffusion crossed the 1997). To code enacting mobilization, we coded all of
chasm. The issues to be addressed in the four RFID actions the IE performed to motivate and involve other
phases we identified are congruent with the issues and actors in materially realizing the innovation (Swanson &
the discursive activities we identified by synthesizing Ramiller, 1997; Wang & Swanson, 2007). To code
the work on IDT (Rogers, 2003), and on phase models enacting legitimation, we coded all of the actions the IE
from IET (Morrill, 2006; Purdy & Gray, 2009) and OVT performed to develop, test, refine, and legitimate the
(Nielsen et al., 2014; Wang & Ramiller, 2009). innovation (Swanson & Ramiller, 1997; Wang &
Swanson, 2007).
In each of these phases and in each article, we searched
for the theoretical concepts we identified during our
review of theoretical foundations. We read each article in 5 Case Analysis and Theoretical
our data line by line, coded them, and summarized them Framework
based on these concepts and their relationships (presented
in Table 1) (Wang & Swanson, 2007). Table 1 provides 5.1 Institutional Change from UPC Bar
coding examples for the concepts of interest. To identify
the actors who played the role of an IE, the IE strategies
Codes to RFID in Retailing
they used, and the OV function they enacted, we focused Barcode technology has been the dominant institutional
on the actions of the four key organizational actors logic for supply chain management in retailing for the last
including RFID research institutes, retailers of consumer four decades (Harford, 2017). A barcode consists of
packaged goods, suppliers of consumer packaged goods, machine-readable black and white stripes printed on each
and RFID vendors because we found that these four product package (Harford, 2017). It encodes the universal
actors were deeply involved in the early-stage diffusion product code (UPC) that uniquely identifies a product
and shaping of the RFID-in-retailing institutional field. (Dodge & Kitchin, 2014). Upon scanning the code using
Table 4 presents the actors who were part of the RFID-in-
a laser scanner, a user can look up a product’s type and
retailing institutional field during the four phases of our
price from the stock database (Dodge & Kitchin, 2014).
study. We identified the actions performed by each actor
Retailers use barcode scanners to more accurately and
in these four phases and coded which of the three IE
strategies (rationale development, relationship efficiently track and automate inventory. With just half a
development, and resource mobilization) the actor cent spent on printing each barcode label, suppliers can
performed, and which of the three OV actions utilize inventory and sales data from retailers to improve
(interpretation, legitimation, and mobilization) the IE their demand forecasts and better plan their production
enacted by performing these actions. To attribute an (McCathie, 2004). By encoding traceability information
action to an actor, we determined the cited contributor for on the barcode, suppliers can enhance shipment tracking,
each coded text using the categories of research institutes, quality control, and product recalls (McCathie, 2004).
retailers, suppliers, and vendors (Abrahamson & However, barcodes have limitations, such as a minimal
Fairchild, 2001). In the cases where the coded text information-holding capacity, line-of-sight optical
belonged to white papers and analysis and press reports, scanning requiring human involvement, and low
we recorded the category as the organization that durability (McCathie, 2004). For supply chain
published these articles. To code3 rationale development management, this could mean that barcodes are labor-
we coded all actions performed by an actor where they intensive, prone to human errors, restricted in traceability
used the discursive realm to promote the IT innovation. and visibility, incapable of item-level tracking, and
To code resource mobilization, we coded all actions sensitive to environmental conditions (McCathie, 2004).

3 4
To reduce repetition, we briefly describe the strategies here, To reduce repetition, we briefly describe the functions here,
and provide more examples of each strategy in Section 2. See and provide examples of each function in Section 2. See
Table 1 for definitions of these strategies and coding Table 1 for definitions of these functions and coding
examples. examples.

395
The Early-Stage Diffusion of Codependent IT Innovations

Table 4. Actors in the RFID-in-Retailing Innovation Community in the US (1999–2019)


Retailers:
Retailers, members of the coadopter community that adopted one complementary part (RFID readers), served as the hub of the
institutional field of RFID in retailing because it was their mandates that triggered RFID diffusion. Though Walmart was the
first to issue the mandate and trigger RFID diffusion, other retail giants including Best Buy, Target, Tesco, Carrefour, Royal
Ahold, and Metro contributed to the diffusion by issuing mandates. Smaller retailers found the need to evaluate the impending
RFID revolution in retailing.
Suppliers:
Consumer packaged goods suppliers, members of the coadopter community that adopted one complementary part (RFID tags),
served as another pillar in the institutional field of RFID in retailing. Manufacturing giants like P&G, Gillette, and Kimberly
were the pioneers aiming for rewards from RFID-based inventory control, theft prevention, product promotion, and other
applications. Smaller suppliers aimed to improve their logistics management or were trying not to lose their contracts from
retailers who mandated RFID tagging.
RFID Vendors and Consultants:
Like other technology innovations, RFID diffusion provided a great business opportunity for technology vendors. New ventures
such as Intermec, Impinj, and Alien emerged and were supported by venture capitalists who trusted RFID’s potential. Established
semiconductor giants like NXP, NEC, Intel, Motorola, and Texas Instruments, as well as software providers like Sun, Oracle,
SAP, HP, and Microsoft entered the RFID business. Consulting companies like IBM, Ernst & Young, Accenture, A. K. Kearney,
and McKinney prepared for consulting opportunities related to RFID diffusion.
Research Institutes, Industry Associations, and Standard-setting Bodies:
Research institutions obtained funds to solve engineering problems, compile business cases, and study the business strategy of
RFID. Research institutes set up special research programs, opened seminars, and published academic papers focused on RFID.
Industry associations in retailing, such as Grocery Manufacturers Association (GMA) and Global Commerce Initiative (GCI),
invited universities and consulting companies to survey and predict the RFID trend. Standard-setting bodies strived to create a
mutual platform. As the most successful standards management institution in retailing, the UCC and EAN set up a joint venture,
EPCglobal, to extend their barcode expertise to RFID. The International Standard Organization (ISO) wanted to merge RFID
standards with its vast body of standards.
Media:
Media played the role of an information channel to convey the progress and frenzy in RFID diffusion. Mainstream business
journals like The New York Times, The Wall Street Journal, and Computerworld regarded RFID as a key venture to attract more
audiences. Internet media, such as RFID journals and RFID news blogs, profited from RFID-related commercials. Conference
organizers, such as the EPCglobal Annual Conference, the Retailing Industry Conference, and RFID Journal Live, profited from
organizations that wanted to either exhibit their RFID products or to gain knowledge about the latest technology and business
information about RFID.

Compared to barcodes, RFID technology is a stocks, fewer expired products, increased labor
significant advancement in representing and productivity, faster replenishment rates, and faster
communicating product information. RFID tags unloading times (Kaplan, 2018b; McCathie, 2004).
consist of an embedded circuit that encodes product The costs of RFID adoption for the retailers were a
information (such as price, unique identification, stock fixed cost of installing RFID readers and the
number, batch number, date of production, etc.) and an supporting IT infrastructure. Suppliers could benefit
antenna that broadcasts this data (Dodge & Kitchin, from RFID from promotions tracking, shipments
2014). An RFID reader can query the information tracking, theft prevention, accurate sales data from a
transmitted by the tag from a considerable distance, retailer, targeted recalls, and counterfeit protection
even if the RFID reader is not in the line of sight. The (McCathie, 2004). The costs of RFID adoption for
RFID technology addresses all the limitations of the suppliers are the recurring cost of attaching RFID tags
barcode technology. It is durable, reliable, accurate, to each product. However, their benefits are not
carries rich information that is updateable throughout immediate. Only economies of scale in the RFID tag
the supply chain, eliminates line-of-sight scanning, production resulting from huge demand can reduce the
affords item-level tracking and simultaneous costs for suppliers. In addition, suppliers bear the costs
automatic scanning, and requires less labor (McCathie, of tagging as they are most likely to attach the RFID
2004). tags in almost all supply chain implementations
because item-level tags need to be attached to the
RFID’s superior capability offered retailers and
carton wrapping of consumer packaged goods or sewn
suppliers a coadopter relative advantage over the
into clothing material (Basker & Simcoe, 2021;
barcode technology based on different sets of benefits
Gaukler et al., 2007). It is not economically feasible to
and costs. Retailers could benefit from RFID through
tag at the retailer location because the tagging and
enhanced product visibility, traceability, and
encoding equipment must be present at each location
availability, accurate sales estimates, reduced out-of-

396
Journal of the Association for Information Systems

(Basker & Simcoe, 2021; Gaukler et al., 2007). In introducing RFID-in-retailing was an institutional field-
essence, although RFID-in-retailing offered a shaping project for Walmart.
coadopter relative advantage to both retailer and
Introducing RFID meant that Walmart had to tackle
supplier communities, the coadopter relative
resistance from suppliers because Walmart’s coadopter
advantage was systematically higher for the retailer
relative advantage from adopting RFID readers would
community beginning with the early stage.
only bear fruit if suppliers started affixing RFID tags to
We noted earlier that an organization in the coadopter their product packaging. In the early stages, suppliers
community that perceived the highest degree of might have perceived risk in embracing a revolutionary
coadopter relative advantage (i.e., the retailer technology for one retailer only. Typically, one supplier
community) sensed the opportunity of an institutional supplies the same products to multiple retailers, many of
change from UPC bar codes to RFID in retailing. them large ones (Malone, 2012). If the rest of the
Among all the retailing players, Walmart perceived a retailers do not adopt RFID, the supplier has to
high coadopter relative advantage in adopting RFID-in- implement the new RFID technology for Walmart but
retailing from a very early stage. Walmart’s strategy is continues using the barcode technology with perhaps
to offer low prices to customers, reflected in its slogan even a larger share of their business (Malone, 2012).
“Everyday Low Prices” (Hyde, 2019). Cutting Such an investment might be too much of a risk because
the supplier’s returns from RFID may be for naught if a
operational costs through innovative retail supply chain
newer technology surpasses the RFID technology and
technologies has been one way Walmart can provide
retailers leapfrog to the new technology (Goldenberg &
low prices to customers—in fact, it was Walmart that
Oreg, 2007; Guo & Chen, 2018; Huang, 2011; Malone,
pioneered the diffusion of barcode technology (Hyde, 2012). For example, retailers might leapfrog to wireless
2019). Given all the potential benefits discussed above, sensors by sensing their superiority over RFID
Walmart sensed that RFID technology would help them technology (Mayberry, 2005). Because RFID provides
provide lower prices to their customers. Walmart felt no short-term benefits and higher costs for suppliers,
that it could leverage its market dominance, cutting- research institutes might find it challenging to obtain
edge expertise, and power over suppliers to diffuse funding from suppliers. Vendors might hesitate to
RFID. More importantly, Walmart perceived lower develop and market RFID solutions until a critical mass
costs with higher benefits compared to suppliers who of retailers and suppliers adopt the RFID tags and
had higher costs with no immediate benefits (Kaplan, readers (Basker & Simcoe, 2021; Gaukler et al., 2007;
2018b). Thus, Walmart perceived a high degree of Malone, 2012). Such a stalemate would deter standard-
coadopter relative advantage in RFID and focused on setting bodies, resulting in a lack of accepted standards
introducing RFID as a next-generation supply chain to move the technology forward. Given these tensions
technology in retailing. inherent to bringing about an institutional change in
support of RFID in retailing, we conducted a research
Walmart faced several challenges in introducing RFID project on Walmart’s efforts to influence the early-stage
in retailing. Barcodes had been firmly embedded in diffusion of the RFID-in-retailing innovation.
retailing for four decades. It was the supply chain
management standard around which retailers and 5.2 Process Model of the Early-Stage
suppliers trained their personnel and built their
infrastructure and processes (McCathie, 2004). For Diffusion of Codependent Innovations
suppliers, barcodes were easy to use and affordable Based on our case analysis of RFID diffusion in
(McCathie, 2004). It was a mature technology. Despite retailing we arrived at a four-phase process model of
that, there were ongoing innovations to advance the the early-stage diffusion of codependent innovations.
technology (McCathie, 2004). Universally accepted and We divided 5 the early stage into four phases 6 and
understood standards existed for encoding labels and the named them: emergence, structuralization, evolution,
design and development of barcode equipment and chasm. Figure 2 presents our abstract process
(McCathie, 2004). Most importantly, all actors accepted model. We present the case description of Walmart’s
the benefits of barcode technology in retailing. Thus, promotion of RFID in retailing below.

5 We agree with the IET literature that institutional field vein, Morrill (2006), whose work we adapt to designate the
conditions and the strategies used by organizations are in phases of early-stage diffusion, characterizes these phases as
constant flux, meaning that one is shaped by the other. The “conceptually distinct, yet empirically interpenetrating
OV literature posits that the diffusion process for an historical” phases.
6 We use the term “stage” to denote early-stage diffusion. We
innovation is always in flux as the discourse surrounding the
innovation evolves. However, we argue that there is still a use the term “phase” to denote the four phases of our process
sequential progression of the diffusion process into distinct model including emergence, structuralization, evolution, and
phases as the roles of the actors, particularly that of the IE, chasm. These phases occur within the early-stage diffusion
change to enact different functions of the OV. In the same of an innovation.

397
The Early-Stage Diffusion of Codependent IT Innovations

Legend: Rationale – rationale development institutional entrepreneurship


R – retailers (with Walmart being the dominant retailer) strategy
S – suppliers Resource – resource mobilization institutional entrepreneurship
V – vendors strategy
A – academia/research institutes/standard-setting bodies Relationship – relationship development institutional
IE – institutional entrepreneur entrepreneurship strategy
OV – organizing vision Green shaded circles – actual IE based on our theory
I – interpretation OV function Red shaded circles – expected IE based on prior theory
M – mobilization OV function Blue arrow – influential action
L – legitimation OV function Double-headed arrow – collaborative action

Figure 2. Four-Phase Process Model for the Early-Stage Diffusion of Codependent Innovation: Based on
the Case of Walmart’s Promotion of RFID in Retailing

In each phase, we presented the expected IE and their surrounding the innovation (Liao, 2016; Rogers, 2003;
actions based on our synthesis of the work on IDT Wang & Ramiller, 2009) and are expected to play the
(Rogers, 2003), and on phase models from IET role of an IE during this phase. At this point in the
(Morrill, 2006; Purdy & Gray, 2009) and OVT innovation process, it is these entities who have the
(Nielsen et al., 2014; Wang & Ramiller, 2009). Then, most knowledge about and the most interest in the
based on case analysis, we identified the actor that potential applications of the new technology for
turned out to be the IE, the institutional solving business problems. Therefore, it is these
entrepreneurship strategies they used, and the OV players who are expected to promote and spur the
function they enacted. community discourse toward the potential applications
and benefits of the innovation in the business world
5.2.1 Phase 1: Emergence (Rogers, 2003; Wang & Ramiller, 2009). These
The emergence phase in the early-stage diffusion entities perform research and development activities to
process for a codependent innovation is when an OV show the feasibility of the innovation and its business
around a new technology such as RFID starts taking value, draft standards, and develop hardware and
shape, and a community begins to form around this software for the innovation. They team up with
vision (Wang & Ramiller, 2009). It is during this phase adopters who fund these entities’ activities. Vendors
when concrete ideas for the application of a technology also observe these entities to understand how they can
in commercial endeavors start being developed build and sell these innovations (Nielsen et al., 2014;
(Greenwood et al., 2002). External influencers— Swanson & Ramiller, 1997). Such events, championed
research institutes and inventors—are accordingly by research institutions, signal the emergence of an
expected to take center stage in the institutional field institutional field. But in the case of RFID diffusion,
we found that it was Walmart, the coadopter of one

398
Journal of the Association for Information Systems

complementary part of the innovation that played the played an influential role in the Auto-ID Center’s quest
role of an IE and led the emergence of the RFID-in- for the RFID rationale (Roberti, 2002). Walmart
retailing institutional field. conducted both public and secret pilots (Roberti,
2003b, 2014) with interested suppliers to test RFID
Until October 1999, the RFID-in-retailing institutional
technology (Gilbert, 2003; Semilof, 2001). Walmart
field did not exist. There was neither an actor
created a buzz in the public discourse by using terms
community involved in the RFID-in-retailing
such as “smart shelves” (Gilbert, 2003).
innovation nor was there a public discourse around this
innovation. Early in the RFID development, it was Sanjay Sarma, a professor at the Massachusetts Institute
Walmart that realized the benefits of RFID. Walmart of Technology and a co-founder of the Auto-ID Center,
envisioned RFID as the next-generation supply-chain mentioned four turning points in RFID diffusion, one of
technology through which it could find efficient ways which happened during the emergence phase and used
to reduce its operational costs and build newer trade Walmart as the IE. Referring to Walmart joining the
models (Haberman, 1999). Suppliers were excited Auto-ID Center, he wrote:
about RFID as a tool for tracking shipments and
In 2001, Wal-Mart joined the Auto-ID
managing cargo theft but were concerned about the
Center, and this move would prove to be a
high cost of tagging (Violino, 2002). In spite of the
very important event. For a company
massive costs of installing RFID readers and the
renowned for its innovative but grounded
supporting infrastructure, Walmart decided to leverage
attitude to join a research effort meant that
RFID, given its cutting-edge IS expertise and prior
the research was getting close to
experience in pioneering several important supply
practicality. (Garfinkel & Rosenberg, 2006;
chain management technologies (Violino, 2002). So,
Wehr, 2005)
Walmart perceived the highest degree of coadopter
relative advantage. Walmart coordinated the knowledge-building
activities during this phase even though research
In Table 5 below, we present the tensions in the
institutes and inventors played a key role by serving as
retailer’s calculus of coadopter relative advantage in
a hub for creating knowledge about RFID. The Auto-
the RFID adoption decision. In Table 6 below, we
ID Center was looking at all possible applications of
present the tensions in the supplier’s calculus of
RFID, but Walmart pioneered the idea and tested the
coadopter relative advantage in the RFID adoption
use of RFID in retailing. Referring to Walmart’s
decision. We can note that retailers and suppliers
planned trials (which were later shelved) (Violino,
assessed different costs and benefits. Figure 3a, 3b, 3c,
2003b), Kevin Ashton, a Procter & Gamble executive
and 3d presents the interactions among actors in the
and executive director of the Auto-ID Center, said:
four phases—emergence, structuralization, evolution,
and the chasm—of early-stage diffusion of Until now, the Auto-ID Center’s tests have
codependent innovations in our case, namely involved identifying and tracking the
Walmart’s promotion of RFID in retailing. In order for location of shipping containers in
Walmart’s coadopter relative advantage to bear fruit, warehouses. The tests starting this month at
suppliers need to buy into the RFID-in-retailing Wal-Mart and Tesco will be among the first
codependent innovation. Walmart focused on building major trials of the technology developed
a case for RFID’s use in retail in the emergence phase with the Auto-ID Center to track individual
because Walmart’s initial RFID pilots with IBM in the items on store shelves ... It’s the last Russian
early 1990s faded out (Violino, 2002). RFID in doll. It’s the hardest thing, and it’s what
retailing was never commercialized due to the high we’ve been working toward. (Gilbert, 2003)
cost of technology, low sales volumes, lack of open
The role of vendors was minimal at this stage because
standards, and IBM’s financial troubles (Roberti,
they were still analyzing the feasibility of building and
2005a). The Auto-ID Center, a nonprofit research
selling RFID technology. The other efforts in this
institute (Roberti, 2002) located at the Massachusetts
period included RFID journal startups (Roberti,
Institute of Technology (MIT) (Roberti, 2002) was
2003a), RFID improvement efforts of scholars and
founded with support from suppliers, Walmart, other
universities (Lee et al., 2004), RFID startups (Impinj,
retailers, and standard-setting organizations. The
2000), and business media selling the idea of RFID
Auto-ID Center has researched topics such as RFID
(Anderson, 2002; 2004; Bacheldor, 2004; Collins,
standards, RFID production, building supporting
2003; Johnson, 2005; Jones et al., 2005; Kevin, 2005).
infrastructure, and harnessing methods (Quinn, 2003).
Overall, Walmart’s actions in funding research
Walmart was the first retail member of the Auto-ID
institutes and conducting trials in collaboration with
Center (Ashton, 2003; Violino, 2002). By funding and
research institutes and suppliers reflect the institutional
collaborating with the Auto-ID Center, Walmart
entrepreneurship strategy of rationale development.

399
The Early-Stage Diffusion of Codependent IT Innovations

Table 5. Tensions in the Retailer Calculus of Coadopter Relative Advantage in RFID Adoption Decision
Tensions Benefits Costs
Adopt • Reduce operational costs • Disruption in current processes based on barcode
• Build newer trade models technology
• Improve product tracking • Fixed costs of investment in RFID readers and supporting
• Improve inventory management Infrastructure
• No recurring cost of tagging • Sever relationship with suppliers
• Capture rich information on each • Negative publicity from press
product
• Uniquely identify products
• Reduce labor and time
• Increase inventory counts per year
• No line-of-sight requirement for
product scanning
• Reduced expired products
• Faster replenishment rates
• Reduce out-of-stocks
• Accurate sales estimates
• Faster unloading times
• Supplier’s dependence
Do not • No disruption in current operations, • Lost opportunity in championing a supply chain revolution
Adopt which is based on the standardized and • Labor-intensive inventory due to use of barcode
mature barcode technology • High human errors due to the use of barcode
• Maintain relationship with suppliers • Inventory count sensitive to environmental conditions due
• No investments needed in adopting to the use of barcode
• No investments needed in researching • Less frequent Inventory counts due to the use of barcode
and developing new technology • Line-of-sight requirement for product scanning due to the
use of barcode
• Impossible to achieve item-level tracking due to the use of
barcode
• Impossible to capture more information on products while
scanning due to the use of barcode
• Restricted product traceability
• Restricted product visibility
• Other retailers could leapfrog to another technology

Table 6. Tensions in the Supplier Calculus of Coadopter Relative Advantage in RFID Adoption Decision
Tensions Benefits Costs
Adopt • Tracking shipments • Recurring costs of tagging products with RFID tags
• Managing cargo-theft • Immature technology
• Long-term supply-chain optimization • High cost of RFID tags
• Short-term maintenance of revenue from • Risk of RFID becoming obsolete
Walmart • Disruption to current operations, which is based on the
• Reduce out-of-stocks standardized and mature barcode technology
• Promotions management • RFID less reliable compared to barcode
• Less chargebacks due to proof-of- • No short-term returns
shipment • Other retailers could leapfrog to another technology
• Accurate sales data from retailer • Lack of ROI
• Targeted recalls • Maintain two separate product encoding processes—one
• Counterfeit protection for RFID-based retailers and another for barcode-based
retailers
Do not • No increase in immediate cost • Loss of revenue due to losing business with Walmart
Adopt • No disruption in current operations, • Lost opportunity in long-term supply-chain optimization
which is based on the standardized and • Risk becoming technologically ineffective and obsolete
mature barcode technology • Long-term risk of becoming incompatible with RFID-
based retailers

400
Journal of the Association for Information Systems

Figure 3a. The Emergence Phase Figure 3b. The Structuralization Phase

Figure 3c. The Evolution Phase Figure 3d. The Chasm Phase
Legend: Green shaded circles – actual IE based on our theory
R – retailers (with Walmart being the dominant retailer) Red shaded circles – expected IE based on prior theory
S – suppliers Blue arrow – influential action
V – vendors Double-headed arrow – collaborative action
A – academia/research institutes/standard-setting bodies

Figure 3. Interactions among Actors in the Early-Stage Diffusion of Codependent Innovation: Based on the
Case of Walmart’s Promotion of RFID in Retailing

401
The Early-Stage Diffusion of Codependent IT Innovations

All of Walmart’s activities, such as research funding, consumer commodities has been the high
standards setting, and pilot testing, were geared toward cost of the equipment. For the tags alone,
providing an understanding of the RFID innovation and manufacturers can expect to pay 30 cents
its uses in retailing. Walmart’s efforts focused on two per tag at a minimum, according to one
main questions: What is RFID good for? How does developer, Alien Technology in Morgan
RFID work? Through these efforts, Walmart Hill, Calif., which is supplying Gillette with
accumulated and aggregated common knowledge about its tags. At that price, it’s unfeasible to
the emerging RFID-in-retailing institutional field. embed the tags on, say, a 50-cent bar of
Walmart’s actions led to buzzwords in the public soap. That’s why Alien Technology and
discourse. For example, when Walmart planned pilot others are working to lower the cost of tags
tests, one headline from a reputed online news magazine to 5 cents to 10 cents over the next several
said, “Major retailers to test ‘smart shelves’” (Gilbert, of years. Even so, Ashton said widespread
2003). Another expert called RFID tags “license plates” adoption of RFID could take until 2010—if
(Nelson, 2001). Chana Schoenberger of Forbes titled her it ever does occur. (Gilbert, 2003)
article “RFID: The Internet of Things.” Interpreting
Walmart’s pilots, she wrote, “Stores have eyes. Now 5.2.2 Phase 2: Structuralization
they’re getting ears and brains. Soon tiny wireless chips
The structuralization phase in the early-stage diffusion
stuck on shampoo bottles and jeans will track all that
process for a codependent innovation starts after the
you wear and buy” (Schoenberger, 2002). Walmart’s
initial discourse in the innovation community during the
actions during this phase represent an actor enacting the
emergence phase has led to the formation of the
OV function of interpretation.
fundamental ideas and notions about the potential
There were also a few instances where Walmart’s applications of the new technology for solving business
actions represented enacting mobilization and problems. While discourse will continue in the
legitimation. Walmart mobilized the actors in the community throughout the innovation life cycle, now is
institutional field by increasing their awareness about the time when the various actors in the innovation
RFID and encouraging them to act on this awareness. community need to take action to further the diffusion
However, the project was only getting close to agenda (Morrill, 2006; Purdy & Gray, 2009). During
practicality. Walmart championing the RFID project this phase, the vendors who produce the technologies
lent legitimacy to RFID, but the legitimation of RFID in that are part of the innovation are expected to play the
retailing was not fully there. Not convinced of the merits role of an IE as external influencers because they want
of RFID, Andy Robson, CHEP’s business development to sell their products to the adopter community (Rogers,
manager for supply chain information services, said: 2003; Wang & Ramiller, 2009). Based on the guidelines
“Certainly, we believe the Intermec/Marconi/Savi from research institutes and inventors, technology
solution that we’re now working with is the best vendors build and test their products (Nielsen et al.,
available in the market, but because it’s in a pilot 2014). Using marketing campaigns, vendors promote
context, we still believe there is further and better the benefits of the technology and try to sell the
technology coming down the line” (Douglas, 2002; innovation to adopters through marketing avenues such
Gilbert, 2003). This leads to our first proposition: as tech fairs, press releases, and advertisements (Paré et
al., 2020; Rogers, 2003; Swanson & Ramiller, 1997).
Proposition 1: The IE for a codependent innovation
Vendors also team up with research institutes,
will use the strategy of rationale development and
consultants, and adopters to learn how to improve the
enact the OV function of interpretation during the
technology and its production (Nielsen et al., 2014). The
emergence phase of the early-stage diffusion
innovation is realized during this phase in the form of
process for that codependent innovation.
vendor-developed products that are ready to be adopted
The extent of joint adoption at the end of this phase by prospective adopters. Adopters, some of which
could be best described as one restricted to pilot tests experiment with the vendors’ technology (Rogers,
by powerful retailers with research institutes and 2003), also contemplate the widespread adoption of the
interested suppliers. Suppliers would not adopt RFID innovation based on community discourse. But in the
technology until the cost of tagging was reduced and case of RFID diffusion, we found that it was Walmart,
they saw evidence of RFID benefits. Vendors would the coadopter of one complementary part of the
not achieve economies of scale in production for them innovation, that played the role of the IE and
to be able to provide low-cost RFID solutions until championed the structuralization of the RFID-in-
there was sufficient joint adoption. One CNET report retailing institutional field.
aptly summarized the extent of diffusion:
The ongoing pilots and other interpretation activities
A major barrier to the broad adoption of helped Walmart reassess the costs and benefits of
smart-shelf technology in the thin-margin RFID (InformationWeek, 2001; TalkBusiness, 2004;
businesses of mass-market retail and Murphy, 2003). In this phase, Walmart further
crystalized the benefits of RFID—better product

402
Journal of the Association for Information Systems

tracking and inventory management—perceived in the hand, Walmart encouraged the supply, affordability,
prior phase (TalkBusiness, 2004). As a downstream and maturity of the technology through vendor
site in the retail supply chain, Walmart did not have to motivation, investments in research, and standard-
bear the recurring cost of RFID tagging (Dutta et al., setting efforts. Walmart’s actions during this phase
2007). Walmart just needed to bear the fixed costs of reflect the institutional entrepreneurship strategy of
installing RFID readers and the supporting IT resource mobilization.
infrastructure. Being the largest retailer, Walmart had
Although vendors took the required steps at this stage,
suppliers depending on them for their business
Walmart played the influencer role in the diffusion of
(Roberti, 2003a). Suppliers were enthusiastic about
RFID during this phase. In fact, RFID vendors built
shipment tracking, promotions management, and
their RFID plans around Walmart, mainly through
reducing out-of-stocks (InformationWeek, 2001;
testing and cost cutting. Sun Microsystems, which
Murphy, 2005). However, suppliers had concerns
built a Walmart-compliant test center, was one such
about immature technology (TalkBusiness, 2004;
example (Sun, 2003). Suppliers had to comply even if
Violino, 2003c), the risk of RFID becoming obsolete
they did not want to because they were highly
(Violino, 2003c), disruption to current operations
dependent on Walmart. As Pete Abell, co-founder of
(Violino, 2003c), lesser reliability compared to
the EPC Group Ltd., an independent consulting
barcodes (Wired, 2003; Violino, 2003c), the high cost
company, frames it, “You can’t [ignore the mandate] if
of tags (Violino, 2003c), and the lack of short-term
10 percent to 40 percent of your business is going
returns given huge investments in RFID tagging
through Wal-Mart” (Roberti, 2003a). Even with
(TalkBusiness, 2004; Wehr, 2004). So, Walmart
research institutes and standard-setting bodies, it was
perceived the highest degree of coadopter relative
Walmart that played an influential role. Sanjay Sarma
advantage.
identified the mandate as the final turning point in
Walmart understood that actions on its part were RFID diffusion:
needed to get suppliers to adopt RFID tags in order for
The final, and perhaps most powerful, step
its coadopter relative advantage to bear fruit. At a
in the gathering momentum behind RFID
retailing conference in June 2003, Walmart issued a
was the Walmart “mandate” announced in
mandate asking its top suppliers to adopt RFID
the middle of 2003… Starting from there,
technology, which was a substantial push for the
the expansion would encompass the entire
diffusion of RFID-in-retailing innovation
supply chain. Nothing riveted attention
(RFIDJournal, 2003, 2004b). Walmart focused on
more than the largest corporation on the
pallet- and case-level RFID in its mandate with the
planet making an announcement of this
objective of improving the inventory accuracy in its
magnitude. (Garfinkel & Rosenberg, 2006;
warehouses as opposed to item-level tagging, which
Wehr, 2005)
would have meant disruption to its in-store operations
and even heavier investments in installing readers for In the structuralization phase, Walmart actively
store shelves and points of sale (O’Connor, 2014; contributed to the discourse by communicating its RFID
Roberti, 2010a). At the same time, using its massive adoption strategy. “By 2006, we will roll it out with all
resources, Walmart deployed RFID systems in almost suppliers,” said Walmart spokesman Tom Williams in
12 distribution centers and 1,000 stores. With the 2003 when Walmart first spelled out its mandate
tagged cases and pallets, Walmart conducted several (Roberti, 2003a). Such statements sketching the
pilots to evaluate RFID in real scenarios (Whitcomb & diffusion plan set the tone and trajectory for the actors
Gallagher, 2004). Eight consumer packaged goods in the structuralization phase. Walmart shaped the
manufacturers participated in an RFID trial in public discourse not only through official statements
Walmart’s distribution centers in April 2004 to assess about RFID mandates, but also in more tangible ways,
the technology in the real world (Whitcomb & such as involving resource commitment in its pilot tests
Gallagher, 2004). Walmart joined the board of and participating in EPCglobal. When some of
directors of EPCglobal, an organization chosen by Walmart’s suppliers were not able to meet the January
actors in the retail industry to develop standards for 2005 adoption deadline, Walmart tried to lower the
RFID (WalmartFact, 2005). Walmart urged vendors to proportion of tagging to 65% and postponed the
exhibit their products in industry conferences, deadline to keep its RFID initiative alive (Sliwa, 2004 ).
collaborate on pilot tests, and to give price discounts Also, when suppliers complained about the lack of
and technology support to suppliers. To ensure benefits, Walmart kept downplaying the cost of tagging
continued RFID diffusion, Walmart reduced the to avoid any hindrance to the diffusion (Vijayan &
number of pallets that required tagging and extended Brewin, 2003). Walmart indirectly shaped the discourse
the mandate deadline when suppliers complained by urging its suppliers to organize a forum to share the
about the lack of benefits (Wailgum, 2004). Overall, lessons learned from the field operation.
on one hand, Walmart substantially raised the demand
for RFID technology through mandates. On the other

403
The Early-Stage Diffusion of Codependent IT Innovations

The mandate effectively mobilized the actors in the 2005b). Walmart’s market power downplayed the
institutional field surrounding RFID. Walmart’s concerns about the mandated diffusion model. For
influence on suppliers and other retailers meant that example, when confronted with the lack of common
vendors envisioned a large and sustainable market for protocols and standards around RFID, Keith Mahoney,
RFID. This market potential attracted vast sums of Walmart’s food logistics controller, said, “We could
capital that were invested in RFID start-ups. Soon after not allow the lack of them to hang up the project”
Walmart imposed the mandate, other major retailers (Vijayan & Brewin, 2003). This leads to our second
such as Target (RFIDJournal, 2004a), Best Buy proposition:
(Roberti, 2004), Albertsons (PRNewswire, 2004),
Tesco (IDTechEx, 2005a), and Metro (RFIDUpdate, Proposition 2: The IE for a codependent innovation
2005) issued similar mandates and RFID pilot will use the strategy of resource mobilization and
requests. Giant manufacturers also became eager to enact the OV function of mobilization during the
carry out RFID pilot tests. This market potential also structuralization phase of the early-stage diffusion
stimulated RFID vendors to release new products process for that codependent innovation.
(Capgemini, 2004) and provide a substantial discount
for suppliers who purchased RFID tags in bulk, with The extent of joint adoption at the end of this phase
discounts in some cases ranging up to 70% (Swedberg, was characterized by suppliers adopting RFID tags
2007a). The competition among RFID vendors also purely to satisfy retailers’ mandates. Suppliers faced
emerged because of Walmart’s strategic actions. problems such as the immaturity of RFID standards
Vendors started soliciting input from suppliers in and the high costs of RFID products. For small
devising their solutions and began offering attractive suppliers, tight budgets and lack of expertise made
discounts. The slap-and-ship7 solution was invented to compliance impossible, even by the deadline. The
attract suppliers who wanted to comply with retailers, including Walmart, installed RFID
Walmart’s mandate at the lowest cost (Swedberg, equipment in their stores and distribution centers. In
2007a). Small suppliers also began seeking affordable March 2005, Walmart made a statement to
RFID compliance solutions. The standard-setting summarize its mandate: 127 suppliers sent almost 5
body, EPCglobal, was set up and began to release million tags to Walmart (Whitcomb & Gallagher,
standards to meet Walmart’s pace (RFIDUpdate, 2005). Walmart itself deployed 14,000 sets of RFID
2004). The media started following the RFID equipment in its 104 stores and 35 Sam’s Club
compliance movement in a more concerted way and locations (IDTechEx, 2005b). Patrick Ervin, Vice
started guessing the result of the 2005-Sunrise mandate President of Business Development, Retail Supply
(IDTechEx, 2005b; Feder, 2004; Hill, 2004 ; Chain Vertical, Alien Technology’s aptly
Incucomm, 2005). summarized the extent of joint adoption:

Based on the above discussion, Walmart’s actions It’s clear that Wal-Mart is more committed
than ever. They’re going to another 200
during this phase represent an actor enacting the OV
vendors. They’re adding store sites and
function of mobilization. “The fact that the largest
they’re increasing the number of SKUs that
company in the world is publicly adopting EPC open
they’re tagging. They have a large
standards should give companies confidence that the investment and they’re not backing off.
day of a single, interoperable RFID system is close at (RFIDJournal, 2005)
hand,” said Kevin Ashton, executive director of the
Auto-ID Center (Vijayan & Brewin, 2003). Walmart’s 5.2.3 Phase 3: Evolution
mandated diffusion model provided an answer for how
RFID diffusion can happen in retailing. Walmart’s The excitement and action that was generated during
actions denoting interpretation and legitimation were the structuralization phase may taper when
less prevalent in this phase. Rather than focusing on implementation complexities and technology glitches
are found during early implementations of the
what RFID technology could do, Walmart was keen on
technology, and if the return on investment (ROI) is
mobilizing the field to get RFID adopted. It did so by
not in line with what was touted. This dissonance
focusing on the feasibility and success of its mandate.
between vision and reality marks the transition of the
At the end of this phase, reflecting on the RFID early-stage diffusion process into the new evolution
strategy, Linda Dillman, executive vice president and phase. Now, is the time to focus on increasing the
chief information officer for Walmart, summed up the normative appeal of the innovation and convincing the
mandate: “I don’t have to stand up here and debate innovation community that these early problems are
whether it will happen, because it did” (IDTechEx, nothing to be disheartened about and that the ROI can

7 Slap-and-ship is a method for consumer packaged goods required little data integration, it was preferred by suppliers
manufacturers to implement RFID by affixing tags on goods who only wanted to meet the mandate at the lowest cost.
just before being shipped to retailers. Since this method

404
Journal of the Association for Information Systems

be generated for all parties by making suitable with internal assessments shifted Walmart’s RFID focus
improvements to technologies, implementation from distribution-center use to in-store use (Songini,
methods, and organizational processes and practices 2007b) and expanded its RFID deployment plan
(Marsan et al., 2020; Nielsen et al., 2014). (WalmartFact, 2006). Second, Walmart collaborated
Accordingly, during the evolution phase, suppliers are with another retailer, Target for a joint pilot to share
expected to play the role of an IE due to their own vested EPC data with suppliers (Roberti, 2005d). Third,
interests at this point in terms of reducing their own costs Walmart pursued actions to mature, standardize and
and even gaining advantage over their competitors from spur demand for the technology. Walmart upgraded its
the innovation that has now been accepted as a valid facilities to next-generation RFID standards and asked
innovation in the field (Rogers, 2003; Wang & Ramiller, suppliers to upgrade to these RFID standards citing
2009). Suppliers are expected to work with academia better performance and lower cost (O’Connor, 2006c).
and research institutes to devise standards and improve They polled suppliers to learn about their readiness in
the innovation’s ROI (Purdy & Gray, 2009; Tolbert & adopting the new standards (O’Connor, 2006c). Inspired
Zucker, 1983). Suppliers are also expected to convey the by this move, some RFID chip manufacturers began
results of pilot tests to the community to attract more introducing new RFID products. “We wanted to kick-
retail customers and more retail business from current start the industry, so to speak, which would help spur
customers touting the benefits of the RFID technology. demand for our silicon,” said Impinj president and chief
As more suppliers foster the innovation, vendors can executive officer William Colleran during the unveiling
offer RFID tags at lower costs. However, contrary to of next-generation tags and readers (O’Connor &
these expectations, we found that in the case of RFID Roberti, 2005). However, these actions never
diffusion, it was Walmart, the coadopter of one established the ROI for suppliers. Except for a few
complementary part of the innovation, who played the suppliers like Kimberly-Clark, Unilever, and P&G, who
role of an IE and led to the evolution of the RFID-in- declared benefits from RFID (KimberlyClark, 2006;
retailing institutional field. Songini, 2007a), most suppliers still suffered from
unstable technology, high costs, implementation
Among the adopters, it was again Walmart that had the
problems, and few benefits (Fabris, 2007; Wailgum,
highest degree of coadopter relative advantage from
2004).
RFID diffusion. Walmart already significantly invested
in the RFID project by installing RFID equipment in There was less cooperation from other actors in the
stores, distribution centers, and Sam’s Club locations. institutional field. Proactive suppliers began
Walmart could get more out of the investments only withdrawing radical RFID implementation plans, and
when more suppliers adopt RFID because there would many smaller suppliers only minimally adopted RFID.
be more data to capture, share, and leverage within and The installed base resulting from Walmart’s mandate
across organizations (Malone, 2012). Suppliers also saw was not big enough to achieve economies of scale since
benefits in RFID. The benefits for suppliers included most suppliers took a passive attitude and purchased as
maintenance of business with Walmart, and long-term little as they could (Fabris, 2007). Many vendors began
supply chain optimization if the RFID technology postponing new product releases, citing the costs
sustains and matures. In addition, a supplier might lose involved in producing equipment according to new
its competitive advantage by getting technologically standards and minimal adoption from suppliers
obsolete if other suppliers start adopting RFID and (O’Connor, 2006b; Roberti, 2005b). “It’s a classic Catch
reduce their supply chain costs. However, compared to 22 situation. In order for prices on radio frequency
the structuralization phase, suppliers sensed even higher identification (RFID) tags to fall to what has been
costs and the lack of ROI in an immature technology that described as the ‘magic 5-cent’ mark, manufacturers
is primarily going to benefit the retailers. Walmart need to see more wide-spread adoption of the
understood further actions were needed to convince technology”, wrote Mel Duvall, a veteran business and
suppliers that there is a good ROI from adopting RFID- technology journalist (Duvall, 2007a). Other retailers
in-retailing technology because its coadopter relative retreated from an aggressive implementation plan even
advantage was contingent upon suppliers also adopting after they had confirmed the benefits of RFID through
RFID tags. pilot tests (Duvall, 2007b). Complaints about the RFID
mandate became a public secret (Gardner, 2004;
Walmart’s actions during this phase reflect the
Hudson; McWilliams, 2007). Rapid introduction of
institutional entrepreneurship strategy of relationship
newer standards (O’Connor, 2006a), patent wars,
development. First, Walmart took several steps to
royalty declaration issues, and a short time gap to
identify and assess the ROI in RFID implementation for
transition between standards made the above challenges
suppliers. Walmart commissioned a research study with
more daunting (O’Connor, 2006b; Roberti, 2005b).
the University of Arkansas. The result of the study—
reduction in out-of-stocks, faster replenishment rate, and Walmart understood that some concrete actions were
reduction in manual orders—encouraged the entire again needed to ensure that their RFID initiative did not
community (WalmartFact, 2005). This finding coupled fizzle out. It saw the potential benefits of RFID for itself

405
The Early-Stage Diffusion of Codependent IT Innovations

but also understood the hesitation of other members of category (O’Connor, 2007). Dubbing this move as a
the RFID community, particularly those of suppliers and “change in focus,” Carolyn Walton, Walmart’s VP of
RFID vendors, in investing in RFID due to a lack of IT said: “We’re coming at [RFID] from a different
ROI. To break this deadlock, Walmart came up with angle” (O’Connor, 2007). This leads to our third
three initiatives to continue extending RFID diffusion. proposition:
In the middle of 2007, it further pushed the agenda of
Proposition 3: The IE for a codependent innovation
RFID-in-retailing by unveiling the three new initiatives:
will use the strategy of relationship development
mandating RFID use on entire specific product and enact the OV function of legitimation during
categories, deploying whole-process RFID in Sam’s the evolution phase of the early-stage diffusion
Club, and applying RFID on weekly product process for that codependent innovation.
promotions (O’Connor, 2007). The initiative in Sam’s
Club was rolled out in January 2008 (RFIDUpdate, In January 2008, in order to take a firmer stance on
2008). During this phase, Walmart pursued its RFID RFID, Walmart followed up its new mandate by
initiative using its technical capabilities, resources, and imposing deadlines, and for the very first time, it
dominant power to provide legitimacy to the RFID-in- charged a service fee of up to $3 per nontagged pallet
retailing innovation and to continue promoting the (RFIDJournal, 2009b). Compared to how they reacted
to Walmart’s 2003 mandate, suppliers did not show
diffusion of this codependent innovation when doubts
great urgency, considering how accommodating
about it started emerging.
Walmart had been for noncompliant suppliers in the
Walmart’s actions during this phase reflect an actor prior mandate (RFIDJournal, 2008a). Suppliers
enacting the OV function of legitimation. Walmart’s wanted to evaluate the benefits of the new mandate
actions focused more on answering the following requirements (RFIDJournal, 2008a). Vendors were
question for the actor community: How can the value getting demand for RFID tagging mainly from
of RFID be realized? The community asked for Walmart and Sam’s Club suppliers (RFIDJournal,
financial justification for their RFID adoption. 2008b). However, in January 2009, Walmart
Walmart projected the next-generation Gen 2 significantly relaxed its mandate requirements,
technology as the answer to the financial justification reducing the fee and indefinitely postponing the
(RFIDUpdate, 2004). Walmart hosted a chief deadline after considering the lack of benefits for
suppliers and the time required to comply
information officer summit and announced the change
(RFIDJournal, 2009b). Sam’s Club established a
from Gen 1 to Gen 2 tags. Simon Langford, Walmart’s
supplier council to help suppliers identify benefits
manager of RFID strategies, called the new tags a
from each type of RFID tagging (RFIDJournal,
“breakthrough” (O’Connor, 2006c), thereby shaping 2009b). As another big blow to Walmart’s campaign,
the discourse. To add legitimacy, he quashed the P&G withdrew from the pilot programs it was
current reports about ROI and added, “I think a lot of conducting with Walmart (RFIDJournal, 2009a;
the [tag] studies that have been done prior [to the SCDigest, 2009)—an event that industry experts felt
introduction of near-field tags] are invalid, and they might have informed the community about the lack of
need to be done again” (O’Connor, 2006c). He benefits of RFID technology (FierceRetail, 2009).
concluded by saying, “The whole [playing] field has Overall, Walmart overestimated its ability to drive
been changed now” (O’Connor, 2006c). By forcing the institutional change—it overpromised but
Gen 1 to Gen 2 standards change, Walmart promoted underdelivered by setting up a very ambitious timeline
chip manufacturers to design and display their for its RFID vision.
products at major summits (O’Connor, 2006c). The
At the end of the evolution phase, all actors were
promotions were aimed at how the new products would
dormant with respect to their RFID activities. Walmart
reduce the cost of RFID. During one of their product chose not to unveil any new initiative until it worked
demonstrations, RFID manufacturer Impinj’s closely with the suppliers and found RFID benefits for
president and chief executive officer William Colleran them. Suppliers, vendors, and research institutes were
mentioned, “An RFID reader is no more complicated waiting for the next definitive move from Walmart.
than an 802.11 Wi-Fi router. … And we’ve seen the The extent of joint adoption at the end of this phase
price of Wi-Fi routers fall from many hundreds of could be best described as one where retailers
dollars to less than $100” (O’Connor & Roberti, 2005). (including Walmart) were cautious about the further
Walmart tried hard to lead the whole community installation of RFID systems in their distribution
toward novel scenarios in which the benefits of RFID centers. Suppliers adopted the minimum required
might be realized. At the beginning of 2008, at an technology. Compared to what it had originally
annual member conference, Walmart spelled out three envisioned, Walmart installed fewer RFID systems in
new initiatives: A mandate to Sam’s Club’s suppliers, its own distribution centers and influenced fewer
applying RFID in weekly promotions, and pilots to test suppliers to embrace RFID. One industry expert said:
if RFID can improve sales across an entire product

406
Journal of the Association for Information Systems

Five years into the RFID revolution, it’s 1997). Eventually, vendors build a marketplace and a
clear that implementation is occurring support infrastructure around the innovation that appeals
more slowly than Walmart originally to the mainstream adopters. Once acquired, these growth-
envisioned…. Today, 600 of Walmart’s stage adopters show loyalty to the innovation by
60,000 suppliers (plus 750 Sam’s Club standardizing it in their operations, appreciating it, and
suppliers) have deployed RFID. On influencing other suppliers, thereby accelerating diffusion
Walmart’s end, RFID has been rolled out at (Akhlaghpour & Lapointe, 2018). However, contrary to
1,000 of the roughly 4,000 Walmart and these expectations, we found that in the case of RFID
Sam’s Club stores in the United States. diffusion, Walmart and other retailers, the members of
Walmart’s original goal was to have 12 of one complementary part of the innovation, who played
its roughly 120 distribution centers outfitted the role of an IE in the early phases also helped the RFID-
for RFID by 2006. Today, only five are set in-retailing institutional field cross the chasm (Moore,
up for RFID, as the company has shifted its 2019).
RFID focus to in-store implementations.
(Webster, 2008) After two years in oblivion, Walmart took a new
approach. Based on pilot tests, Walmart realized that
Walmart abandoned the RFID mandate considering the RFID benefits lie in item-level tracking in stores,
cost of alienating the suppliers and the flak it received particularly in the apparel section (Roberti, 2010a). Item-
from the press (Roberti, 2010b). Walmart did not want to level tracking uniquely identifies a product and
force suppliers with additional expenses given the global distinguishes product varieties. Walmart was able to
financial crisis (Roberti, 2010b). The innovation diffusion perform frequent, efficient, and cheaper inventory
literature shows a chasm between the early stage and the management, and trace a product from the backroom to
growth stage of the diffusion process (Moore, 2014). It, the store shelf and checkout, leading to inventory
therefore, does not come as a surprise to know that the accuracy, on-shelf availability, and granular product
diffusion of RFID in retailing suffered a similar fate visibility in higher orders of magnitude than manual
(Malone, 2012; Roberti, 2017). The diffusion went into counting and barcoding (Roberti, 2010a). This new
an abyss and near death only to rise again.
approach played to the strengths of RFID, as opposed to
the pallet-level tagging enforced by Walmart in its earlier
5.2.4 Phase 4: Chasm
mandates. Pallet-level tagging focused on maintaining
The innovator suppliers’ success stories to legitimize the inventory accuracy for shipments arriving at warehouses,
innovation do not serve as good references for suppliers where the accuracy was already at 99% (O’Connor,
in the growth stage because the latter are risk averse, price 2014). It is at the stores where accuracy is impacted due
sensitive, fear change, and look for incremental to various factors such as customers, nonrepeatable
performance improvements (Moore, 2014; Rogers, 2003; processes, employee distractions, weather conditions, and
Xu et al., 2017). These adopters buy whole products that buying patterns (O’Connor, 2014). Other retailers like
are easily implementable, backed by mature technology, Macy’s, Zara, Target, and JC Penney conducted pilots
and sold by leading vendors. Innovations often do not and understood the benefits and future use cases of item-
cross the early stage, which results in a chasm because level RFID (Studio98Test, 2009; University of Arkansas,
growth-stage adopters are a self-referencing group; they 2009; Kay, 2010; O’Connor, 2009). Speaking about the
will not adopt until the ROI is confirmed by adopters from benefits of item-level RFID at a retailer conference, Peter
their category (Xu et al., 2017). During this phase, Longo, president of logistics and operations at Macy’s
vendors are the external influencers and act as IEs said, “no matter how good the bar code is, it is woefully
because they can gain substantial profit and growth by short of what you can do with RFID” (Roberti, 2011).
winning over growth-stage adopters (Mola et al., 2022). Regarding the benefits from Macy’s RFID trials, he
They devote resources to finding a use case for a niche stated: “Inventory accuracy went from 70 to 75 percent to
adopter segment, around whom they develop the 95 to 99 percent. … That’s practically stunning. We see
innovation and its support infrastructure (manuals, that as the most important benefit” (Roberti, 2011).
service partners, and additional products) (Mola et al., Summarizing his views on the technology’s future, he
2022; Moore, 2014). They establish relationships with said: “A year ago, we used to talk about whether to use
adopters to convince them and to directly sell their RFID or not” and added, “The debate is over. Our
innovation to them, often working closely with them for position is that the technology is here, and it is time. RFID
installation (Gombault et al., 2016; Moore, 2014). They is no longer a jazzy technology being tested by academics
forge tactical alliances with other vendors and consultants and consultants. It’s not acceptable to have 30 percent
for delivering sales support and related products (Marsan inventory inaccuracy” (Roberti, 2011).
et al., 2020). They dominate the public discourse to
Suppliers had higher benefits and lower costs in this phase
convey their market leadership, the innovation’s ROI, and
compared to the previous phases. Unlike barcodes, item-
the availability of customized solutions tailored to suit
level RFID benefits suppliers by letting them track
each adopter (Marsan et al., 2020; Swanson & Ramiller,
promotions, automate assembly, lower chargebacks from

407
The Early-Stage Diffusion of Codependent IT Innovations

reduced shrinkage and shipping errors, derive sales Warry, 2015; Zaino, 2016). Retailers continued to make
insights using retailer’s real-time sales data, and increase a case by trying more RFID use cases such as closed-loop
sales due to product availability in the retail sales floor supply chains, omnichannel sales, smart dressing rooms
(Kay, 2010; Roberti, 2011; Swedberg, 2011a). The costs (Moore, 2019; Swedberg, 2011c), nonreplenishment
were lower because of the technology maturity items (Swedberg, 2014a), pick-to-the-last-unit programs
(Wailgum, 2010), the emergence of established vendors (Swedberg, 2016c), one-hour delivery (Swedberg,
(Kay, 2010; Roberti, 2010a), and the retailer’s 2016a), and smart product displays in showrooms
willingness to share the tagging costs, subsidize the (Swedberg, 2016d).
sensor costs and work closely with the former in RFID
Retailers’ actions also reflected the institutional
installation (Bustillo, 2010; Roberti, 2010a). Suppliers
entrepreneurship strategy of relationship development. In
could see the benefits this time due not to the cost of
collaboration with research institutes, retailers conducted
losing business with influential retailers for
pilots (O’Connor, 2009), identified use cases, developed
noncompliance but due to RFID’s superiority compared
standards (Swedberg, 2011c), and tailored technology for
to the barcode and their cheaper cost (Bustillo, 2010; Kay,
specific use cases (Roberti, 2016a). Retailers worked
2010). Suppliers sensed that they would lose out on the
closely with suppliers by troubleshooting (Swedberg,
competitive advantage if they did not invest in RFID
2011a), providing tagging guides and standards manuals
(Kay, 2010).
(Beck, 2018), and involving business managers in
Although retailers and suppliers had their respective decision-making (Roberti, 2010a). Retailers got the best
coadopter relative advantage, it was again the retailers price on tags for suppliers by forecasting the volume of
that had the highest degree of coadopter relative tags to be bought for a particular product category and
advantage. The pilot studies were retailer initiated and sharing the tagging costs with a cost-of-goods model
retailer centric. The supplier benefits were primarily (Roberti, 2010a). In the process, retailers gained the
industry predictions based on the retailer pilots, with supplier’s trust, which had earlier been lacking (Kaplan,
limited references within the supplier community (Kay, 2018a). Retailers formed industry groups with other
2010). Suppliers had the cost of maintaining two supply retailers and invited participation from vendors and
chains—building a new one for retailers beginning their suppliers (Swedberg, 2010).
RFID rollouts and another for retailers using barcodes.
Retailers’ actions also reflected the IE strategy of resource
Moreover, the cost-sharing arrangements and the
mobilization. Retailers invested in RFID rollouts by
supplier-side installation were still in flux (Kay, 2010).
installing RFID readers, redesigning processes, and
Thus, despite their optimism, suppliers looked up to
building data handling capability (Bustillo, 2010). Early
retailers for the concrete actions needed for the coadopter
adopters like Walmart and Macys served as opinion
advantage to materialize for both parties. As an expert in
leaders for the early majority to join the RFID bandwagon
RFID in retail aptly summarized:
(Bustillo, 2010; Kriz, 2018; Thau, 2017). They invested
Simply put, this is an instance where everyone in research and development activities by funding
is a winner. Even apparel suppliers currently research centers (Studio98Test, 2012), sharing data
concerned about whether retailers will pay an (Beck, 2018; Lopez, 2018), conducting field trials
appropriate share of tagging costs would (Swedberg, 2010), and piloting new use cases (Moore,
have difficulty disputing the fact that RFID- 2019; Swedberg, 2011c). They channeled their human
based processes are superior to barcode- resources by training employees, appointing ambassadors
based processes. (Kay, 2010) (Greengard, 2015), creating specialized job positions
such as chief omnichannel officer (Prince, 2013), and
The actions of Walmart and the other retailers represent
showing senior leadership commitment (Zaino, 2017).
all three IE strategies. Retailers’ actions reflect the
Retailers brought in vendors to sponsor such RFID
institutional entrepreneurship strategy of rationale
programs, letting them have a say in standards
development. When RFID diffusion went to a standstill,
(Swedberg, 2011c). By using their power in the retailing
the retailers conducted pilots and identified RFID’s
field, they publicly backed RFID (Bustillo, 2010), issued
niche—in-store item-level tracking in the apparel
mandates to suppliers (Anderson, 2016; Zaino, 2017) and
section—in the retail supply chain and attracted the field’s
influenced standards development (Edwards, 2012). To
attention (Bustillo, 2010; Studio98Test, 2009; University
get suppliers’ support and cooperation in RFID adoption,
of Arkansas, 2009). Retailers formed and co-chaired
retailers invested in significant outreach efforts and
industry groups that served as knowledge hubs drafting
provided resources (McBeath, 2018).
detailed tagging guidelines and sharing them publicly on
their websites (Swedberg, 2010). These resources helped Vendors played a vital role in this phase. For instance, the
adopters in their RFID rollouts. Retailers shared their RAIN RFID alliance formed by vendors (such as Google,
success stories and lessons learned in conferences Impinj, Intel, and Smartrac) was a pivotal event aimed at
(Anderson, 2016), research studies, and case studies, accelerating RFID adoption (RAINRFID, 2014;
which served as a reference for other retailers and spurred Greengard, 2015). Vendors kept advancing the
RFID adoption (McKevitt, 2016b; Swedberg, 2016b; technology (Cook, 2014; Swedberg, 2013; Warry, 2015),

408
Journal of the Association for Information Systems

built the support infrastructure (GS1, 2009b; Greengard, (Hardgrave, 2016b), “one-hour delivery” (Swedberg,
2015), and tailored solutions to suit adopters (Woodie, 2016a), and “smart dressing rooms” (Moore, 2019) to
2016). With the increase in RFID demand, vendors frame the use cases. These actions represent an actor
earned profits, achieved economies of scale in production enacting the OV function of interpretation.
(Warry, 2015), and offered solutions at affordable prices
In the prior phases, RFID was a technology solution
(Kriz, 2018; Swedberg, 2016b). Industry leaders emerged
looking for a business problem. However, the IE’s ROI
in the RFID market (Zaczkiewicz, 2016). They published
studies in the chasm established item-level RFID as a
case studies for potential adopters (Kay, 2010), piloted
technology solution for actual business problems,
new use cases (Zaino, 2016), offered new RFID solutions
many of which remained unsolved by barcodes until
and services (Woodie, 2016), and collaborated for
that point (Bustillo, 2010; Kay, 2010; O’Connor,
standards development, research (Swedberg, 2011c), and
2009). As a result, coadopters focused on tailoring
installations (Prince, 2013). Vendors gained experience
RFID to their operations instead of pondering whether
with RFID implementations, learning best practices, and
RFID would deliver ROI, signaling a mindset change
common pitfalls (Kay, 2010; McBeath, 2018).
compared to prior phases (Greengard, 2015). Retailers’
Suppliers primarily played a supporting role by public announcements about extensive RFID rollout
embracing RFID (Bustillo, 2010; Kay, 2010), fulfilling plans and their rationale lent legitimacy to item-level
tagging requirements (Zaino, 2017), collaborating on tracking in retail (Bustillo, 2010; Greengard, 2015;
research (Swedberg, 2011b), and actively participating in Kay, 2010). For instance, one headline stated: “Macy’s
industry groups (Swedberg, 2011c). There are also to RFID tag everything” (Anderson, 2016). Suppliers
instances where suppliers played an influential role. A envisioned RFID as a mainstay of retailing due to such
Johnson & Johnson representative was elected as the public backing by retailing giants (Bustillo, 2010;
vice-chairman of the Board of Governors of GS1 Greengard, 2015). To lend further legitimacy, retailers
EPCglobal, a nonprofit RFID standards organization played a leadership role in standards development and
(GS1, 2012). Realizing RFID’s future, some suppliers maturation (Swedberg, 2011c). They participated in
big conferences (Anderson, 2016) and announced their
even tagged products for retailers who had yet to adopt
post-implementation success (McKevitt, 2016b).
RFID, even influencing those retailers resisting RFID
These actions represent an actor enacting the OV
(Zaino, 2017). Suppliers like Nike (Cosgrove, 2019),
function of legitimation.
Reebok (Swedberg, 2016d), Macy’s (Swedberg, 2014a),
and Rebecca Minkoff (Swedberg, 2014b) tried novel use Retailers fostered joint adoption by installing RFID and
cases and announced positive outcomes (Roberti, 2016b; simultaneously issuing mandates to suppliers (Bustillo,
Wollenhaupt, 2018). 2010; Kay, 2010), working closely with them in
fulfilling the tagging requirements (Anderson, 2016;
Research institutes collaborated with other actors to Roberti, 2016a). To activate all actors in the institutional
publish ROI studies, case studies, manuals, and future field, they forged social networks to obtain sponsorship
use cases (Beck, 2018; Studio98Test, 2009; University from vendors, involvement from suppliers, and funding
of Arkansas, 2009; GS1US, 2018). With support from for research institutes (Swedberg, 2010). Through
other actors, they lead standards development, research studies and industry groups, they created
upgradation, and ratification (GS1, 2012). They knowledge (e.g., websites, social media, manuals,
supported adopters by providing education, training, recommendations, and case studies) for the would-be
tools, and community support (Edwards, 2012). coadopters (Swedberg, 2011c). Retailers regularly
Besides, they invested resources by opening new announced RFID roadmaps and successes using
dedicated RFID research centers (Studio98Test, 2012). compelling and future-oriented frames such as “a very
Finally, they performed mergers with other institutes big part of our future” that nudged other actors to
to better collaborate on standards development identify the commercial opportunities in RFID (Prince,
(Edwards, 2012). 2013). Some of the articles headlined “RFID ready for a
The retailer’s actions in this phase indicate that they lift-off,” (Kay, 2010), “Why RFID is (finally) here to
enacted all three OV functions. When RFID adoption stay,” (Woodie, 2016), and “Macy’s: Inventory will be
came to a standstill, the pilot studies shifted the field’s 100% RFID-tagged by 2017” (McKevitt, 2016a).
attention from pallet-level to item-level RFID Speaking of changing the future of RFID, one vendor
(Bustillo, 2010; Kay, 2010). By doing so, retailers said, “I don’t see it [RFID] slowing down at this point.
provided a revised understanding of RFID’s Acceleration has been pretty consistent the last two
usefulness, which encouraged other retailers to test and years” (Woodie, 2016). These actions reflect an actor
adopt RFID and motivated suppliers to cooperate. enacting the OV function of mobilization.
Retailers kept the field enthusiastic throughout the Proposition 4: The IE for a codependent innovation will
chasm by visualizing and testing new use cases in use all three IE strategies and enact all three OV
different product segments, often using buzzwords functions during the chasm phase of the early-stage
such as “buy online pickup in store,” “ship from store” diffusion process for that codependent innovation.

409
The Early-Stage Diffusion of Codependent IT Innovations

Advancements in RFID technology and other fields adopter communities, they are external to the coadopter
made RFID more useful. RFID technology advanced, community of suppliers resisting adoption of RFID tags
matured, and became cheaper (Wailgum, 2010), due to high recurring costs and lower perceived benefits,
meaning that choosing a tag for a specific application as compared to retailers. Therefore, for the diffusion of
became easier (Cooduvalli, 2017). Big data analytics RFID-in-retailing, Walmart acted as an internal-external
afforded the number-crunching ability not available in influencer in the first three phases, and the other retailers
the early phases (Lopez, 2017). The rise of cellular joined in during the chasm. This leads to our fifth
tracking technologies rendered RFID’s privacy issues proposition:
trivial in the minds of consumers (Woodie, 2016). ERP
Proposition 5: Instead of external influencers of a
systems evolved to ingest real-time data and are hence
codependent innovation playing the role of an IE,
suited for RFID (Wollenhaupt, 2018). The proliferation
the internal-external influencers who perceive the
of handheld devices and mobile apps meant that retailers
highest coadopter relative advantage of the
could employ handheld readers for their inventory count
codependent innovation will play the role of an IE
(Michel, 2015). Technologies like drones, Wi-Fi, and
in the early-stage diffusion process for that
the Internet of Things complemented RFID
codependent innovation.
(Baydakova, 2018). Adhesives became more resilient to
environmental conditions, improving read rates (Moore,
2019). With the popularity of social media, omnichannel 6 Discussion
sales—where customers buy online and pick up at
stores—emerged as a new RFID use case (Swedberg, 6.1 Key Insights about the Processual
2019). Dynamics
The joint adoption of RFID increased between 2010, Our findings show that the organization in the
when retailers introduced item-level tracking, and 2019 coadopter community that perceives the highest degree
(Hardgrave, 2016a). A rising number of retailers moved of coadopter relative advantage between the two
from pilots to deployment (Hardgrave, 2016a) during different coadopter communities influences the early-
this time, with the retail sector driving the majority of stage diffusion of a codependent innovation. In our
sales volume for the vendors (Wollenhaupt, 2018). The RFID casework, retailers perceived the highest degree
apparel segment within the retail industry continues to of coadopter relative advantage throughout the early
dominate RFID adoption, with other segments gradually stage and influenced the diffusion of the RFID-in-
catching up. The forecasts for RFID adoption by retailing innovation. This notion is similar to the notion
academic scholars and industry pundits paint a bright of relative advantage in prior research on single-whole
future (Kriz, 2018; Thau, 2017). There is consensus that innovation diffusion, which posits relative advantage
RFID fell into the “chasm” after Walmart’s mandate as the key predictor of adoption decisions
(Kriz, 2018; Roberti, 2017; Thau, 2017). Experts opine (Akhlaghpour & Lapointe, 2018; Rogers, 2003). In
that RFID is at the tipping point and ready for prime line with this literature, retailers perceive relative
time, with mainstream adoption on the horizon (Kriz, advantage from RFID adoption due to orders of
2018; Roberti, 2017; Wollenhaupt, 2018). Other experts magnitude improvements in inventory management,
feel that RFID adoption has now reached the early product tracking, and reducing operational costs.
majority (Swedberg, 2019). The discourse around RFID Suppliers perceive relative advantage in RFID due to
has moved from technology issues to business cases better shipment tracking and theft prevention.
(Warry, 2015). However, the notion of relative advantage perceived
by each adopter community is limited in explaining
Our overall case analysis discussed above shows that the
diffusion outcomes in codependent innovations
key influencers in the early stages of the codependent
because two different adopter communities are
innovation diffusion need not be an actor external to the
adopting two different parts of the innovation with
adopter community, as the traditional innovation
different relative advantages, and members from both
diffusion literature suggests (Bass, 2004; Jeyaraj &
communities need to adopt their respective parts for
Sabherwal, 2014; Strang & Soule, 1998). Walmart, an
their respective relative advantages to bear fruit. In the
actor internal to the overall adoption phenomenon but
case we examined, suppliers voiced concerns over the
external to the coadopter community of suppliers,
recurring costs of attaching RFID tags, despite their
perceived the highest degree of coadopter relative
enthusiasm for future benefits. But retailers perceived
advantage in RFID and played the IE role throughout the
a higher coadopter relative advantage because of the
early-stage diffusion process. During the chasm, the
fixed costs of installing RFID readers and the
other retailers played the role of an IE as well. Although
supporting infrastructure, while suppliers bore the high
Walmart and the other retailers are internal to the overall
and ongoing RFID tagging costs. The coadopter
adoption phenomenon, which involves two different
relative advantage was thus higher for retailers than for

410
Journal of the Association for Information Systems

suppliers beginning with the early stage, specifically Based on our analysis of the early-stage diffusion of
for Walmart, who factored into its calculus of RFID in retailing, we show how the diffusion process
coadopter relative advantage its dominance over unfolds during the early stages of a codependent
suppliers and its prior experience in supply-chain innovation. We divided the early stage into four
revolutions. Therefore, it adopted RFID readers. phases: emergence, structuralization, evolution, and
However, retailers cannot benefit from RFID without chasm. In each phase, we identified the actor that
joint adoption of the RFID technology by suppliers. played the role of an IE, the institutional
Thus, Walmart attempted to influence the suppliers entrepreneurship strategies they used, and the OV
who resisted RFID due to a lower coadopter relative functions they enacted.
advantage for them from this technology. Therefore,
6.1.1 Phase 1: Emergence
we add the notion of coadopter relative advantage to
IDT to answer who plays the role of influencers during In the emergence phase, which lasted from 1999 to
the early-stage diffusion process for codependent 2003, Walmart, as the internal-external influencer,
innovations. performed IE actions by developing and testing the
RFID technology. Our findings contrast with prior IDT
Our findings reveal the role of an internal-external literature that posits that during this phase, external
influencer in the early-stage diffusion of codependent influencers (Rogers, 2003), such as research analysts
innovations. In our RFID case, Walmart played this and their firms, play the role of IE by seeking
role in the early-stage diffusion of the RFID-in- opportunities for and developing IT innovations
retailing innovation. In contrast, prior research on (Greenwood et al., 2002; Liao, 2016; Morrill, 2006;
single-whole innovations suggests that external actors Wang & Ramiller, 2009). They collaborate with
play the influencer role in the early stages of the innovative adopters to research, experiment, and test
diffusion process (Bass, 2004; Jeyaraj & Sabherwal, the IT innovation (Morrill, 2006; Nielsen et al., 2014;
2014; Strang & Soule, 1998). The notion of the Rogers, 2003). However, our case analysis shows that
internal-external influencer is similar to the notion of Walmart performed these actions by funding the Auto-
internal and external influencers in prior research on ID Center and conducting pilots with interested
single-whole innovation diffusion. Walmart is an suppliers, representing the IE strategy of rationale
internal influencer as a member of the adopter development. Moreover, Walmart’s actions reflected
community. Walmart influenced other retailers by their enactment of the OV function of interpretation.
sharing RFID benefits found by its pilots and success Further, they dominated the discourse by using frames
stories gained as an adopter. Walmart is also an such as “smart shelves” and by discussing the RFID’s
external influencer in terms of its influence on the emergence and applicability and its future in retailing,
coadopter community of suppliers who must adopt calling it the next-generation supply chain technology.
In contrast, prior work on OVT suggests that in this
RFID tags in order for the benefits of RFID adoption
phase, research institutes dominate the discourse to
to accrue to any party. Walmart acted as an external
present the new IT as superior to current practices,
influencer for suppliers by extolling the benefits of
thereby attracting community attention and shaping the
RFID, framing RFID as the retailing industry’s future benefits for potential adopters (Liao, 2016; Wang &
technology, and preparing resources (such as manuals Ramiller, 2009). At the end of this phase, diffusion was
and tagging guides). Walmart was able to disseminate restricted to pilot tests conducted by retailers (mainly
information to many suppliers due to the sheer number Walmart) in collaboration with suppliers.
of suppliers tied to them and their use of mass media
outlets and conferences. Despite its similarity to the 6.1.2 Phase 2: Structuralization
notion of influencer in the IDT on single-whole
innovations, Walmart’s role adds nuance to the In the structuralization phase, which lasted from 2003 to
influencer concept because it influenced the diffusion 2005, Walmart as the internal-external influencer
performed IE actions by mobilizing the retailing field
of RFID, a new class of innovation that we term
toward RFID diffusion. Our findings contrast with prior
codependent innovation. Walmart is an internal-
IDT literature that posits that in this phase, external
external influencer because it is internal to the overall
influencers (Rogers, 2003), such as technology vendors,
RFID adoption phenomenon (as the adopter of RFID perform the actions of an IE by producing and releasing
reader) but external in terms of its influence on the products (Wang & Ramiller, 2009). Vendors work with
coadopter community of suppliers adopting the RFID visionary adopters by deploying resources to realize the
tags. Therefore, we add the notion of internal-external innovation (Purdy & Gray, 2009) and learn the new
influencer to IDT to answer the question of who plays technology to solve business problems (Fayard et al.,
the role of influencer during the early-stage diffusion 2016). However, our case analysis shows that Walmart
process for codependent innovations. influenced RFID diffusion by deploying RFID systems in
their retail stores and distribution centers and mandating
suppliers to adopt RFID tags. To activate the market,

411
The Early-Stage Diffusion of Codependent IT Innovations

retailers influenced vendors to reduce prices, offer retailers performed IE actions by focusing on a
discounts, and display their products, indicating the IE business problem that best suited the strengths of
strategy of resource mobilization. Walmart’s actions RFID. Compared to their focus on pallet-level RFID in
reflected their enactment of the OV function of the initial phases, Walmart and the other retailers
mobilization. Further, they dominated the discourse by focused on item-level RFID in the apparel category,
issuing and summarizing their mandate and providing which conferred a coadopter relative advantage for
constant updates on their mandate and RFID deployment. retailers and suppliers. In this phase, they worked
In contrast, prior work on OVT suggests that in this phase, closely with suppliers for sharing costs and
vendors dominate the discourse through their marketing implementation, simultaneously transforming their
campaigns to shape the benefits for potential adopters own sales floors by installing RFID readers. Our
(Swanson & Ramiller, 1997; Wang & Ramiller, 2009). findings contrast with prior literature that posits that
At the end of this phase, the diffusion can be characterized vendors act as external influencers and perform the
by retailers aggressively adopting RFID systems and actions of an IE during the chasm to shape the relative
suppliers adopting only the minimum required RFID tags advantage for the pragmatic potential adopters (Moore,
to fulfill Walmart’s mandate. 2014). They deploy resources to develop a complete
product and set up support services, working closely
6.1.3 Phase 3: Evolution with the adopter and forging other tactical alliances
(Gombault et al., 2016; Marsan et al., 2020; Moore,
In the evolution phase, which lasted from 2005 to
2014). However, our case analysis shows that retailers
2008, Walmart as the internal-external influencer
performed such actions, essentially utilizing all three
performed IE actions by focusing on developing and
IE strategies. Retailers’ actions reflected their
conveying the ROI of RFID to other actors. Our
enactment of all three OV functions. They dominated
findings contrast with prior literature on IDT that
the discourse to pitch item-level RFID as a new
posits that in this phase, the adopters of an innovation,
direction, sketch their implementation plans, visualize
i.e., the internal influencers (Rogers, 2003), are
novel use cases, showcase the maturity of standards
expected to perform IE actions to get more adopters on
and technology, and share success stories. In contrast,
board in order to benefit from network externalities.
prior work on OVT suggests that in this phase, vendors
Suppliers’ actions focus on perfecting the innovation
dominate the public discourse to demonstrate ROI and
to reduce costs and leveraging it to gain an advantage
market leadership and position the innovation in a
over their competitors because the innovation has now
niche market segment (Moore, 2014). Overall, retailers
been validated (Fayard et al., 2016; Wang & Ramiller,
helped RFID in retailing cross the chasm, which is
2009). However, our case analysis shows that Walmart
critical to deeming the early-stage diffusion successful,
focused on adding legitimacy to RFID use in retailing
consequently moving the diffusion to the growth stage
by sharing data with suppliers and conducting pilots in
(i.e., the mainstream markets).
real scenarios and standards development in
collaboration with research institutes, indicating the IE
strategy of relationship development. Walmart’s
6.2 Theoretical Contributions
actions reflected their enactment of the OV function of Understanding codependent innovation diffusion is
legitimation. Further, they dominated the discourse by critical to tapping into the transformative potential of
framing RFID ROI, standards maturity, technology this class of innovations. A case in point is the RFID-
maturity, and backing the technology. In contrast, prior in-retailing innovation, which promised a supply-chain
work on OVT suggests that in this phase, suppliers revolution in retailing and involved massive
dominate the discourse to incorporate their success investments and process changes from all actors in
stories and justify a potential adopter’s investment in retailing. But its diffusion fell into a chasm and needed
innovative IT (Wang & Ramiller, 2009). Despite these almost a decade to enter the growth stages of diffusion
efforts, RFID diffusion came to a standstill due to a (Kriz, 2018; Roberti, 2017). There is a lack of
lack of coadopter relative advantage for suppliers. understanding of how codependent innovations
Other retailers did not follow an aggressive (Bucklin & Sengupta, 1993) diffuse in the IS
implementation, limiting RFID adoption to pilots. innovation diffusion literature, compared to the
Vendors did not influence because they had few extensive research on the diffusion of single-whole
suppliers and had to bear the costs of updating innovations (Angst et al., 2010; Wang & Swanson,
standards. Our results show evidence of what scholars 2007). To address this, we study codependent
call a chasm between the early and the growth stages innovations with two complementary parts and
of diffusion (Moore, 2014; Xu et al., 2017). examine how they diffuse during their early stage. Our
paper makes three distinct contributions to the IS
6.1.4 Phase 4: Chasm literature by expanding the scope of IDT to the
In the chasm, which lasted between 2008 and 2019, as codependent innovation context. Table 7 summarizes
the internal-external influencers, Walmart and other our key contributions.

412
Journal of the Association for Information Systems

Table 7. Contributions of Our Study

Concept Phase 1: Emergence Phase 2: Structuralization Phase 3: Evolution Phase 4: Chasm


(1999–2003) (2003–2005) (2005–2008) (2008-2019)

Who perceived the Retailers (specifically Retailers (specifically Retailers Walmart and other
coadopter relative Walmart): Walmart Walmart): Based on (specifically retailers:
advantage? What could lead supply pilots, Walmart further Walmart): Reap
were the key chain innovation in crystallized the benefits RFID benefits by Item-level tracking
benefits? retailing, build new perceived in the last leveraging RFID improves inventory
trade models, and find phase. They need not bear already deployed in accuracy, on-shelf
ways to lower the recurring cost of RFID stores and availability, and
operational costs. tagging. distribution centers. granular product
The more the visibility in higher
suppliers adopt, the orders of magnitude
more benefits are than manual counting
accrued. and barcoding.

Was the influencer Internal-external Internal-external Internal-external Internal-external


internal or influencer influencer influencer influencer
external?

Who will act as the Retailers (specifically Retailers (specifically Retailers Walmart and other
IE? Walmart) Walmart) (specifically retailers
Walmart)

Which Rationale Resource mobilization Relationship Rationale


institutional development development development,
entrepreneurship Resource
strategy will the IE mobilization, and
use? Relationship
development

Which OV Interpretation Mobilization Legitimation Interpretation,


function will be Mobilization, and
enacted by the IE? Legitimation

Extent of joint Joint adoption takes Joint adoption grows: Joint adoption goes Joint adoption crosses
adoption off: Retailers aggressively into a chasm: the chasm: A rising
deployed RFID systems in Retailers (including number of retailers
Adoption was stores and distribution Walmart) were moved from pilots to
restricted to pilot tests centers. Suppliers adopted cautious about deployment.
conducted by retailers only the minimum further installing Suppliers, convinced
in collaboration with required RFID tags to RFID systems. by the benefits, are
suppliers. fulfill Walmart’s mandate. Walmart installed working closely with
Some suppliers missed the fewer RFID systems retailers to fulfill
deadline. in its distribution tagging requirements.
centers and lowered Some suppliers even
its mandate tagged products for
expectations retailers yet to adopt
compared to what it RFID. The apparel
originally segment within the
envisioned. retail industry
Suppliers’ continues to dominate
complaints on the RFID adoption, with
lack of benefits and other segments soon
high cost of catching up. Pundits
implementation from academia and
grew exponentially, industry paint a bright
and they adopted the future.
minimum required
RFID tags to satisfy
the retailers.

413
The Early-Stage Diffusion of Codependent IT Innovations

First, we add the notion of coadopter relative where the focus is on different perceptions within a
advantage to the IDT literature and posit that the community resulting from the technology component,
organization in the coadopter community that i.e., comparing the new technology with the incumbent
perceives the highest degree of coadopter relative one (Moore & Benbasat, 1991).
advantage between the two different adopter
In addition, our conceptualization focuses on the relative
communities will emerge as an IE and will influence
advantage differences between the two coadopter
the early-stage diffusion of the codependent
communities. Further, we show that another adopter’s
innovation. By doing so, we provide an enhanced
adoption likelihood (or lack thereof) is factored into the
understanding of the notion of relative advantage in
calculation of relative advantage and determines
IDT from two perspectives.
whether an actor plays the role of an influencer.
(1) Prior work conceptualizes relative advantage by Moreover, we add a new understanding that a high
assuming that all members of the adopter community relative advantage by itself is insufficient to foster the
draw from the same set of costs and benefits (Junglas diffusion of codependent innovations. Our notion adds
et al., 2019). This conceptualization is insufficient to value to researchers studying the diffusion of
explain the diffusion of codependent innovations codependent innovations, as they may direct their efforts
because it cannot explain the asymmetry in relative to develop instruments for measuring the coadopter
advantage between the two different coadopter relative advantage for the different coadopters involved.
communities and how this asymmetry affects the Further, our work offers a novel construct that
diffusion of the codependent innovation in its early researchers can theorize as predicting codependent
stages. We offer a nuanced conceptualization of innovation diffusion. Scholars focusing on strategies
relative advantage that we term coadopter relative like mandates and incentives (as we did in our paper)
advantage. Each coadopter community has its own (Rogers, 2003), and interventions in the IS design,
coadopter relative advantage, which is calculated promotion, and surrounding processes need to focus on
based on the unique set of benefits and costs in this novel construct as a dependent variable when
adopting their respective part of the innovation; studying codependent innovation diffusion.
however, the realization of this advantage is contingent
Second, we add the notion of an internal-external
upon joint adoption of the entire codependent
influencer to the IDT literature and posit that the IE as
innovation. This also helps explain why codependent
an influencer may be an actor internal to the overall
innovations generally favor one of the two coadopter
adoption phenomenon, which involves two different
communities. While we challenge the adequacy of the
adopter communities, who is, however, external in
notion of relative advantage in IDT in the context
terms of their influence for the coadopter community
codependent innovations, we also add further nuance
that has a lower coadopter relative advantage and thus a
to this notion, given its limitations in understanding the
lower motivation to adopt their part of the codependent
costs and benefits perceived by the two coadopter
innovation. By doing so, we provide an enhanced
communities in codependent innovations (Agarwal &
understanding of the notion of influencers in the IDT
Prasad, 1997).
from two perspectives.
(2) The conceptualization of prior work is limited in
(1) the concepts of internal or external influencers are
explaining how diffusion unfolds in the light of a
inadequate in conceptualizing the form of influence
hindrance to joint adoption due to the difference in the
needed for the diffusion of codependent innovations.
relative advantage for the two different coadopter
External influencers might find it hard to persuade
communities. Our conceptualization of coadopter
adopters due to the asymmetric coadopter relative
relative advantage factors in the adoption likelihood of
advantage and the consequent stalemate in
the other coadopter community, and that calculation
codependent innovation diffusion. A member of the
determines who performs the role of an IE to influence
coadopter community with the higher degree of
diffusion. We propose that the organization in the
coadopter relative advantage might influence the
coadopter community with a higher coadopter relative
coadopter community with the lower degree of
advantage that perceives the highest degree of coadopter
coadopter relative advantage. While this member, as
relative advantage will influence the diffusion by acting
an adopter in the overall codependent innovation
as an IE because their higher coadopter relative
adoption phenomenon, is an internal influencer, this
advantage will bear fruit only if the other coadopter
member is also external in terms of their influence on
community also embraces and adopts their part of the
the coadopter community with the lower degree of
codependent innovation. Through this understanding,
coadopter relative advantage. To conceptualize this
we add to the prior work in IDT that posits relative
influence, our work proposes a new type of influencer
advantage as the key notion in explaining the diffusion
in terms of internal-external influencer and adds a new
of single-whole innovations (Akhlaghpour & Lapointe,
understanding to the influence processes in IDT
2018). Our conceptualization of coadopter relative
research on communication channels (Loh &
advantage aligns with prior work on relative advantage
Venkatraman, 1992; Strang & Soule, 1998).

414
Journal of the Association for Information Systems

(2) We add a new understanding that internal-external (2) We add a finer-grained understanding of the diffusion
influencers play the role of an IE in all the phases of the phenomenon by identifying the four unique phases
early-stage diffusion process for codependent during the early-stage diffusion process and discussing
innovations. In contrast, prior research on single-whole which specific institutional entrepreneurship strategies to
innovations suggests that external actors play the use and OV functions to enact to address the distinct
influencer role in the early stages of the diffusion issues that arise in each of those phases to foster the
process (Bass, 2004; Hu et al., 1997). Our notion adds diffusion of codependent innovations. Based on our
value for researchers studying the diffusion of RFID-in-retailing casework, we propose that the IE for a
codependent innovations, who might develop new codependent innovation will: (a) use the rationale
instruments to measure internal-external influence in the development strategy and enact the interpretation OV
context of codependent innovation diffusion. Further, function during the emergence phase, (b) use the
researchers can theorize the effect of diffusion resource mobilization strategy and enact the mobilization
interventions involving the internal-external influencer OV function during the structuralization phase, (c) use
to speed up the diffusion process because these the relationship development strategy and enact the
influencers are most likely to obtain buy-in from legitimation OV function during the evolution phase, and
potential adopters. Such interventions can be compared (d) use all the three institutional entrepreneurship
with those involving external influencers, whose strategies and enact all the three OV functions during the
influence might slow down the diffusion. Our work aids chasm phase. By doing so, our work extends prior work
researchers in theorizing the timing and the nature of on phase models in the IT diffusion literature that looks
influence at various phases in the early-stage diffusion at dynamic patterns in effective communication channels
so that communication channels are utilized in the ideal (Loh & Venkatraman, 1992; Strang & Soule, 1998) and
sequence as the diffusion of codependent innovations in an actor’s discursive (Wang & Ramiller, 2009) and
progresses from one phase to another. material practices in innovating with IT (Nielsen et al.,
2014). We extend this work to the context of
Third, we divide the early-stage codependent innovation
codependent IT innovations.
diffusion process into four phases—emergence,
structuralization, evolution, and chasm—and identify In addition, our work adds to the small but growing body
the institutional entrepreneurship strategies used and the of work on the strategies used by actors to enact the OV
OV functions enacted by the IE during each phase. This functions of interpretation (Liao, 2016), legitimation
contribution provides a more nuanced understanding of (Kaganer et al., 2010; Marsan et al., 2020), and
influencers’ actions in the IDT from two perspectives. mobilization (Amadoru et al., 2021; Wang & Swanson,
2007). Further, compared to prior work that primarily
(1) We add a new understanding that influencers need to
focuses on legitimation strategies (Kaganer et al., 2010;
engage in institution building actions to promote the Paré et al., 2020), our work also focuses on interpretation
diffusion of the new class of innovation termed and mobilization strategies (Amadoru et al., 2021). Our
codependent innovations. The diffusion of codependent work affords researchers additional mechanisms—i.e.,
innovations requires influencers to change the existing institutional entrepreneurship strategies and enacting OV
unsupportive institutional arrangements, which lowers functions—with which to theorize the role of influencers
the coadopter relative advantage and deters the key in fostering the diffusion of codependent innovations.
actors’ attention and resources from the new technology. Our four-phase process model serves as a baseline for
However, prior work posits that external and internal future researchers to test and expand the understanding
influencers motivate potential adopters by relying on the of the process of codependent innovation diffusion in
communication of generic and experiential information, different contexts.
respectively (Angst et al., 2010; Jeyaraj & Sabherwal,
Our theoretical contributions offer originality through
2014). We add to this line of work and posit that
several revelatory (i.e., surprising) insights into our
influencers foster diffusion by building new institutions understanding of the diffusion of codependent
by using institutional entrepreneurship strategies and by innovations, offer scientific utility through a clear
enacting OV functions. We also add to the growing presentation of the research model, present boundary
body of work that integrates OVT and IET to explain the conditions for further theory building and testing, and
diffusion process by extending it to the context of provide practical utility through the presentation of the
codependent IT innovations (Paré et al., 2020; Wang & managerial implications in a subsequent section (Corley
Swanson, 2007). In contrast to prior work focusing on & Gioia, 2011). We made our best efforts to adhere to
the strategies used by vendors and other external the key considerations of theory development outlined
influencers (Kim & Miranda, 2018; Paré et al., 2020; by Rivard (2014) by identifying a gap, defining and
Wang & Swanson, 2007), we focus on the strategies extending the key theoretical concepts driving our
used by adopters, specifically internal-external phenomenon, explicating our contributions, and
influencers, who enact different OV functions to spur improving the presentation clarity by using various tools
the innovation process. like figures, tables, and propositions. In terms of

415
The Early-Stage Diffusion of Codependent IT Innovations

theoretical integration, our work falls into the type “close whether they need to carry out the role of a “visionary”
areas of research, compatible underlying assumptions” in organization championing an innovation. Second,
the typology proposed by Okhuysen and Bonardi (2011). visionary organizations can engage in three types of
All three theories—IDT, OVT, and IET—represent strategies: (1) developing the rationale for the
close research areas because they examine a common innovation by means of press statements, online news
phenomenon—the diffusion of new practices. Further, releases, white papers, conferences, promotional
these three theories also have compatible underlying activities, buzzwords, and keynotes; (2) mobilizing
assumptions, as they all assume multiple actors with resources by using penalties, financial clout,
motives and self-interests and clarify that an innovation investments in new equipment, research funding,
might not benefit everyone. political resources for lobbying, and reward systems
for supporters of the innovation; and (3) developing
We believe that our theoretical insights are generalizable
relations through coalitions, alliances, research
(see Appendix A for more details) within the boundary collaborations, think tanks, standard-setting bodies,
of codependent innovations in retailing settings and/or governmental organizations. Third, our notion
involving big players. In terms of the Lee and Baskerville that an IE can enact certain OV functions informs
(2003) generalizability framework, ET generalizability organizations that their actions have industry-level
(from empirical statements to theoretical statements) diffusion implications. In other words, their actions
applies to our study. To further strengthen the face can, directly and indirectly, activate the industry to
validity of our findings, we show our model’s perform interpretation, mobilization, and legitimation
applicability in the case of another codependent functions and thereby aid IT diffusion. Organizations
innovation in retail. We analyzed the early-stage interested in leading the community toward
diffusion of EMV technology in US retailing. EMV interpretation should engage in research and
chip-based credit\debit cards and EMV chip readers are development to understand innovations, fund research
the two complementary parts. The two different adopter institutes, and use public discourse to come up with
communities who adopt this innovation are banks who buzzwords. Organizations interested in leading the
issue EMV cards to their customers and merchants and community toward mobilization need to help vendors
retailers who install chip readers at point-of-sale manufacture IT solutions, urge them to display new
terminals, respectively. We found that the early-stage products in forums, mandate the adoption of other
diffusion of this technology followed the process actors, call others to deploy resources to implement the
model we discerned in our study of RFID diffusion in innovation, and use discourse to outline
retailing. In the early stages of EMV card diffusion, implementation plans. Finally, organizations
banks perceived a higher degree of coadopter relative interested in leading the community toward
legitimation need to endorse innovations, help set
advantage than retailers and played the role of an IE.
standards, manage the discourse, legitimize the
Merchants in the US resisted EMV cards mainly due
innovation, and promote success stories. Fourth, our
to costs associated with implementation. As a
four-phase model for early-stage diffusion of a
countermeasure, banks announced a deadline beyond codependent innovation shows organizations how to
which EMV noncompliant merchants would be held time their strategies when innovating with
liable for fraud committed by consumers at the point codependent innovations. In the emergence phase, the
of sale—a move very similar to the mandates issued by IE needs to work closely with research institutes by
Walmart during the structuralization phase (Pettey, funding them and drafting standards. In the
2015). Banks utilized public discourse to position structuralization phase, the IE needs to influence the
EMV technology as a solution for security breaches coadopter community and vendors simultaneously and
and to outline EMV transition deadlines (Visa, 2019; use various techniques, including mandates, to help
Orem, 2019). These efforts led to the joint adoption of achieve economies of scale. In the evolution phase, the
EMV cards by banks and chip readers by retailers IE should ensure that adopters get a good ROI, vendors
(Orem, 2019). save on costs, and research institutes develop new
standards to improve constituent technologies. In the
6.3 Practical Implications chasm phase, the IE needs to find a niche target
segment for the innovation, work closely with the
Our theoretical contributions have important practical coadopter community for sharing costs and
implications for organizations’ IT innovation implementation, and collaborate with vendors and
practices. First, organizations engaging in the adoption research institutes for technology maturation and
and diffusion of a codependent IT innovation should standards development. Finally, the IE interested in
constantly perform internal analyses of the costs and codependent innovation diffusion needs to perform
benefits of the innovation over existing actions to enact the institutional functions of
technology/practices and the external analysis of costs interpretation, mobilization, and legitimation during
and benefits to the members of the other coadopter the emergence, structuralization, and evolution phases,
community. This analysis will inform the organization

416
Journal of the Association for Information Systems

respectively. In the chasm phase, the IE needs to Second, future studies could also extend our
perform actions to enact all three OV functions. framework to include the later stages of codependent
innovation diffusion and study who plays the
6.4 Limitations and Future Work influencer role in later stages. Future studies might also
delve deeper into the mechanisms of internal-external
Our narrative and interpretation of RFID diffusion are influence. For instance, IDT posits that prestige and
based on our analysis of internet-sourced discourse. competition influence adopters to imitate other
While internet sources can play an essential and key adopters (Jeyaraj & Sabherwal, 2014; Strang & Soule,
part in the discourse surrounding an innovation, they 1998). Given that an internal-external influencer might
could be restricted in the extent to which they represent not be an actor from the same adopter community, it
the diversity of opinions by each actor (Wang & would be interesting to study how such factors
Ramiller, 2009). However, internet-based resources influence imitation.
are the biggest discursive arena and have also been
used in prior RFID (Fish, 2006; Quaadgras, 2005), Third, future research could also focus on the structure
innovation diffusion (Davidson et al., 2015), and of the OV. Business use cases are one of the key
institutional entrepreneurship (Garud et al., 2002) components of an OV (Miranda et al., 2015). We
studies. We collected data from diverse, reputed expect the structure of an OV for codependent
sources to ensure the credibility of our findings. Future innovations to be different because two different
research could focus on other sources like industry adopter communities are involved, and the overall
expositions, social media sources, and personal innovation tends to favor one coadopter community. In
interviews to gain an insider perspective, and use the case of the RFID-in-retailing diffusion, we noticed
analytical tools like text mining and linguistic analysis that the business use cases focused on retailers fueled
to gain additional insights (Amadoru et al., 2021; by their interests. Studying the structure of OV in
Davidson et al., 2015; Miranda et al., 2015). codependent innovations could yield a nuanced
structure of OV (Miranda et al., 2015). As another
Also, while we studied the actions of all actors in the fruitful future research avenue, codependent
retailing field, we analyzed OV around codependent innovations may add further nuance to the notion of
innovations primarily from a Walmart-centered community reception of an OV (Davidson et al., 2015;
perspective. As a potential extension to OV research in Ramiller & Swanson, 2003; Swanson & Ramiller,
codependent innovations, scholars could explore the 2004). The business use cases, legitimating claims, and
OV shaping activities of other adopting parties and the implementation guidelines and challenges are centered
tensions in the conversations surrounding coadopter around one coadopter community and thus cannot be
relative advantage. Further, although suppliers in our readily received by another coadopter community.
case study initially complied with the mandate, they Although the business use cases are centered on one
eventually abandoned it because they would have been coadopter community, studying the profile and
no worse off had they followed the mandate and experience of actors in the other coadopter community
adopted RFID (Hardy & Maguire, 2017). Moreover, could explain those actors’ reception of the OV
retailers were unable to develop relationships in the (Marsan et al., 2012). A related concept in IET is an
evolution phase to the extent needed to collectively actor’s social position in the institutional field
fulfill the mandates (Hardy & Maguire, 2017). Future (Battilana et al., 2009), which confers upon them
studies could extend IET with the notion of mandates financial, political, or social resources and grants them
by an IE to bring about institutional change and the power to influence the field (Hardy & Maguire,
research the extent to which they can be effective. 2017). Future research could explore the dynamic
We propose three additional directions for future patterns in an actors’ profile, experience (Marsan et al.,
research. First, future research can explore how 2012), and social position in the institutional field at
differential levels of other important innovation different phases of the diffusion of codependent
attributes, like compatibility, complexity, trialability, innovations (Battilana et al., 2009).
and observability, play a role in diffusing codependent
innovations (Rogers, 2003; Sun et al., 2018). Experts Acknowledgments
have opined that one of the barriers to RFID diffusion
is the complexity inherent in RFID deployment—a We are grateful to the senior editor and the three
process involving the integration of complicated IT anonymous reviewers for a constructive and
infrastructure and business process changes for which developmental review process, which substantially
there is no cohesive solution available in the market improved our manuscript. We are also thankful to Dr.
(Kaplan, 2018b; Ricci, 2010). Studying the role of Sutirtha Chatterjee (University of Nevada, Las Vegas)
relative complexity between two different adopter for his pointers during the revision process on how to
communities and how it impacts their adoption more clearly articulate the contribution that this
decisions is an exciting avenue for future research. qualitative work makes to the IS field.

417
The Early-Stage Diffusion of Codependent IT Innovations

References universal product code. Journal of Political


Economy, 129(4), 1252-1286.
Abrahamson, E., & Fairchild, G. (2001). Knowledge
industries and idea entrepreneurs: New Bass, F. M. (2004). Comments on “A new product
dimensions of innovative products, services, growth for model consumer durables: The bass
and organizations. Stanford University Press. model.” Management Science, 50(12_supple.),
1833-1840.
Agarwal, R., & Prasad, J. (1997). The role of
innovation characteristics and perceived Battilana, J., Leca, B., & Boxenbaum, E. (2009). How
voluntariness in the acceptance of information actors change institutions: Towards a theory of
technologies. Decision Sciences, 28(3), 557- institutional entrepreneurship. The Academy of
582. Management Annals, 3(1), 65-107.

Akhlaghpour, S., & Lapointe, L. (2018). From placebo Baydakova, A. (2018). Walmart explores blockchain
to panacea: Studying the diffusion of IT for connecting automated delivery drones.
management techniques with ambiguous Coindesk. https://www.coindesk.com/walmart-
efficiencies: The case of capability maturity explores-blockchain-for-connecting-
model. Journal of the Association for automated-delivery-drones
Information Systems, 19(6), 441-502. Beck, A. (2018). Measuring the impact of RFID in
Amadoru, M., Fielt, E., & Kowalkiewicz, M. (2021). retailing: Key lessons from 10 Case-study
Organizing visions in the digital world: The companies. https://www.gs1uk.org/sites/default/
case of the blockchain discourse on Twitter. files/inline-files/gs1_uk_the_impact
Proceedings of the Forty-Second International _of_rfid_report.pdf
Conference on Information Systems. Becker, P. (2004). Wal-Mart converse suppliers RFID
Anderson, C. (2002). Everything you always wanted to summit in Bentonville. Digital ID World, 2(1),
know about RFID, but were afraid to ask. 62-65.
Logistics Management, 43(9), 57-60. Benbasat, I., Goldstein, D. K., & Mead, M. (1987). The
Anderson, G. (2016). Macy’s to RFID tag everything. case research strategy in studies of information
RetailWire. Retrieved 11-7-2011 from systems. MIS Quarterly, 11(3), 369-386.
https://www.retailwire.com/discussion/macys- Brancheau, J. C., & Wetherbe, J. C. (1990). The
to-rfid-tag-everything/ adoption of spreadsheet software: testing
Angst, C. M., Agarwal, R., Sambamurthy, V., & innovation diffusion theory in the context of
Kelley, K. (2010). Social contagion and end-user computing. Information Systems
information technology diffusion: The adoption Research, 1(2), 115-143.
of electronic medical records in US hospitals. Bucklin, L. P., & Sengupta, S. (1993). The co-
Management Science, 56(8), 1219-1241. diffusion of complementary innovations:
Anonymous. (2004). RFID: Powering the Supply Supermarket scanners and UPC symbols.
Chain. Logistics Management, 43(8), R3-R16. Journal of Product Innovation Management,
10(2), 148-160.
Ashton, K. (2003). Ready for the auto-ID revolution:
an interview with Kevin Ashton. Supply Chain Bustillo, M. (2010). Wal-Mart radio tags to track
Management Review, 7(3), 44-52 clothing. The Wall Street Journal. https://www.
wsj.com/articles/SB100014240527487044213
Bacheldor, B. (2004, 01/05). Ready for RFID? 04575383213061198090
Information Week, 970, 20-22.
Capgemini. (2004). CGE&Y and Sun Microsystems
Bacheldor, B. (2007, 04/16). EPCglobal Ratifies help top suppliers meet Wal-Mart RFID
EPCIS Standard. RFID Journal. http://www. compliance.. http://www.ca.capgemini.com/
rfidjournal.com/article/articleprint/3237/-1/1 news/current_news.asp?ID=347&PRyear=2003
Bartis, E., & Mitev, N. (2008). A multiple narrative Chandler, S. (2014). The dysfunctional state of
approach to information systems failure: a America’s credit cards. CNBC. https://www.
successful system that failed. European cnbc.com/2014/01/13/the-dysfunctional-state-
Journal of Information Systems, 17(2), 112- of-americas-credit-cards.html
124.
Clarysse, B., Wright, M., Bruneel, J., & Mahajan, A.
Basker, E., & Simcoe, T. (2021). Upstream, (2014). Creating value in ecosystems: Crossing
downstream: Diffusion and impacts of the the chasm between knowledge and business

418
Journal of the Association for Information Systems

ecosystems. Research Policy, 43(7), 1164- Duvall, M. (2007a, 10/04). RFID price tags stay high,
1176. holding back adoption. BaseLine. http://www.
baselinemag.com/article2/0,1397,2192217,00.asp
Collins, J. (2003, 09/15). IBM Introduces New RFID
Services. RFID Journal. http://www.r Duvall, M. (2007b, 10/03). Wal-Mart’s Faltering
fidjournal.com/article/view/583 RFID Initiative. BaseLine. http://www.
baselinemag.com/print_article2/0,1217,a=216
Cooduvalli, U. (2017). Top considerations for RFID
467,00.asp
tag selection in your retail project.
RISnews.com. https://risnews.com/top- Edwards, T. (2012). GS1 US and voluntary
considerations-rfid-tag-selection-your-retail- interindustry commerce solutions association
project to merge. GS1 US. https://www.gs1us.org/
details-page/articleid/113/gs1-us-and-
Cook, J. (2014). Impinj bounces back as RFID heats
voluntary-interindustry-commerce-solutions-
up, expands in old Seattle grocery store. Geek
association-to-merge
Wire. https://www.geekwire.com/2014/impinj-
bounces-back-rfid-market-heats-expands-old- EMV. (2019, August 3). In Wikipedia
seattle-grocery-store/ https://en.wikipedia.org/w/index.php?title=
EMV&oldid=909110808
Corley, K. G., & Gioia, D. A. (2011). Building theory
about theory building: what constitutes a EPCGlobalInc. (2003). Specifications.
theoretical contribution? Academy of http://www.epcglobalinc.org/standards/specs/
Management Review, 36(1), 12-32.
Fabris, N. (2007, 06/17). End-user survey confirms
Cosgrove, E. (2019). Nike reaps benefits of short lead that RFID is moving beyond supply chain
times, RFID. Supplychaindive.com. management. ABI Research. http://www.abi
https://www.supplychaindive.com/news/Nike- research.com/abiprdisplay.jsp?pressid=886
lead-times-speed/551227/
Fayard, A.-L., Gkeredakis, E., & Levina, N. (2016).
David, J. E., & Best, I. (2014). Target: Stolen Framing innovation opportunities while staying
information involved at least 70 million people. committed to an organizational epistemic
CNBC. stance. Information Systems Research, 27(2),
https://www.cnbc.com/2014/01/10/target- 302-323.
stolen-information-involved-up-to-70-million-
Feder, B. J. (2004, December 27). Despite Wal-Mart’s
people.html
edict, radio tags will take time. New York
Davidson, E. J., Østerlund, C. S., & Flaherty, M. G. Times. http://www.nytimes.com/2004/12/27/
(2015). Drift and shift in the organizing vision technology/27rfid.html?ex=1261803600&en=
career for personal health records: An dc8bfdf3fd86e222&ei=5090&partner=geartest
investigation of innovation discourse dynamics.
FierceRetail. (2009). P&G’s decision to pull back from
Information and Organization, 25(4), 191-221.
Wal-Mart RFID trial quite understandable.
DiMaggio, P. J. (1988). Interest and agency in https://www.fierceretail.com/operations/p-g-s-
institutional theory. In L. G. Zucker (Ed.), decision-to-pull-back-from-wal-mart-rfid-trial-
Institutional patterns and organizations: quite-understandable
culture and environment (pp. 3-21). Ballinger.
Fish, L. A. (2006). The state of RFID implementation
Dimyan, L. (2016). No chips: A slow go for new credit 2006 for retail supply chains (WP 2006-02).
card technology. CNBC. http://www.cnbc. Canisius College. http://www.canisius.edu/
com/2016/03/04/no-chips-a-slow-go-for-new- images/userImages/wsbweb/Page_9979/WP20
credit-card-technology.html 06-02.doc
Dodge, M., & Kitchin, R. (2014). Barcodes and Gardner, W. D. (2004, December 21). Analyst: Wal-
RFIDs. Oxford University Press. Mart's RFID suppliers are resisting.
InformationWeek. http://www.information
Douglas, M. (2002). Playing tag with pallets and
week.com/shared/printableArticleSrc.jhtml?art
containers. Inbound Logistics. http://www.
icleID=558
inboundlogistics.com/cms/article/playing-tag-
with-pallets-and-containers/ Garfinkel, S., & Rosenberg, B. (2006). RFID:
Applications, security, and privacy. Pearson
Dutta, A., Lee, H. L., & Whang, S. (2007). RFID and
Education India.
operations management: Technology, value,
and incentives. Production and Operations Garud, R., Jain, S., & Kumaraswamy, A. (2002).
Management, 16(5), 646-655. Institutional entrepreneurship in the

419
The Early-Stage Diffusion of Codependent IT Innovations

sponsorship of common technological www.gs1us.org/industries/apparel-general-


standards: The case of Sun Microsystems and merchandise/standards-in-use/project-zipper
Java. Academy of Management Journal, 45(1),
Guo, Z., & Chen, J. (2018). Multigeneration product
196-214.
diffusion in the presence of strategic
Gaukler, G. M., Seifert, R. W., & Hausman, W. H. consumers. Information Systems Research,
(2007). Item‐level RFID in the retail supply 29(1), 206-224.
chain. Production and Operations
Haberman, A. L. (1999). Twenty-five years behind
Management, 16(1), 65-76.
bars: The proceedings of the twenty-fifth
Gilbert, A. (2003). Major retailers to test “smart anniversary of the U.P.C. Harvard University
shelves”. CNET. http://www.cnet.com/news/ Press.
major-retailers-to-test-smart-shelves/
Hardgrave, B. (2016a). Greater RFID awareness and
Gillo, N. (2016). Target’s RFID and fulfillment acceptance lead to accelerated adoption. R
investments lead to massive increase in digital RFID Journal. https://www.rfidjournal.com/
sales. RISnews.com. https://risnews.com/ articles/view?15040
targets-rfid-and-fulfillment-investments-lead-
Hardgrave, B. (2016b). Retailers must reframe their
massive-increase-digital-sales
thinking. RFID Journal. https://www.
Goldenberg, J., & Oreg, S. (2007). Laggards in rfidjournal.com/retailers-must-reframe-their-
disguise: Resistance to adopt and the thinking-2
leapfrogging effect. Technological Forecasting
Hardy, C., & Maguire, S. (2017). Institutional
and Social Change, 74(8), 1272-1281.
entrepreneurship and change in fields. In R.
Gombault, A., Allal-Chérif, O., & Décamps, A. Greenwood, C. Oliver, T. B. Lawrence, R. E.
(2016). ICT adoption in heritage organizations: Meyer (Eds.), The SAGE handbook of
Crossing the chasm. Journal of Business organizational institutionalism (pp. 261-280).
Research, 69(11), 5135-5140. SAGE.
Greengard, S. (2015). RFID adoption heats up. RFID Harford, T. (2017). How the barcode changed retailing
Journal. https://www.rfidjournal.com/rfid- and manufacturing. BBC News. https://www.
adoption-heats-up bbc.com/news/business-38498700
Greenwood, R., Suddaby, R., & Hinings, C. (2002). Hill, K. (2004, December 22). Analyst: Wal-Mart
Theorizing change: The role of professional RFID pilot closely watched. NewsFactor
associations in the transformation of http://www.crm-daily.com/story.xhtml?story_
institutionalized fields. The Academy of id=29218
Management Journal, 45(1), 58-80.
Hinings, B., Gegenhuber, T., & Greenwood, R. (2018).
GS1. (2009a). GS1 EPCglobal RFID-based Electronic Digital innovation and transformation: An
Article Surveillance (EAS) guides released by institutional perspective. Information and
GS1. https://www.gs1.org/sites/default/files/ Organization, 28(1), 52-61.
docs/media_centre/gs1_pr_131009_epc_rfid_b
Hu, Q., Saunders, C., & Gebelt, M. (1997). Diffusion
ased_eas.pdf
of information systems outsourcing: A
GS1. (2009b). Microsoft BizTalk RFID successfully reevaluation of influence sources. Information
passes the GS1 EPCglobal Certification Systems Research, 8(3), 288-301.
Program for compliance with the EPCIS
Huang, C.-Y. (2011). Rethinking leapfrogging in the end-
Standard. https://www.gs1.org/sites/default/
user telecom market. Technological Forecasting
files/docs/media_centre/gs1_pr_101109_EPCI
and Social Change, 78(4), 703-712.
S_Microsoft_BizTalk.pdf
Hudson, K. (2006, December 29). Wal-Mart blames
GS1. (2012). MIT and Johnson & Johnson
short-term woes. Wall Street Journal.
representatives elected to lead GS1 EPCglobal
http://chinese.wsj.com/gb/20061228/hrd17285
board. https://www.gs1.org/docs/media
8.asp?source=NewSearch
_centre/gs1_pr_080312_epc_global_board.pdf
Hyde, R. (2019). How Walmart model wins with
GS1. (2013). New version of EPC air interface
“everyday low prices.” Investopedia.
standard “Gen2v2” is ratified. https://www.
https://www.investopedia.com/articles/persona
gs1.org/docs/media_centre/gs1_pr_071113.pdf
l-finance/011815/how-walmart-model-wins-
GS1US. (2018). EPC/RFID data exchange study everydaylowprices.asp
project zipper executive summary. https://

420
Journal of the Association for Information Systems

IDTechEx. (2005a). The RFID knowledge base: Kay, M. (2010). Ready for liftoff: RFID in the apparel
Sample case studies. https://www.idte industry. RIS News. https://risnews.com/ready-
chex.com/research/online-market-intelligence- liftoff-rfid-apparel-industry
services-rfid-knowledgebase.asp
Kevin, F. (2005, Dec 12, 2005). The RFID revolution.
IDTechEx. (2005b). RFID progress at Wal-Mart. Telephony, 246(24), 28-30.
http://www.idtechex.com/products/en/articles/
Kim, I., & Miranda, S. M. (2018). 20 years old but still
00000161.asp
a teenager? A review of organizing vision
Impinj. (2000). About us. http://www.impinj.com/ theory and suggested directions. Proceedings of
about/default.aspx the Pacific Asia Conference on Information
Systems.
Incucomm. (2005). Wal-Mart’s RFID deployment -
how is it going http://www.incucomm.com/ KimberlyClark. (2006). Kimberly-Clark begins
releases/Wal-Mart%20Jan%202005%20Status shipping Gen2 RFID-tagged cases of products
%20- %20Executive%20Summary.PDF to Wal-Mart. http://investor.kimberly-
clark.com/ReleaseDetail.cfm?ReleaseID=1930
InformationWeek. (2001). RFID chips put to the test.
99
https://www.informationweek.com/rfid-chips-
put-to-the-test/d/d-id/1010978 Klein, H. K., & Myers, M. D. (1999, 03). A set of
principles for conducting and evaluating
Jeyaraj, A., & Sabherwal, R. (2014). The bass model
interpretive field studies in information
of diffusion: Recommendations for use in
systems. MIS Quarterly, 23(1), 67-93.
information systems research and practice.
Journal of Information Technology Theory and Kossman, S. (2016). The big chip card switch: Living
Application, 15(1), 5-30. with EMV. Credictcards.com. https://www.
creditcards.com/credit-card-news/emv-chip-
Johnson, J. R. (2005). The Promise of RFID: all that
cards-one-year-later-consumer.php
glitters... DC Velocity, (9), http://www.
dcvelocity.com/print/?article_id=489 Kossman, S. (2017). 8 FAQs about EMV credit cards.
Credictcards.com. https://www.creditcards.
Jones, M. A., Wyld, D. C., & Totten, J. W. (2005). The
com/credit-card-news/emv-faq-chip-cards-
adoption of RFID technology in the retail
answers-1264.php
supply chain. The Coastal Business Journal,
4(1), 29-42. Kriz, E. (2018). Did 2017 mark a tipping point for
RFID adoption by retailers? Freedonia Group.
Junglas, I., Goel, L., Ives, B., & Harris, J. (2019).
https://www.freedoniagroup.com/Content/Blo
Innovation at work: The relative advantage of
g/2018/01/11/Did-2017-Mark-a-Tipping-
using consumer IT in the workplace.
Point-for-RFID-Adoption-by-Retailers
Information Systems Journal, 29(2), 317-339.
Lai, K.-h., Wong, C. W. Y., & Cheng, T. C. E. (2006).
Kaganer, E. A., Pawlowski, S. D., & Wiley-Patton, S.
Institutional isomorphism and the adoption of
(2010). Building legitimacy for IT innovations:
information technology for supply chain
the case of computerized physician order entry
management. Computers in Industry, 57(1), 93-98.
systems. Journal of the Association for
Information Systems, 11(1), 1-33. Langley, A. (1999). Strategies for theorizing from
process data. Academy of Management Review,
Kaplan, D. A. (2018a). The rise, fall and return of
24(4), 691-710.
RFID. Supply Chain Dive. https://www.
supplychaindive.com/news/RFID-rise-fall- Leca, B., Battilana, J., & Boxenbaum, E. (2008).
and-return-retail/530608/ Agency and institutions: A review of
institutional entrepreneurship (Working
Kaplan, D. A. (2018b). The rise, fall and return of
paper). HBS Working Paper Series).
RFID. Supply Chain Dive. https://www.supply
https://www.hbs.edu/faculty/Pages/item.aspx?
chaindive.com/news/RFID-rise-fall-and-
num=32451
return-retail/530608/
Lee, A., & Baskerville, R. (2003). Generalizing
Katz, M. L., & Shapiro, C. (1985). Network
generalizability in information systems
externalities, competition, and compatibility.
research. Information Systems Research, 14(3),
American Economic Review, 75(3), 424-440.
221-243.
Katz, M. L., & Shapiro, C. (1986). Technology
Lee, A., & Baskerville, R. (2012). Conceptualizing
adoption in the presence of network
generalizability: new contributions and a reply.
externalities. Journal of Political Economy,
MIS Quarterly, 36(3), 749-761.
94(4), 822-841.

421
The Early-Stage Diffusion of Codependent IT Innovations

Lee, Y. M., Cheng, F., & Leung, Y. T. (2004). https://losspreventionmedia.com/the-roi-for-


Exploring the impact of RFID on supply chain rfid-technology-in-retail/
dynamics. Proceedings of the 2004 Winter
McCathie, L. (2004). The advantages and
Simulation Conference.
disadvantages of barcodes and radio frequency
Liao, T. (2016). Is it “augmented reality”? Contesting identification in supply chain management
boundary work over the definitions and [University of Wollongong Faculty of
organizing visions for an emerging technology Engineering & Information Sciences Honours
across field-configuring events. Information Thesis]. https://ro.uow.edu.au/thesesinfo/9/
and Organization, 26(3), 45-62.
McKevitt, J. (2016a). Macy’s: Inventory will be 100%
Libai, B., Mahajan, V., & Muller, E. (2008). Can you RFID-tagged by 2017. Supplychaindive.com.
see the chasm? Innovation diffusion according https://www.supplychaindive.com/news/Macy
to Rogers, Bass, and Moore. In N. K. Malhotra s-RFID-inventory-tracking/428937/
(Ed.), Review of Marketing Research (Vol. 5,
McKevitt, J. (2016b). Target links RFID-enabled
pp. 38-57). Emerald Publishing Limited.
inventory accuracy with sales growth.
Loh, L., & Venkatraman, N. (1992, 12). Diffusion of Supplychaindive.com. https://www.supply
information technology outsourcing: Influence chaindive.com/news/target-rfid-invetory-
sources and the kodak effect. Information growth/432291/
Systems Research, 3(4), 334-358.
McKevitt, J. (2017). Macy's RFID effort boosts sales,
Lopez, E. (2017). Report: A decade later, 96% of fulfillment. Supplychaindive.com. https://www.
retailers deploying RFID tags on apparel. supplychaindive.com/news/RFID-Macys-
Supply Chain Dive. https://www.supply success-inventory-fulfillment-markdown-
chaindive.com/news/rfid-retail-technology- Platt/440827/
implementation/435285/
McWilliams, G. (2007, February 15). Wal-Mart’s
Lopez, E. (2018). RFID can achieve near 100% order radio-tracked inventory hits static. Wall Street
accuracy, study says. Supplychaindive.com. Journal. http://online.wsj.com/article_email/
https://www.supplychaindive.com/news/RFID SB117150681979009405-lMyQjAxMDE3N
-100-accurate-ROI-Auburn/539449/ zExNTUxMDU2Wj.html
Maguire, S., Hardy, C., & Lawrence, T. B. (2004). Michel, R. (2015). Is RFID ready for a reinvention?
Institutional entrepreneurship in emerging Modern Materials Handling. https://www.
fields: HIV/AIDS treatment advocacy in mmh.com/article/is_rfid_ready_for_a_reinvention
Canada. Academy of Management Journal,
Miles, M. B., Huberman, A. M., & Saldana, J. (2018).
47(5), 657-679.
Qualitative data analysis: A methods
Malone, M. (2012). Did Wal-Mart love RFID to sourcebook. SAGE.
death? ZDnet. https://www.zdnet.com/article/
Miles, M. B., & Huberman, M. A. (1994). Qualitative
did-wal-mart-love-rfid-to-death/
data analysis: An expanded sourcebook.
Marsan, J., Carillo, K. D. A., & Negoita, B. (2020). SAGE.
Entrepreneurial actions and the legitimation of
Miranda, S. M., Kim, I., & Summers, J. D. (2015).
free/open source software services. Journal of
Jamming with Social Media. MIS Quarterly,
Information Technology, 35(2), 143-160.
39(3), 591-614.
Marsan, J., Paré, G., & Beaudry, A. (2012). Adoption
Misangyi, V. F., Weaver, G. R., & Elms, H. (2008).
of open source software in organizations: A
Ending corruption: The interplay among
socio-cognitive perspective. The Journal of
institutional logics, resources, and institutional
Strategic Information Systems, 21(4), 257-273.
entrepreneurs. Academy of Management
Mayberry, R. (2005). Beyond RFID, A new era of Review, 33(3), 750-770.
wireless sensors. TechTarget. https://search
Mola, L., Carugati, A., Giangreco, A., & Da Cunha, J.
mobilecomputing.techtarget.com/feature/Beyo
V. (2022). Learning from unexpected
nd-RFID-A-new-era-of-wireless-sensors
technological success: An extended model of
Mayfield, K. (2002). Radio ID tags: Beyond
supply-side diffusion. European Journal of
bar codes. Wired. https://www.wired.com/
Information Systems, 31(5), 597-616.
2002/05/radio-id-tags-beyond-bar-codes/
https://doi.org/10.1080/0960085X.2021.19072
McBeath, B. (2018). The ROI for RFID technology in 32
retail. Loss Prevention Magazine.

422
Journal of the Association for Information Systems

Moore, G., & Benbasat, I. (1991). Development of an O’Connor, M. C. (2006c). Wal-Mart Specifies Gen 1
instrument to measure the perceptions of Sunset, Forklift Pilot. RFID Journal.
adopting an information technology innovation. http://www.rfidjournal.com/article/view/2271/
Information Systems Research, 2(3), 192-222. 1/1
Moore, G. A. (2014). Crossing the chasm: Marketing O’Connor, M. C. (2007). Wal-Mart, Sam's Club Push
and selling high-tech products to mainstream RFID Further Along. RFID Journal.
customers. Harper Collins. http://www.rfidjournal.com/article/articleprint/
3666/-1/1/
Moore, L. (2019). Walmart and RFID: The
relationship that put RFID on the map. Atlas O’Connor, M. C. (2009). Bloomingdale’s Tests Item-
RFID Store. https://www.atlasrfidstore. Level RFID. RFID Journal. https://www.
com/rfid-insider/walmart-and-rfid-the- rfidjournal.com/bloomingdales-tests-item-
relationship-that-put-rfid-on-the-map level-rfid
Morrill, C. (2006). Institutional change through O’Connor, M. C. (2014). Can RFID save brick-and-
interstitial emergence: The growth of mortar retailers after all? Fortune.
alternative dispute resolution in American Law, https://fortune.com/2014/04/16/can-rfid-save-
1965-1995. In W. W. Powell & D. L. Jones brick-and-mortar-retailers-after-all/
(Eds.), How institutions change. University of
O’Connor, M. C., & Roberti, M. (2005). Impinj
Chicago Press.
announces Gen 2 tags, reader. RFID Journal.
Munir, K. A., & Phillips, N. (2005). The birth of the http://www.rfidjournal.com/article/articleprint/
“Kodak moment”: Institutional 1482/-1/1/
entrepreneurship and the adoption of new
Okhuysen, G., & Bonardi, J.-P. (2011). The challenges
technologies. Organization Studies, 26(11),
of building theory by combining lenses.
1665-1687.
Academy of Management Review, 36(1), 6-11.
Murphy, C. (2005). Real-World RFID: Wal-Mart,
Olson, J. (2017). A brief history of EMV technology.
Gillette, and others share what they're learning.
Fattmerchant. https://fattmerchant.com/blog/
InformationWeek. https://www.information
brief-history-emv-technology/
week.com/real-world-rfid-wal-mart-gillette-
and-others-share-what-theyre-learning/d/d- Orem, T. (2019). EMV chip adoption credited with
id/1033036 76% drop in card-present fraud. Credit Union
Times. https://www.cutimes.com/2019/06/05/
Murphy, J. V. (2003). Get ready! Wal-Mart Mandate
emv-chip-adoption-credited-with-76-drop-in-
puts RFID, smart tags on fast track. Supply
card-present-fraud/?slreturn=20190704224502
Chain Brain. https://www.supplychainbrain.
com/articles/1085-get-ready-wal-mart- Paré, G., Marsan, J., Jaana, M., Tamim, H., &
mandate-puts-rfid-smart-tags-on-fast-track Lukyanenko, R. (2020). IT vendors’
legitimation strategies and market share: The
Myers, M. D. (1997, June 1997). Qualitative research
case of EMR systems. Information &
in information systems. MIS Quarterly, 21(2),
Management, 57(5), Article 103291.
241-242.
Patton, M. Q. (2014). Qualitative research and
Nelson, M. G. (2001). The fast track.
evaluation methods. SAGE.
InformationWeek. http://www.information
week.com/the-fast-track/d/d-id/1010866 Pentland, B. T. (1999). Building process theory with
narrative: From description to explanation.
Nielsen, J., Mathiassen, L., & Newell, S. (2014).
Academy of Management Review, 24(4), 711-724.
Theorization and translation in information
technology institutionalization: Evidence from Pettey, C. (2015). EMV won’t end card fraud, and may
Danish home care. MIS Quarterly, 38(1), 165-186. slow you down at the checkout. Gartner.
https://www.gartner.com/smarterwithgartner/e
O’Connor, M. C. (2006a). Gen 2 EPC Protocol Approved
mv-wont-end-card-fraud-and-may-slow-you-
as ISO 18000-6C. RFID Journal.
down-at-the-checkout/
http://www.rfidjournal.com/article/view/2481/1/1
Phillips, N., Lawrence, T. B., & Hardy, C. (2004).
O’Connor, M. C. (2006b). RFID Consortium Names
Discourse and institutions. Academy of
Patent-Pool Administrator. RFID Journal.
Management Review, 29(4), 635-652.
http://www.rfidjournal.com/article/articleprint/
2636/-1/1/ Prince, P. (2013). RFID a “very big part of Macy’s future.”
RFID Journal. https://www.rfidjournal.com/rfid-a-
very-big-part-of-macys-future

423
The Early-Stage Diffusion of Codependent IT Innovations

PRNewswire. (2004). Albertsons launches RFID RFIDJournal. (2005). Four lessons of the Wal-Mart
technology program to improve merchandising mandate. https://www.rfidjournal.com/four-
and supply chain management efficiencies. lessons-of-the-wal-mart-mandate
http://www.prnewswire.com/cgi-bin/
RFIDJournal. (2008a). Sam’s Club RFID mandate no
stories.pl?ACCT=104&STORY=/www/story/
big deal? http://www.rfidjournal.com/articles/
03-05-2004/0002122904&EDATE=
view?6921
Purdy, J. M., & Gray, B. (2009). Conflicting logics,
RFIDJournal. (2008b). Sam's Club suppliers may face
mechanisms of diffusion, and multilevel
RFID fines from Wal-Mart. http://www.rfid
dynamics in emerging institutional fields.
journal.com/articles/view?6896
Academy of Management Journal, 52(2), 355-
380. RFIDJournal. (2009a). RFID leader P&G steps back
from promotions tracking. http://www.rfid
Quaadgras, A. (2005). Who joins the platform? The
journal.com/articles/view?7122
case of the RFID business ecosystem.
Proceedings of the Hawaii International RFIDJournal. (2009b). Sam’s club letter outlines
Conference on System Sciences. changes to RFID requirements. http://www.rfid
journal.com/articles/view?7109
Quinn, F. (2003, 01-MAY-03). Ready for the auto-ID
revolution: Interview with RFID pioneer Kevin RFIDUpdate. (2003). Is this the future of retailing?
Ashton. Supply Chain Management Review, http://www.rfidupdate.com/articles/index.php?
7(3), 44-51. id=274
RAINRFID. (2014). Google, Impinj, Intel, Smartrac & RFIDUpdate. (2004). RFID standard “GEN 2”
AIM Global launch RAIN RFID alliance to ratified. http://www.rfidupdate.com/articles/
promote UHF RFID industry growth. index.php?id=699
https://rainrfid.org/google-impinj-intel-
smartrac-aim-global-launch-rain-rfid-alliance- RFIDUpdate. (2005). Intermec and metro getting read
rates of 99%. http://www.rfidupdate.com/
to-promote-uhf-rfid-industry-growth/
articles/index.php?id=735
Ramiller, N., & Swanson, E. (2003). Organizing
RFIDUpdate. (2008). Sam’s club suppliers may face
visions for information technology and the
RFID fines from Wal-Mart. http://www.
information systems executive response.
Journal of Management Information Systems, rfidupdate.com/articles/index.php?id=1431
20(1), 13-50. Roberti, M. (2002). Why the Auto-ID Center matters.
RFID Journal. https://www.rfidjournal.com/
RedPrairie. (2003). RFID: Just the facts
articles/view?137
http://www.werc.org/assets/1/workflow_stagin
g/Publications/395.PDF Roberti, M. (2003a). Analysis: RFID: Wal-Mart’s
network effect. CIO Insight, https://www.
Ricci, C. (2010). Retailers Buy into item-level tagging.
Apparel magazine. https://apparelmag.com/ cioinsight.com/news-trends/analysis-rfid-wal-
retailers-buy-item-level-tagging marts-network-effect/1(30)
http://www.cioinsight.com/print_article/0,366
Riggins, F. J., & Slaughter, K. T. (2006). The role of 8,a=61672,00.asp
collective mental models in IOS adoption:
Roberti, M. (2003b). The Real Scandal. RFID Journal.
Opening the black box of rationality in RFID
https://www.rfidjournal.com/articles/view?654
deployment. Proceedings of the Hawaii
International Conference on System Sciences. Roberti, M. (2004). Best Buy to Deploy RFID. RFID
Rivard, S. (2014). Editor’s comments: the ions of theory Journal. http://www.rfidjournal.com/article/
construction. MIS Quarterly, 38(2), iii-xiv. articleprint/1104/-1/1/
Roberti, M. (2005a). The history of RFID technology.
RFIDJournal. (2002). Auto-ID Center adds RFID vendors.
RFID Journal. http://www.rfidjournal.com/
http://www.rfidjournal.com/articles/view?7
articles/view?1338
RFIDJournal. (2003). Wal-Mart draws line in the sand.
http://www.rfidjournal.com/article/view/462/1/1 Roberti, M. (2005b). A Look Back at 2005. RFID
Journal. http://www.rfidjournal.com/article/
RFIDJournal. (2004a). Target issues RFID mandate. articleview/2048/1/2/
http://www.rfidjournal.com/article/view/802/1/1/
Roberti, M. (2005c). Retailers say RFID will take time.
RFIDJournal. (2004b). Wal-Mart details RFID RFID Journal. http://www.rfidjournal.com/
requirement. http://www.rfidjournal.com/ article/articleview/1344/1/1/
article/articleprint/642/-1/1/

424
Journal of the Association for Information Systems

Roberti, M. (2005d). Target, Wal-Mart share EPC Songini, M. L. (2007b, February 26). Wal-Mart shifts
data. RFID Journal. http://www.rfid RFID plans. Computerworld. http://www.
journal.com/article/articleprint/1928/-1/1 computerworld.com/action/article.do?comman
d=printArticleBasic&articleId=284115
Roberti, M. (2007). Metro pushes pallet tagging. RFID
Journal. http://www.rfidjournal.com/article/ Soon, C.-B., & Gutiérrez, J. A. (2008). Effects of the
articleview/3341/ RFID mandate on supply chain management.
Journal of Theoretical and Applied Electronic
Roberti, M. (2010a). Wal-Mart relaunches EPC RFID
Commerce Research, 3(1), 81-91.
effort, starting with men's jeans and basics.
RFID Journal. https://www.rfidjournal.com/ Strang, D., & Soule, S. A. (1998). Diffusion in
articles/view?7753/ organizations and social movements: From
hybrid corn to poison pills. Annual Review of
Roberti, M. (2010b). Wal-Mart takes a new approach
Sociology, 24(1), 265-290.
to RFID. RFID Journal. https://www.rfid
journal.com/articles/view?7756 Studio98Test. (2009). Dillard’s, U. of Ark. study
quantifies RFID’s superiority to manual
Roberti, M. (2011). Item level RFID initiative focuses
inventory counts. https://www.studio98test.
on supplier benefits. RFID Journal.
com/ahmad-new-test-articles/dillards-u-of-ark-
https://rfid.grandcentr.al/articles/item-level-
study-quantifies-rfids-superiority-to-manual-
rfid-initiative-focuses-on-supplier-benefits
inventory-counts
Roberti, M. (2014). When did Wal-Mart start using
Studio98Test. (2012). University of Arkansas’ new
RFID? RFID Journal. https://www.rfidjournal.
RFID research center parallels growth of RFID
com/blogs/experts/entry?11217
industry. https://www.studio98test.com/
Roberti, M. (2016a). Macy’s to RFID-Tag 100 Percent ahmad-new-test-articles/university-of-
of Items. RFID Journal. https://www.rfid arkansas-new-rfid-research-center-parallels-
journal.com/macys-to-rfid-tag-100-percent-of- growth-of-rfid-industry
items
Su, Y.-f., & Yang, C. (2010). Why are enterprise
Roberti, M. (2016b). Rebecca Minkoff adds RFID to resource planning systems indispensable to
more stores, boosts sales. RFID Journal. supply chain management? European Journal
https://www.rfidjournal.com/rebecca-minkoff- of Operational Research, 203(1), 81-94.
adds-rfid-to-more-stores-boosts-sales
Sun Microsystems. (2003). SUN to open a Wal-Mart
Roberti, M. (2017). Is retail approaching the tipping point compliant RFID test center. http://www.sun.
for RFID? RFID Journal. com/smi/Press/sunflash/2003-
https://www.rfidjournal.com/articles/view?16668 11/sunflash.20031105.4.xml
Rogers, E. (2003). Diffusion of innovations (4th ed.). Sun, S., Cegielski, C. G., Jia, L., & Hall, D. J. (2018).
Free Press. Understanding the factors affecting the
organizational adoption of big data. Journal of
SCDigest. (2009). RFID news: Looking back at the Computer Information Systems, 58(3), 193-203.
Wal-Mart RFID time line. http://www.scdigest.
com/assets/On_Target/09-02-23-1.php Swanson, E. B., & Ramiller, N. C. (1997). The Organizing
vision in information systems innovation.
Schoenberger, C. R. (2002). RFID: The internet of
Organization Science, 8(5), 458-474.
things. Mindfully. http://www.mindfully.org/
Technology/RFID-Things-Forbes18mar02.htm Swanson, E. B., & Ramiller, N. C. (2004). Innovating
mindfully with information technology. MIS
Semilof, M. (2001). Bar Codes in a Chip. Iranscope. Quarterly, 28(4), 553-583.
http://iranscope.ghandchi.com/Anthology/Com
punet/barcode.htm Swedberg, C. (2007a). Movie, Music Suppliers Opt for
Simple RFID Compliance. RFID Journal.
Sliwa, C. (2004, May 19). Wal-Mart updates RFID http://www.rfidjournal.com/article/articleprint/
road map, revises expectations for ‘05 deadline. 2989/-1/1
Computerworld. http://www.computerworld.
com/printthis/2004/0,4814,93266,00.html Swedberg, C. (2007b). Wal-Mart, Best Buy spearhead
DVD-tagging pilot. RFID Journal.
Songini, M. L. (2007a, February 26). Procter & http://www.rfidjournal.com/article/articleprint/
Gamble: Wal-Mart RFID effort effective. 3645/-1/1
Computerworld. http://www.computerworld.
com/action/article.do?command=printArticleB Swedberg, C. (2010). Major retailers, industry groups
asic&articleId=284160 launch item-level RFID guidelines initiative.

425
The Early-Stage Diffusion of Codependent IT Innovations

RFID Journal. https://rfid.grandcentr.al/ TalkBusiness. (2004). RFID could cost average Wal-
articles/major-retailers-industry-groups- Mart vendor $9 million. https://talkbusiness.
launch-item-level-rfid-guidelines-initiative net/2004/04/rfid-could-cost-average-wal-mart-
vendor-9-million/
Swedberg, C. (2011a). Macy’s Inc. to begin item-level
tagging in 850 stores. RFID Journal. Thau, B. (2017). Is the “RFID retail revolution” finally
ttps://www.rfidjournal.com/macys-inc-to- here? A Macy’s case study. Forbes.
begin-item-level-tagging-in-850-stores https://www.forbes.com/sites/barbarathau/201
7/05/15/is-the-rfid-retail-revolution-finally-
Swedberg, C. (2011b). University of Arkansas study
here-a-macys-case-study/#4f46714b3294
finds 60 ways to use RFID in apparel supply
chain. RFID Journal. https://www.rfidjournal. Tolbert, P. S., & Zucker, L. G. (1983). Institutional
com/university-of-arkansas-study-finds-60- sources of change in the formal structure of
ways-to-use-rfid-in-apparel-supply-chain organizations: The diffusion of civil service
reform, 1880-1935. Administrative Science
Swedberg, C. (2011c). VICS item level RFID initiative
Quarterly, 28(1), 22-39.
enters phase II. RFID Journal. https://www.
rfidjournal.com/vics-item-level-rfid-initiative- University of Arkansas. (2009). New Study Shows
enters-phase-ii RFID significantly improves item-level
inventory accuracy. https://news.uark.edu/
Swedberg, C. (2013). Read range for EPC Gen 2 UHF
articles/14256/new-study-shows-rfid-
tags rises 20 percent this year, according to
significantly-improves-item-level-inventory-
EECC. RFID Journal. https://www.rfidjournal.
accuracy
com/articles/view?10959
Vijayan, J., & Brewin, B. (2003). Wal-Mart backs
Swedberg, C. (2014a). Macy’s expands RFID and
RFID technology. Computerworld. http://
beacon deployments. RFID Journal.
www.computerworld.com/article/2570642/ent
https://www.rfidjournal.com/macys-expands-
erprise-resource-planning/wal-mart-backs-rfid-
rfid-and-beacon-deployments
technology.html
Swedberg, C. (2014b). Rebecca Minkoff store uses
Violino, B. (2002). Will Wal-Mart order RFID tagging?
RFID to provide an immersive experience.
Studio98test. https://www.studio98test.com/
RFID Journal. https://www.rfidjournal.com/
ahmad-test/will-wal-mart-order-rfid-tagging
rebecca-minkoff-store-uses-rfid-to-provide-an-
immersive-experience Violino, B. (2003a). The cost of Wal-Mart’s RFID
edict. RFID Journal. http://www.rfidjournal.
Swedberg, C. (2016a). American Apparel, Postmates
com/articles/view?572
use RFID visibility for on-demand delivery.
RFID Journal. https://www.rfidjournal. Violino, B. (2003b). The shelved shelf. RFID Journal.
com/american-apparel-postmates-use-rfid- https://www.rfidjournal.com/articles/view?501
visibility-for-on-demand-delivery
Violino, B. (2003c). Wal-Mart expands RFID
Swedberg, C. (2016b). Large apparel retailers, mandate. RFID Journal. https://www.rfid
wholesalers see sharp rise in RFID adoption, journal.com/articles/view?539
ROI. RFID Journal. https://www.rfidjournal.
Violino, B. (2003d). EAN and UCC form EPCglobal,
com/articles/view?15121
Inc. RFID Journal. https://www.rfid
Swedberg, C. (2016c). Macy’s launches pick to the last journal.com/ean-and-ucc-form-epcglobal-inc
unit program for omnichannel sales. RFID
Visa. (2019). Chip technology helps reduce counterfeit
Journal. https://www.rfidjournal.com/macys-
fraud by 76 percent. https://usa.visa.com/visa-
launches-pick-to-the-last-unit-program-for-
everywhere/blog/bdp/2019/05/28/chip-
omnichannel-sales
technology-helps-1559068467332.html
Swedberg, C. (2016d). Reebok, music festival reach
Wailgum, T. (2004, November 15). Tag, You're Late.
out to visitors with iGotcha RFID solution.
CIO, 18(4), 6-12.
RFID Journal. https://www.rfidjournal.com/
reebok-music-festival-reach-out-to-visitors- Wailgum, T. (2006, September 15). How
with-igotcha-rfid-solution-2 manufacturers can reap the benefits of RFID.
CIO, 18, https://www.cio.com/article/
Swedberg, C. (2019). RFID enters early majority phase in
260622/rfid-how-manufacturers-can-reap-the-
retail. RFID Journal. https://www.rfidjournal.com/
benefits-of-rfid.html
articles/view?18147/
Wailgum, T. (2010). RFID tags arrive on store floor
as retailers go after ROI. Network World.

426
Journal of the Association for Information Systems

https://www.networkworld.com/article/221499 Whitcomb, G. (2004a). Wal-Mart EPC pilot begins


6/rfid-tags-arrive-on-store-floor-as-retailers- transition to implementation. Wal-Mart Store.
go-after-roi.html http://www.walmartfacts.com/articles/2612.as
px
WalmartFact. (2005). Wal-Mart improves on-shelf
availability through the use of electronic product Whitcomb, G. (2004b, 06/17). Wal-Mart expands
codes. Wal-Mart Store. http://www.walmartfacts. electronic product code goals. Wal-Mart Store.
com/articles/1937.aspx http://www.walmartfacts.com/_resources/Tool
Bar/printerfriendly.aspx?id=2596&PageType=
WalmartFact. (2006, 09/12). Wal-Mart continues RFID
1&lang=-1
expansion. Wal-Mart Store. http://www.walmart
facts.com/articles/4435.aspx Whitcomb, G., & Gallagher, C. (2004, 04/30). Wal-
Mart begins roll-out of electronic product codes
Walsham, G. (1993). Interpreting information systems
in Dallas/Fort Worth Area. Wal-Mart Store.
in organizations. Wiley.
http://www.walmartfacts.com/_resources/ToolB
Walsham, G. (1995). Interpretive case studies in IS ar/printerfriendly.aspx?id=2603&PageType=1
research: nature and method. European Journal &lang=-1
of Information Systems, 4(2), 74-81.
Whitcomb, G., & Gallagher, C. (2005, 01/04).
Wang, P., & Ramiller, N. C. (2009). Community Countdown begins to Wal-Mart's Latest EPC
learning in information technology innovation. Milestone. Wal-Mart Store. http://www.
MIS Quarterly, 33(4), 709-734. walmartfacts.com/articles/2658.aspx
Wang, P., & Swanson, B. (2008). Customer Wired. (2003). Wal-Mart, DOD forcing RFID..
relationship management as advertised: https://www.wired.com/2003/11/wal-mart-
Exploiting and sustaining technological dod-forcing-rfid/
momentum. Information Technology & People,
Wollenhaupt, G. (2018). What’s next for RFID?
21(4), 323-349.
Supply Chain Dive. https://www.supply
Wang, P., & Swanson, E. B. (2007). Launching chaindive.com/news/RFID-what-is-
professional services automation: Institutional next/544108/
entrepreneurship for information technology
Woodie, A. (2016). Why RFID is (finally) here to stay.
innovations. Information and Organization,
IT Jungle. https://www.itjungle.com/2016/09/
17(2), 59-88.
28/tfh092816-story02/
Warry, C. (2015). The return of RFID to the retailers’
Wysong, B. J. (2005). Bar codes, RFID or both? Fisher
radar. PwC. https://www.digitalpulse.pwc.
College of Business.
com.au/rfid-retail-1/
Xu, X., Thong, J. Y., & Tam, K. Y. (2017). Winning back
Webster, J. (2008). Wal-Mart’s RFID revolution a
technology disadopters: testing a technology
tough sell. Network World. https://www.
readoption model in the context of mobile internet
networkworld.com/article/2275035/wireless/w
services. Journal of Management Information
al-mart-s-rfid-revolution-a-tough-sell.html
Systems, 34(1), 102-140.
Wehr, J. (2004). The Wal-Mart effect: Retailer
Yin, R. K. (2008). Case study research: Design and
mandates RFID tagging of product shipments
methods. SAGE.
and others follow suit. SecureIDNews.
https://www.secureidnews.com/news-item/the- Zaczkiewicz, A. (2016). RFID adoption gathers
wal-mart-effect-retailer-mandates-rfid- “momentum” amid drive to improve inventory.
tagging-of-product-shipments-and-others- WWD. https://wwd.com/business-news/retail/
follow-suit/ rfid-adoption-increases-retail-10512580/
Wehr, J. (2005). A History of the EPC. SecureIDNews. Zaino, J. (2016). Riding the tails of apparel retailers.
http://www.secureidnews.com/news-item/a- RFID Journal. https://www.rfidjournal.com/
history-of-the-epc/ riding-the-tails-of-apparel-retailers
Whitaker, J., Mithas, S., & Krishnan, M. S. (2007). A Zaino, J. (2017). Tagging retail items at the source.
field study of RFID deployment and return RFID Journal. https://www.rfidjournal.com/
expectations. Production and Operations tagging-retail-items-at-the-source
Management, 16(5), 599-612.

427
The Early-Stage Diffusion of Codependent IT Innovations

Appendix A: Generalizability
The EMV technology is another example of a codependent innovation. The complementary parts are the EMV chip
credit\debit cards and the chip readers. The two different adopter communities are banks who issue EMV cards to their
customers and merchants and retailers who install chip readers at point-of-sale terminals. Our analysis shows that the
early-stage diffusion of EMV technology in US retailing followed the process model we discerned in our study of
RFID diffusion in retailing. EMV cards create a unique transaction code for every transaction at a point of sale, which
makes these cards resistant to counterfeit frauds that magnetic strip-based cards are prone to (Kossman, 2017). Before
its implementation in the US, EMV card implementations in other countries had established massive benefits in
decreasing counterfeit fraud (Visa, 2019; Kossman, 2017). Banks perceived a higher coadopter relative advantage with
EMV cards than retailers and acted as IEs in the early stages of EMV card diffusion. EMV cards benefit banks the
most because they are liable for customer losses due to counterfeit fraud from in-store transactions; retailers and
merchants, especially consumer-oriented ones, face a loss of reputation when a security breach happens at their stores
(Chandler, 2014). However, EMV adoption would require banks to reissue EMV credit cards to all customers and set
up the supporting IT infrastructure, and would require retailers to install chip readers and associated software in all
point-of-sale terminals (Chandler, 2014). For retailers, EMV cards can frustrate customers because the checkout
process takes longer than before (i.e., using magnetic stripes). With security breaches becoming commonplace in point-
of-sale terminals in the US, banks took a leadership role in introducing EMV cards in the US by announcing their
migration plans (David & Best, 2014). Even in Europe, where EMV cards first originated, it was banks\credit card
companies who developed a rationale by inventing the EMV technology and drafting the standards (EMV, 2019;
Olson, 2017). When merchants in the US resisted EMV cards mainly due to costs associated with implementation,
banks issued a shift in liability deadline after which the EMV noncompliant merchants would have to bear the costs of
counterfeit frauds incurred by customers at the in-store point-of-sale—a move very similar to mandates issued by
Walmart at the structuralization phase (Pettey, 2015). This strategy was pivotal in influencing other banks to issue
similar notices and bringing retailers on board. Banks developed relationships by working closely with vendors and
merchants to conduct pilots and make the EMV technology easy to use for customers (Kossman, 2016). Banks utilized
the OV by positioning the technology as a solution for security breaches, outlining EMV transition deadlines, and
publishing research on fraud reductions (Visa, 2019; Orem, 2019). These efforts lead to the widespread joint adoption
of EMV cards by banks and EMC card readers by retailers (Orem, 2019).
By showing our model’s applicability in the case of another codependent innovation in retail, we strengthen the face
validity of our findings. However, it is important to note that the US was one of the last countries adopting EMV
technology, which was already a mature and standardized technology in European retailing before it reached the US
markets. On the other hand, RFID in retailing was in its infancy, with businesses developing the technology and
identifying the use cases in retail. Also, retailing in Europe has fewer banks and credit card companies than the
thousands in its US counterpart.
Although we limited our data collection and analysis to a single case study, we believe that our theoretical insights are
generalizable within the boundary of codependent innovations in the retailing context with large players. In terms of
the Lee and Baskerville (2003) generalizability framework, ET (from empirical statements to theoretical statements)
generalizability applies to our study. We demonstrate the generalizability of our model to a new setting by analyzing
the early-stage EMV technology diffusion in US retailing. Our analysis shows that the early-stage diffusion of EMV
technology in US retailing followed the process model we theorized in our study of early-stage RFID diffusion in
retailing. It is important and ethical to judge and report whether the conditions in settings from which the theory is
generalized from (RFID technology) are similar to new settings (EMV technology) to which theory is generalized (Lee
& Baskerville, 2012). Thus, in the previous paragraph, we outlined how this new setting (EMV in the US) is different
from RFID in retailing, based on which we theorized our process model.

428
Journal of the Association for Information Systems

About the Authors


Srikanth Parameswaran is an assistant professor of management information systems in the School of Management
at Binghamton University. His research interests are in the areas of IT innovation, user-generated and web content,
and technology-mediated health outcomes. His research has been published in Organizational Research Methods,
Information and Management, Journal of the American Medical Informatics Association, Information Systems
Frontiers, and other outlets. He earned his MS in management information systems and PhD in management science
and systems from the State University of New York at Buffalo. He received his bachelor’s degree in computer science
and engineering from Anna University, India. He was featured on the Poets & Quants list of the “Top 50 Undergraduate
Business Professors of 2021.”
Rajiv Kishore is a professor of information systems and the department chair of the Management, Entrepreneurship, and
Technology Department at the University of Nevada, Las Vegas. He received his MS and PhD degrees in computer
information systems from Georgia State University. His current research interests are in the areas of online markets and
platforms, social media, and virtual communities in the context of healthcare; IT sourcing, governance, and impacts; IT
innovation; and corruption and corporate social responsibility. He has published his work in premier journals in the
information systems, medical informatics, operations management, organizational behavior, and strategic management
fields, including MIS Quarterly, Journal of Management Information Systems, Journal of the American Medical
Informatics Association, Production and Operations Management, Journal of Organizational Behavior, and Strategic
Management Journal, among others. He has also received multiyear research grants as PI and Co-PI from prestigious
federal agencies including NSF and AHRQ. Rajiv currently serves as an associate editor at Information and Management
and as an editorial board member at Healthcare. His ORCID ID is 0000-0002-9476-4479.
Xuanhui Yang received his PhD degree from Xiamen University, China. His research interests are in the areas of
innovation and entrepreneurship. He is currently a data scientist at Wilsonart LLC and focuses on effectiveness
measurement and operations optimization.
Zhenyu Liu is a professor of information systems in the School of Management at Xiamen University, China. He
received his bachelor’s degree in automation engineering (1982) and master’s degrees in management information
systems (1985) from Huazhong University of Science and Technology, Wuhan, China. He received his PhD degree in
business informatics from University of Freiburg, Germany in 1998. Zhenyu has taught courses on information
systems, operations management, and electronic commerce. His research interests include interorganizational
information systems, B2B e-commerce, coordination mechanisms, and the economic analysis of networks.

Copyright © 2023 by the Association for Information Systems. Permission to make digital or hard copies of all or part
of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for
profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for
components of this work owned by others than the Association for Information Systems must be honored. Abstracting
with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists requires prior
specific permission and/or fee. Request permission to publish from: AIS Administrative Office, P.O. Box 2712 Atlanta,
GA, 30301-2712 Attn: Reprints, or via email from publications@aisnet.org.

429
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.

You might also like