Cross Platform Analysis of Seller Performance and Churn For Ecommerce Using Artificial Intelligence

Journal of Global Information Management
Volume 31 • Issue 1
Cross-Platform Analysis of Seller

Performance and Churn for Ecommerce
Using Artificial Intelligence
Anuj Batta, Indian Institute of Technology, Delhi, India
https://orcid.org/0000-0002-8583-5781
Arpan Kumar Kar, Indian Institute of Technology, Delhi, India*

Shyamali Satpathy, IILM University, Greater Noida, India
ABSTRACT
Suppliers and sellers play a crucial role in the ecommerce ecosystem. Sellers and ecommerce firms
use social media to increase user engagement, visibility, and sales. Seller ratings are as important as
the product ratings on ecommerce platforms to drive buying decisions. Based on sellers’ actions on
social media, this study examines seller turnover and disengagement on e-commerce platforms. The
study has been supported by the justice theory. Seller reviews and ratings from e-commerce platforms
and conversations from social media platforms have been gathered. Using natural language processing,
machine learning, partial least squares (PLS) path analysis, and statistical inferences, objectives of the
study are met. The study offers recommendations for both practitioners and researchers. The sellers
must focus more on interaction and communication than marketing. Through a longitudinal analysis,
the study also establishes that ecommerce organizations can use seller social media performance as
a predictor of future seller churn and disengagement so they can take the necessary remedial action.
Keywords:
Cross platform Impact, E-Commerce, Machine learning, Natural language processing, Seller Churn, Seller
disengagement, Seller Review Consensus, Social Media Content Type
INTRODUCTION
Almost half of the online purchase decisions are based on online reviews and ratings given by
customers in the ecommerce context (Mosteller & Mathwick, 2016). Customer reviews and ratings
are more trusted and considered highly credible in comparison to the advertisements and promotions
by the brands or sellers (Thakur, 2018). It is not only the product and brand rating or reviews which
are important, however, the seller reviews and ratings also have an impact on the purchase decision
DOI: 10.4018/JGIM.322439 *Corresponding Author

This article published as an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0/) which permits unrestricted use, distribution, and production in any medium,
provided the author of the original work and original publication source are properly credited.
1
by customers (Chen et al., 2008). Thus, retailers, sellers, and suppliers continue to invest heavily
in various techniques and methods, apart from providing the best possible customer experience, to
get better online ratings and reviews. The available literature in this domain indicates that customer
engagement is a primary driver which leads to several effective outcomes, including positive ratings
and reviews leading to a positive buying decision by customers (Brodie et al., 2013; Kaur et al.,
2020; Xiao & Li, 2019).
Sellers and suppliers use social media platforms as one of the primary channels for customer
engagement and interaction. Millions of customer interactions occur every day on social media
platforms such as Facebook, Twitter, YouTube, etc. (Appel et al., 2019; Gupta and Ramachandran,
2021). There is a robust interrelated phenomenon between customer relationship management, social
media technologies, customer engagement, positive word of mouth and brand loyalty (Dewnarain et
al., 2019). With the exponential growth in social media users and the trend reflecting a fundamental
shift from primarily company-customer interactions to customer-customer interactions that impact
company-customer relationships, it is imperative for sellers to have an effective social media customer
engagement plan (Agnihotri, 2020). Social media appeared as, and continues to be a business
phenomenon. Increasingly, current and prospective customers use social media to communicate about
the products and services they plan to buy. At the same time, sellers or suppliers are also using the
social media platform to express their products and services and resolve customer issues (Kaplan &
Haenlein, 2010; Laroche et al., 2013).
Social media customer engagement for most companies began on external social media sites
such as Facebook, Twitter, LinkedIn etc. Today, 90% of mid-size and large enterprises spent at least
11% of their total marketing expenses on these platforms (de Oliveira Santini et al., 2020). As this
is a huge cost for companies, there has been an immense interest in evaluating the impact and return
on investment of this spend, however, it is not an easy and a straightforward mechanism. There are
several issues including limitation of the social media analysis tools to only include keywords, topics
or sentiment analysis, difficulty in identifying the right questions, suggestions, desires etc., and most
importantly the failure in considering the communication in the conversation context (Abbasi et al.,
2018). Due to the high cost and ineffective ways to measure the impact, companies quickly realized
multifold benefits of supporting customer communities on their own website. Customers can raise
questions or complaints about the product either on a regular website from where they buy the product
or on the social media platform. Also, companies are utilizing these platforms to communicate their
brands and product to customers. Most companies that engage with customers online, do so on both
of these avenues, i.e. social media platforms and their own websites (Connell et al., 2019; Kumar
et al., 2022).
Looking at the literature in this domain, it is evident that there is a significant amount of work
done in the area of customer engagement on social media, the impact of product or seller reviews,
word of mouth, customer loyalty, buying intention, and customer retention (de Oliveira Santini et al.,
2020; Rosado-Pinto & Loureiro, 2020). However most of the work has been unidimensional which
looks at the impact of customer engagement independently for a particular objective and on a single
platform (Gligor et al., 2019; Hollebeek et al., 2021; Yen et al., 2020). The current study explores a
very interesting field of work which is at a nascent stage, where the interrelationship between customer
engagement on social media platforms, and reviews or rating intentions on ecommerce platforms is
looked at (Thakur, 2018).
This study contributes to the e-commerce literature and industry practice through a cross platform
behavior analysis. It explores the impact of customer engagement by the sellers through a social media
platform i.e. Twitter, on the reviews and rating intention of customers on an ecommerce platform,
Amazon. Twitter has been chosen for this study because it is an excellent source of opinion data
from customers in large numbers and in the form of microblogging, which can give effective and
efficient insights needed for this study (Giachanou & Crestani, 2016). This study also establishes a
model to predict sellers’ churn on the ecommerce platform. It explores the degree of engagement that
2
a seller demonstrates with the buyers on Twitter and the overall ratings given to the same seller on
Amazon. It will also help to analyze the degree of consensus in the reviews provided by the buyers
to these sellers, both for positive and negative reviews, along with the average review polarity. Most
importantly, it establishes the relationship between the sellers’ behavior or engagement on Twitter
and the sellers’ churn on Amazon. Using a combination of methods from big data analytics, natural
language processing, machine learning, and statistical tests for inferences, the research objectives has
been achieved. To summarize, following are the objectives of this study:
1. To determine the relationship between sellers’ engagement with the customers (including customer
orientation) on Twitter, and the overall seller ratings on Amazon,
2. To identify the degree of consensus in the reviews given by customers for the sellers on Amazon,
3. To identify the relationship between the degree of consensus in the reviews and the overall sellers’
ratings,
4. To establish the relationship between sellers’ behaviour on Twitter, and seller disengagement,
including churn on Amazon.
This study uses Justice theory as the theoretical model, which is shown to be useful in identifying
antecedents to psychological processes and explaining individual’s reaction to conflict situation.
In the literature, three types of justice, namely distributive, procedural, and interactional justice,
are suggested as relevant to explain the psychological process of social interaction. In the context
of seller’s engagement on social media with the customers, or in the buyer-supplier relationship
performance, distribution justice can take the form of acknowledgement of inconvenience or issues
and taking remedial actions (Liu et al., 2012). When customers feel being heard and perceive fairness
in handling their complaints, customers tend to be satisfied, have a higher trust for the seller and the
company. Thus, they tend to spread positive word of mouth about it and give a positive rating to the
seller (Luca, 2017).
The rest of the article is structured as follows. Section 2 represents the contextual background
and literature review, followed by section 3, which is the theoretical background and hypothesis
development. Section 4 describes the research methodology and data collection. Analysis and findings
are discussed in section 5 after which the results are summarized. Discussion on the findings and
insights are in section 6. Finally, the article ends with a conclusion in section 7, and limitations and
further opportunities in this area of research in section 7.
BACKGROUND
This section is divided into the following three subsections: customer engagement through social
media; seller ratings and reviews on ecommerce platforms; customer intention of rating and reviews.
The first subsection highlights the usage of social media platforms by the sellers for customer
engagement. Next subsection introduces the sellers ratings and reviews on ecommerce platforms.
The last subsection defines the importance and intention of customers for giving ratings or writing
reviews for the sellers on ecommerce platforms.
Customer Engagement Through Social Media

Customer Engagement can be defined as the process of identifying, attracting, absorbing, motivating,
and interacting with customers for achieving business objectives (Harrigan et al., 2017; Pansari &
Kumar, 2017). Social media has become the pervasive channel for interactions and connections for
customer engagement (Grover and Kar, 2020). These interactions are not only limited to customers but
multiple stakeholders including organizations, stakeholders, and non-customers (Carlson et al., 2018;
Shawky et al., 2020). Since every brand’s effort on social media involves substantial investments, they
3
are left with a rudimentary question: Is it worth investing into social media and does this investment
translate into better customer acquisition, retention, and profits? However, most of the companies still
struggle in evaluating the impact of maintaining their personalized brand pages or these investments
on all the perceived benefits. They also fail to clearly see or calculate the return on investment (Lal
et al., 2019; Maecker et al., 2016). Consequently, companies following a multichannel customer
management approach (Neslin and Shankar, 2009; Neslin et al., 2006) along with investing heavily in
social media by establishing brand fan pages on which companies convey brand-related content (i.e.,
brand posts) that users can like, comment on, or share. These open, social media-enabled interactions
between a company and its customers help create a community that revolves around the brand and
fosters the brand relationship (McAlexander et al., 2002; Muniz and O’Guinn, 2001).
Past research has examined and suggested possible benefits of customer engagement through social
media. It includes a more profound and positive brand image (Quynh, 2019), brand value (France et
al., 2016), better revenue (Tu et al., 2018) and process improvements (Kushwaha et al., 2021). Recent
research has advocated the value of social media interactions for companies that propose conceptual
frameworks to manage brands (Gensler et al., 2013) and customers (Malthouse et al., 2013), develop
metrics, and assess the marketing potential in the social media context. However, there is a need for
a deeper understanding of the implications of customer engagement through social media, and its
utilization on the impact on company’s performance and customer satisfaction (Hennig-Thurau et al.,
2013), based on empirical evidence. Without knowing how and how much social media influences
customer relationships, companies struggle to assess the return on their investments in terms of
revenue and goodwill.
Seller Ratings and Reviews on E-Commerce Platforms

The innovation in technology and the proliferation of e-commerce companies in recent past has
revolutionized and transformed the retail industry worldwide, and are growing globally at an annual
rate of 23%. However, the real success of these platforms, like Amazon, lies in the performance of
the sellers who create great value for these platforms and the stakeholders by customer engagement,
strategic initiatives, and operational excellence (Mu & Zhang, 2021).
In India, Amazon alone has done sales over 2.1 billion dollars in the financial year 2021 showing
a growth rate of 49% from the previous year (ETech, 2022). With the huge opportunities that lie ahead
for e-commerce companies, there is also a great challenge of competition along with managing the
sellers on their platforms to get the competitive edge. There has been a lot of research that suggests
the benefits of positive brand image, ratings and feedback of sellers on the e-commerce platforms
on sales and revenue (Deshpande & Pendem, 2022; Mu & Zhang, 2021; Rodriguez et al., 2021).
It is evident from the above that seller ratings and positive reviews are as important as the product
reviews if not more. For ecommerce companies, thus, it is critical to get the best possible ratings and
reviews for the sellers. E-commerce companies, like Amazon, spent a lot of resources in managing
the sellers and make sure that it works in the best of their interest. The current study tries to fill the
literature gap and enriches it by looking at how Indian seller ratings and reviews on Amazon are
affected by the customer engagement done by these sellers on a social media platform, like Twitter.
Customer Intention of Rating and Reviews

Majority of literature focuses on the customer intention for purchase and impact of the reviews or
word of mouth on online purchase decisions (Liu & Tang, 2018; Meilatinova, 2021; Wang et al.,
2018). As customer ratings and reviews for sellers are critical for assessing customer engagement
and experiences, it is important to understand these reviews for the improvement of services (Kumar
et al., 2021). It is also crucial to understand the reasons and drivers behind the customer intention
to write a review or rate the sellers, which is an area, to best of our knowledge, where there is huge
gap in the literature. There can be a lot of drivers which are internal and external, which includes
a positive experience by the customers, customers receiving free samples, a negative experience,
4
incentives including rewards points and discounts, helping and guiding others, desire to help brand
improve their products, and validation from other shoppers (Team, 2022).
In the last decade, the electronic word of mouth marketing (e-WOM) has gained prominence
because of the disruptive growth in online users and customers becoming the contributors as well.
As e-WOM is always available in the future for the stakeholders to use and impacts the buying
decisions, it is critical for the industry practitioners to identify drivers which motivate the generation
of e-WOM in the form of reviews and ratings for the sellers (Jain et al., 2022; Rouibah et al., 2021).
While companies focus a lot on the above mentioned reasons for getting more reviews and ratings,
however, there still is a lot of scope to explore new areas in this context. The current study focuses
on one particular such aspect to fill this literature gap to find the relationship and impact of customer
engagement on social media platforms done by sellers, on the intention and score of reviews and
ratings by the customers for these sellers on e-commerce platforms.
THEORETICAL FRAMEWORK AND HYPOTHESES DEVELOPMENT
This study uses Justice Theory as the theoretical model, which is shown to be useful in identifying
antecedents to psychological processes and explaining individual’s reaction to conflict situation. In the
literature, three types of justice, namely distributive, procedural, and interactional justice, are suggested
as relevant to explain the psychological process of social interaction. In the context of seller’s engagement
on social media with the customers, distribution justice can take the form of acknowledgement of
complaints or issues and taking remedial actions (Ye and Tripathi, 2016). When customers feel being
heard and perceive fairness in handling their complaints, customers tend to be satisfied, have a higher
trust for the seller and the company. Thus, they tend to spread positive word of mouth about it and give
a positive rating to the seller and hence it leads to more seller engagement (Luca, 2017). The literature
indicates that the role of justice perception is paramount in customer satisfaction and in the intention for
word of mouth and reviews for the sellers. Distributive, Procedural and Interactional justice have positive
impact on customer satisfaction and thus increases the likelihood of word of mouth (Fu et al., 2015).
In the past, researches have typically emphasized distributive justice in order to predict customer
satisfaction (Lapidus & Pinkerton, 1995; Martínez-Tur et al., 2006). In purchase transactions,
customers invest inputs, such as money, and obtain outcomes such as service quality. The degree
to which the customer feels the transaction was unfair depends on how much it affected him/her
after the transaction. The customer expects reciprocity in terms of tangible matters, such as relating
quality to price. With respect to the distribution of outcomes, individuals seek to maximize gains
and minimize losses. Thus, distributive justice reflects a more outcome-oriented and instrumental
evaluation, given that a satisfactory outcome to input ratio is desired. From a service point of view
and as a seller on ecommerce platform, distributive justice refers to the satisfaction of a customer and
their experience in terms of the fairness with which they have been treated and provided the service
(Mccoll-Kennedy & Sparks, 2003).
The current study extends this reasoning in cross platform ecosystem based on the justice theory,
where the sellers who engage and address their customers’ concerns on social media platform, get
better ratings and more reviews on ecommerce platforms. Further, it explores the effect on the
sellers’ rating and degree of consensus on Amazon if the seller is active on social media platform
like Twitter, with a particular way of communicating and handling customers. It expands the horizon
of customer engagement by identifying the impact on seller disengagement or seller churn. Based
on the literature review, and to fulfil the objectives of this study, the following conceptual model,
shown in Figure 1, is proposed.
Reviews, Ratings, and Justice Theory

In the context of customer engagement and issue resolution, according to the available literature, justice
is perceived as the evaluation done by the customers of fairness of customer complaint handling process
5
Figure 1.
Conceptual relationship between platforms with sellers’ behaviour on social media
and it is divided into four categories: distributive justice, informational justice, interactional justice
and procedural justice (Badawi et al., 2021). There has been a lot of research and available literature
in this area of justice in complaints or issue handling (Ali et al., 2022; Karatepe, 2006; Muralidharan
et al., 2019). From a service point of view and as a seller on ecommerce platform, distributive justice
refers to the satisfaction of a customer and their experience in terms of the fairness with which they
have been treated and provided the service (Mccoll-Kennedy & Sparks, 2003).
The literature indicates that the role of justice perception is paramount in customer satisfaction and
in the intention for word of mouth and reviews for the sellers. Distributive, Procedural and Interactional
justice have positive impact on customer satisfaction and thus increases the likelihood of word of
mouth (Fu et al., 2015). In the context of ecommerce and sellers, a satisfied consumer is more likely
to respond to the seller in a positive way. It has been empirically shown in the past researches that
customer satisfaction has a significant impact on word of mouth in online business context (Manyanga
et al., 2022; Nam et al., 2020). The current study extends this reasoning in cross platform ecosystem
based on the justice theory, where the sellers who engage and address their customers’ concerns on
social media platform get better ratings and more reviews on ecommerce platforms.
Degree of Consensus in Reviews

Consensus in the online reviews has received some attention from the market research, primarily for
the objectives of identifying fake reviews and evaluating the genuineness of overall ratings. Jiménez
& Mendoza (2013) defined the consensus in online reviews as the perceived agreement among the
reviewers in their evaluation of the services of the sellers, that also includes the average ratings.
Benedicktus et al. (2010) evaluated the impact of consensus through the rating score on overall
purchase intention of the buyers. If a particular review, whether positive or negative, is in line with
the other reviews for the seller, it is perceived as accurate and genuine review which leads to changing
the perception and decision of the reader. Consensus in the reviews has been shown as a factor of
trustworthiness (Pavlou & Dimoka, 2006). If the consensus is high in reviews for a seller, it indicates
the genuineness of the reviews, which is a necessary pre-requisite for the present study (Munzel, 2016).
The hypotheses related to degree of consensus propose that there is a high degree of consensus
in the reviews for a particular seller on Amazon.com, and the degree of consensus in positive reviews
is directly proportional to the overall ratings of that seller.
6
H1a. There is a high degree of consensus in the reviews for the sellers on Amazon.com.
The hypothesis H1b looks at the consensus in the positive reviews in order to look at the positive
ratings.
H1b. Degree of consensus in positive reviews is directly proportional to the overall ratings of that seller.
Customer Orientation, Content Type, and Seller Ratings

Customer orientation can be defined as a seller’s ability to identify, acknowledge, and address the
needs of its target customers along with creating value for them (Mukerjee & Shaikh, 2018). There
has been a lot of work done for finding the impact of customer orientation through digital marketing
on buying intention, word of mouth, brand image, and perceiving high value of the product or the
service. Customer focus and customer orientation by the companies lead to positive word of mouth
and ratings (Izogo & Mpinganjira, 2020; Palazón et al., 2022). At the same time, researchers have
found out that the type of the content and timing of posting are equally important for user engagement
and influencing buyers’ behavior (Grover & Kar, 2018).
This study extends this phenomenon by exploring and identifying the impact of customer
orientation and type of content on the overall customer rating across platforms. The tweets and
conversation has been divided into marketing tweets and user centric communication tweets using
machine learning, referred to as content type. Customer orientation on Tweeter is captured by the
ratio of user centric tweets to marketing tweets. For a company to be customer oriented, they must
have more user centric tweets than marketing tweets. The hypotheses test the impact of the different
types of content on the overall ratings for sellers.
H2a. There is no statistical difference between the number of marketing tweets vs. customer focused
communication tweets posted by Amazon sellers on Twitter.
H2b. There is a positive impact of customer focused communication tweets on the overall ratings
of the sellers.
Seller Disengagement and Churn

Seller disengagement and churn is a major focus area for huge ecommerce platforms like Amazon
and even small ecommerce companies, because even 1% churn in sellers leads to 2% to 3% revenue
losses for these platforms (Tandon, 2017). At the same time, sellers must compete with other sellers
as tens of sellers are selling the same product on such platforms. Also, sellers have a big challenge
managing the fast-changing protocols, policies, and algorithms by these ecommerce platforms. For
example, many sellers were offering huge discounts or even free products in exchange of reviews
from customers, which was banned by Amazon in the year 2016. A couple of years ago, Amazon
stopped the option of personal identification for the sellers which created a huge challenge for sellers
as the sellers were heavily dependent on this for their marketing campaigns (Rodriguez et al., 2021).
Even though seller churn is a critical area for ecommerce ecosystem, however, most of the
academic research is focused on customer churn, and it is hard to find much work around sellers’
churn. The current study is a steppingstone in this area by exploring a small part of seller churn by
seeing if there is any relationship between sellers churning from Amazon in India and their customer
disengagement on social media platforms like Twitter. The hypothesis is that sellers who are churning
on Amazon are disengaged with their customers on Twitter.
H3. There is a strong positive correlation between seller disengagement on social media and their
churn on ecommerce platforms.
7
RESEARCH METHODOLOGY
The current study utilizes the textual data extracted from user generated content of two platforms,
i.e. Amazon and Twitter, to achieve the objectives. Thus, a combination of social media analytics,
natural language processing, machine learning and statistical testing has been used as the research
methodology for this study for theory building (Eisenstein, 2019; Emmert-Streib & Dehmer, 2019;
Kar and Dwivedi, 2020; Otter et al., 2021).
Out of the four types of social media analytics, this study uses descriptive and content analysis
to help in understanding basic descriptive statistics about the textual data elements and to convert the
textual qualitative data into more reliable and valid quantitative data (Krippendorff, 2004). Natural
Language Processing (NLP) is the use of machine learning algorithms for computers to understand
and use human languages for various applications. In the present study, various techniques including
term frequency–inverse document frequency, more commonly known as, TF-IDF for NLP have been
used to make sense of the textual data in the customer reviews and seller-customer communication
(Chowdhary, 2020; Trstenjak et al. 2014).
Machine learning has been used for two primary tasks; first, a supervised classification technique
has been used on labelled data for text categorization, and second, for sentiment, polarity, and
consensus analysis (Aphinyanaphongs et al. 2014). Text cleaning, preparation and quantification of
data was done before implementing the machine learning techniques. Partial Least Squares (PLS)
path analysis has been used to find out the relationship between latent variables, and the impact on
seller churn because of ratings on Amazon and the variables retrieved from Twitter, like number
of tweets, tweet category, polarity etc. (Henseler et al., 2016; Sohaib, 2021). For statistical tests,
multiple linear regression model has been used on the tabular data which is combined based on the
sellers, between Amazon and Twitter (Dospinescu et al., 2021). Finally, for seller churn, longitudinal
study was done to identify those sellers who have stopped doing business at Amazon India. Figure 2
provides the overall, high level methodology for the current study. This section is divided into three
subsections. The first part explains the data collection process, the second part provides the details
with data processing and implementation, and the last subsection is about the way data has been
analysed for the study.
Data Collection
The data for this study has been collected through the following four steps:
Step 1: Randomly search sellers on Amazon India with at-least 50, pre-covid period, reviews. This
was done to ensure there is no bias of the pandemic on the analysis, and second there should
be sufficient transactions for a seller to be part of this study. From this list of sellers, only those
sellers were considered who manage an active Twitter page with at least 100 Tweets. Only those
sellers who fulfil these two conditions were chosen for this study. It is a very time consuming
and critical task to identify the same seller on Twitter based on their presence on Amazon. 260
such sellers were identified and selected for the study from over 1000 initial sellers which were
identified from Amazon. These sellers are from multiple locations across India covering presence
across nation, and thus the study is not limited to a particular geography in the country. Similarly,
the category of products is not limited to a specific subset of product types.
Step 2: For these 260 sellers, all the reviews given by the customers along with the ratings were
captured for pre-covid pandemic time period. The data was stored in Excel files to be later read
into Python environment for processing. A total of 12950 reviews were collected.
Step 3: Using Tweeter API and Tweepy in Python, the tweets were extracted from Tweeter for these
260 sellers. A total of 256887 tweets were collected which included marketing and interaction
tweets.
8
Figure 2.
Overall high-level methodology for the present study
Example of marketing tweet: “1083306762227204096|2019-01-10 10:17:13|b’New Styles Added

and Huge Discounts Dropped; what are you waiting for?\n\nhttps://t.co/pawf3X44uW https://t.
co/QivY4Wpct1”
Example of interaction tweet: “@5TXLLc8WdjcbMvN sorry about this. will ask the customer
care to resolve.”
Step 4: After two years, the status of these 260 sellers was relooked at. Those sellers, who haven’t
got any reviews on Amazon for more than recent six months were tagged as inactive and churned.
44 out of 260 sellers were tagged as churned.
Data Pre-Processing and Data Preparation

The textual data collected, i.e. the tweets collected from Twitter and Seller reviews from Amazon,
suffers from numerous syntactical and semantic issues like spelling mistakes, punctuation, incorrect
grammar, usage of slang, abbreviations, and emojis. The objective is to remove this noise from the
data to ensure unbiased and accurate outcome. A series of steps have been followed during the pre-
processing of the texts. The URLs and hashtags were deleted from the tweets and comments during
the first data cleaning step. In the second step, all emoji were removed from the data, and all text
was converted in lowercase. In the third step, slang from the tweets and messages were removed
through data pre-processing. Lemmatization process has been used to convert multiple words with the
same root or meaning into single words to different inflected forms of an expression. The next stage
was to reveal and utilize the emotions present in the text. The feelings in the text were transformed;
for example, ‘:)’ is replaced with ‘Happy’ and ‘:(‘ is replaced with ‘sad.’ Further, stop words were
eradicated from texts as these word do not contribute in the analysis. All this was done using the
Spacy library in Python along with regular expressions and inbuilt Python features.
Data Analysis
Various types of analysis has been used on the two sets of textual data such as content analysis, and
social media analytics. Content analysis has been used to categorize the total number of 256887
tweets into communication and advertisement/ marketing tweets using machine learning models.
First, around 8000 tweets were manually labelled and then using machine learning, multiple models
were trained for splitting the tweets into two categories. 25% of the tweets were kept for testing using
k-fold validation. After trying multiple machine learning models, Logistic Regression resulted in the
best accuracy of 99.20%. The tweets were converted into a numerical matrix using TF-IDF, and then a
9
model using Logistic Regression on 135 features was created. Using the trained model, the remaining
tweets were divided into two buckets, i.e. marketing and communication tweets.
Amazon reviews were clustered into three buckets to check the consensus on the reviews. Three
clusters were formed based on the ratings scores; first for rating 1 and 2, second for rating 3 and third
for rating 4 and 5. The Silhouette coefficient is calculated to analyse the consensus amongst reviews
of different ratings (Tang et al., 2020). The high value of coefficient indicates that the clusters are
well formed and the degree of consensus is high. In the next cycle, all the common words and stop
words were removed to ensure there is no bias due to the commonly used words in the reviews. The
Silhouette coefficient is still above the recommended threshold and confirmed high consensus in
the reviews for the sellers.
Finally, Partial Least Squares (PLS) path analysis was used to find out the relationship between
latent variables, and the impact on seller churn because of ratings on Amazon and the variables
retrieved from Twitter, like number of tweets, tweet category, polarity etc. It has helped as confirmatory
statistical analysis using iterative ordinary least squares regression for PLS. Linear regression was
also used for conforming the impact of customer engagement by sellers on overall rating scores. For
the seller churn with longitudinal data, Mann Whitney U test was conducted to find the statistically
significant dissimilarity between the two groups of sellers, i.e. sellers who are active vs. the sellers
who have churned.
FINDINGS
This section has been divided into three subsections. The first subsection covers the analysis on
reviews and ratings to find the consensus between reviewers’ feedback. Next subsection presents the
impact of customer engagement through Twitter on the seller ratings on Amazon followed by the last
subsection which covers the seller disengagement and churn.
Degree of Consensus in Reviews

Literature indicates a direct impact of consensus in the reviews of customers on the purchase intention.
Customers are more likely to buy from a seller who has received feedback with a consensus in positive
reviews, as it increases the trustworthiness for the seller (Benedicktus et al., 2010). There is a strong
positive correlation between agreement in the positive reviews for a seller and the purchase intention
from the seller in ecommerce (Jiménez & Mendoza, 2013). Hypothesis H1a proposes that the reviews
on Amazon are genuine and have a high consensus.
H1a. There is a high degree of consensus in the reviews for every seller on Amazon.com.
After pre-processing step, i.e. applying regular expression, cleaning noise and stop words, the
clean textual data is available for every sellers’ reviews on Amazon along with the rating given against
each review. After applying TF-IDF on this data, k-mean clustering was applied with the rating score
as one of the variables. Three clusters were formed; one for rating 1 and 2, second for rating 3, and
third for rating 4 and 5. To find out whether the words used are same and there is a consensus, the
Silhouette coefficient was calculated for each cluster. With the score of 0.82, it is evident that the
clusters are very well representing similarity in the words (Shahapure & Nicholas, 2020). Further,
more stop words and commonly used words were identified, e.g. those words which were showing
emotions, to eliminate any possible bias towards high Silhouette coefficient, and the process was
repeated to calculate Silhouette coefficient on the new clusters. It still gives a silhouette coefficient
of 0.74. It suggests that there is a high consensus in the reviews (Larose & Larose, 2015). Thus, this
data can be used for further processing and analysis.
10
As an example, here are three reviews from different customers which are in one cluster, showing
high consensus:
1. “Packagingwas safe n secure.”

2. “Nice packing”
3. “Nice Product and Packaging.”
It would be interesting to see the relationship and impact of consensus on overall rating of the
seller. Hypothesis H1b proposes this relationship as follows:
H1b. Degree of consensus in positive reviews is directly proportional to the overall ratings of that seller
For a seller, the consensus score for positive reviews was calculated based on the Silhouette
coefficient for positive review clusters, i.e. for ratings 4 and 5 and compared against the overall rating
score of the seller. For the 260 sellers, the correlation coefficient is 0.84, which indicates a very strong
correlation between consensus in positive reviews and overall ratings of the seller. Thus, based on the
previous researches and literature as mentioned previously, it can be concluded that higher consensus
in positive reviews has higher overall ratings that eventually leads to positive buying intentions.
Customer Orientation, Content Type, and Seller Ratings

Literature suggests that customer orientation has a positive impact on the buying intentions and
is a critical attribute in e-commerce context (Hamidi & Moradi, 2017). With the help of dataset
collected from Twitter for these 260 sellers, this particular analysis has been started with the following
hypothesis to see if sellers focus more on marketing tweets on social media or they use it for customer
orientation and engagement.
H2a. There is no statistical difference between the number of marketing tweets vs. customer focused
communication tweets posted by Amazon sellers on Twitter.
At α = 0.05 (assumption), n number of observations at 260, and k = number of sample at 2,

degree of freedom is (1, 259). According to the decision rule, if calculated value of F is greater than
the Statistical table value of F, the null hypothesis can be rejected. With the current sample, the F
value at 5% level of significance is 6.28, thus the null hypothesis is rejected, and there is a significant
difference between marketing tweets and customer oriented tweets. With a higher number of marketing
tweets in the sample, it can be concluded that sellers use social media platforms, like Twitter, for
marketing significantly more than customer orientation. Figure 3, Word clouds show the common
words used in interaction vs. marketing tweets.
H2b. There is a positive impact of customer focused communication tweets on the overall ratings
of the sellers.
This is one of the most valued analysis of the current work. A linear regression was calculated
to find the impact of customer focused communication tweets, representing customer orientation,
on the ratings of the sellers. The combined data from Amazon and Twitter based on the sellers, with
average rating as the dependent variable, was used for the regression model. A significant regression
equation was found with these parameters: F = 57.87, p < 0.000, and R2 = 0.72. Higher the number
of customer focused tweets, better is the rating score of the sellers.
Here is the functional relationship of the model used:
11
Figure 3.
Word clouds for interaction vs. marketing tweets by sellers
Seller Rating Score on Amazon = f (Count of marketing tweets on Tweeter, Count of

communication tweets on Tweeter, Subjectivity of the Tweets, Polarity of the Tweets, Count of
Reviews, Business category of the sellers, Years of operations of sellers on Amazon, Place of business),
where the last three dependent variables have been used as control variables.
Seller Disengagement and Churn

Seller churn has a major loss for e-commerce giants like Amazon. While there can be several reasons
for seller churn, however, it would be interesting to see the impact and relationship between lack of
customer engagement and being inactive on social media, on seller churn on e-commerce platform.
H3. There is a strong positive correlation between seller disengagement on social media and their
The null hypothesis is that there is no significant relationship between the number of customer-
oriented tweets by a seller and the seller’s churn on ecommerce platform. For the 44 sellers who
churned, against the 216 sellers who continue to do well in terms of sales and reviews, Mann Whitney
U test is used to find if there is any statistically significant difference between the number of customer-
oriented tweets. The test is used because the variance is not same between the two groups and it is not
normally distributed, thus Mann Whitney U test is the best choice (Ghobadi et al., 2015; Zimmerman,
1987). At α = 0.05 (assumption), the statistics is equal to 112, with p < .000, thus the null hypothesis
is rejected. The average number of customer-oriented tweets by the churned seller during their active
phase is 119, whereas the average number of customer-oriented tweets by the other sellers in the
same period is 378, thus, it is evident that less customer orientation on social media indicates seller
Table 1 below summarizes the bootstrap algorithm result for path significance analysis. With the
coefficient scores and t statistics, it is evident that the inferential model supports the contextual model
introduced earlier through the paths and hypothesis testing discussed above (Henseler et al., 2016).
12
Table 1. Results of the bootstrap algorithm for path significance testing
DISCUSSION
Sellers and suppliers are an integral part of the ecommerce ecosystem, and the disengagement or
churn makes a huge negative impact on the profitability and competitive advantage for ecommerce
companies (Qin et al., 2021). There are several areas in which the ecommerce companies work to
ensure a healthy environment for the sellers. Customer engagement and orientation is an important
need in every business, however, it is even more important, essential and critical in ecommerce, as
the buying intentions are heavily driven by it.
The only factors that are available to a buyer are the way these sellers engage with them, and the
reviews given to these sellers by other buyers. While both of these factors are independently useful
and have been studied in detail, the present study has looked at the two factors together to see the
impact of one on another across different online platforms. The study explores the unexplored area
of the cross-platform relationship between the customer engagement by the sellers on social media
and performance of these sellers on ecommerce. The study also paves a path for exploring the area
of seller churn on ecommerce platforms further.
To enable the study, customer reviews and ratings from ecommerce website, Amazon were
taken for 260 sellers who had an active presence on social media (Twitter) as well. The tweets and
posts by these sellers on Twitter were also extracted. After using natural language processing and
machine learning, the action and focus of these sellers on social media were analysed for validating
the statistical models.
Hypothesis H1a and H1b are important to ensure the data is valid and useful for testing the most
important findings of this study. H1a and H1b reveal that the data is trustworthy because of high
consensus and the higher consensus in positive ratings also has a positive impact on higher ratings for
the sellers. Sellers and ecommerce platforms must continue their focus on the consensus in reviews
and must drive these based on their strategic and operational excellence.
H2a and H2b focused on the relationship between sellers’ behaviour on social media and their
ratings score and reviews on ecommerce platform. The tests suggest that the sellers are more focused
on marketing in general rather than engaging the customers and proving support to them. However,
it is clearly shown by the data, that more customer engagement and customer-oriented behaviour
the seller demonstrates, better reviews and ratings they receive. As this is a lot easier to manage in
comparison to exceeding the expectation of customers in their service, the sellers must focus more
on taking care of their customers on social media. Ecommerce companies must drive this behaviour
and can offer schemes for these sellers if not already doing it to promote this behaviour.
This is the first study which has tried to explore this cross platform impact of social media
engagement on seller disengagement and churn. Hypothesis H3, which is the highlight of the present
study and uses longitudinal data to find the impact of involvement and activities by the sellers through
social media engagement on the disengagement or churn on ecommerce. As mentioned above, seller
churn significantly impacts the profitability of ecommerce companies in a negative way. Ecommerce
13
companies can monitor the sellers’ social media behaviour, and take appropriate actions to avoid the
churn. The result suggests that lesser customer focus and involvement on social media by a seller
indicates the seller churn in near future.
Theoretical Contributions
The present study tries to address various gaps in this area and in this process, it makes important
theoretical contributions to both customer engagement and customer experience literature (Kumar et
al., 2022). This study has two primary theoretical contributions. First, the study extends the limited
research on the understanding of the categories of communication done by sellers on social media
and its impact on the reviews and ratings these sellers receive on ecommerce platforms. This study
is one of the first studies to consider social media performance of a seller as an important antecedent
of the ratings for the sellers on ecommerce platform.
Second, the study looks at the critical factor of seller churn on ecommerce platform by analyzing
the impact of the sellers’ social media engagement and disengagement. As this is a novel way of
looking at seller churn, the study will open a completely new area of research and will contribute
significantly to the literature in the future. No previous studies to the best of authors’ knowledge and
thorough search have empirically explored the relationship between sellers’ social media performance
and seller churn on ecommerce in an academic setting. However, literature and research show the
impact of sellers’ customer-orientation on the buying intentions, the positive impact of consensus
on the trustworthiness of the reviews, and the negative impact of seller churn on the profitability of
ecommerce companies (Chakravarty et al., 2014; Chang et al., 2003; Clemons, 2007; Cuong, 2020;
Goad & Jaramillo, 2014; Liu et al., 2020).
The theoretical lens for this study is Distribution Justice part of the Justice Theory that supports
the explanation of proposed relationship between social media engagement by the seller and the
factors like ratings, reviews and seller churn on ecommerce. The study would add to the theoretical
development by integrating cross platform relationship between the variables chosen on ecommerce
and social media platform to foster the seller engagement.
Implications for Practice

Literature suggests that users have four primary needs to use social media, viz. cognitive, social,
personal, and effective need. This study proposes better presence and more focus by the sellers on
customer engagement and communication posts on social media to fulfil the needs of the existing and
prospective customers. Hypothesis H2a, H2b and H3 indicates the benefit of the sellers’ appropriate
usage and engagement on Twitter for both sellers and ecommerce platforms. Thus, this study suggests
all the ecommerce sellers and ecommerce companies to consistently track, motivate and invest in
social media engagement with customer orientation for better sales, positive word of mouth, positive
reviews, and better ratings for the sellers. It also helps ecommerce companies to retain the sellers
and will reduce seller churn.
The companies, both sellers and ecommerce companies, can formulate strategies to look at the
timing, frequency, and impact of seasonality in the social media engagement. The study suggests that
both type of these firms must focus on (a) using social media to fulfil the needs of the customers, (b)
have high focus on customer orientation while posting on social media, and (c) track and formulate
strategies to have the maximum impact of social media engagement on the revenue, ratings, and seller
retention. As this study is specifically focused on ecommerce platforms like Amazon and the sellers
on these platforms, anyone from both these types of companies can use the findings of the study for
proactive strategies and actions to reduce seller churn, and to have effective customer engagement
for better ratings and profits.
14
CONCLUSION
This study tries to explore the cross-platform impact of customer-oriented seller engagement through
social media on the revies and ratings for the sellers on ecommerce platform, along with the indication
of seller churn. The focus of sellers on social media has been analyzed by looking at the content type.
The impact has been anaylzed from both user perspective and sellers’ perspective. This study introduces
the conceptual model and uses machine leaning, natural language processing and statistics to validate
the model and assumptions. It has been revealed that the firms, i.e., sellers and ecommerce platforms,
must gauge, focus, and encourage customer-oriented posts on social media and continuously fulfil the
needs of the users by addressing their concerns and communicating with them. While social media
platforms are very helpful marketing channels, however, the sellers must increase the communication
and interaction posts, and must try to have it more than the marketing messages. This is the first
study to explore this cross-platform relationship and unexplored impact of social media behavior by
sellers on their ratings and reviews on ecommerce platform. For ecommerce companies, the sellers’
social media behavior can be used as a key performance indicator to evaluate the performance, and
ecommerce companies should set thresholds and targets for the sellers to have a minimum number
of customer-oriented tweets or posts on social media platforms. Same methodology, which is used
in this study, can be used, and extended for further research and studies in this area.
Limitations and Future Work

Although this study makes a substantial contribution to the understanding of seller engagement,
disengagement, and churn because of sellers’ use of social media, the utilization of exclusively
seller-generated content severely restricts the reach of this study. Future studies should examine
and assess user-generated content to understand its function and significance. There is a possibility
that some sellers were not identified on Twitter out of 1000 sellers from Amazon as it was searched
manually on Twitter through the name of the sellers despite doing multiple rounds of search. Also,
a deeper contextual analysis can be done by bringing in the seller type in the study. Since Twitter
was the source of the data at a certain period for this study, it can be extended to include data from
other social media sites by quantifying the relationship between sellers’ roles on those sites and seller
ratings. This study relies only on social media data and the customer reviews. In future, customer
surveys can be administered to build and validate theoretical models.
ACKNOWLEDGMENT
We would like to thank Quanolytics LLP, India for supporting this research.
COMPETING INTERESTS
All authors of this article declare there are no competing interest.
FUNDING AGENCY
This research and publication were supported by Quanolytics LLP, India.
15
REFERENCES
Abbas, A., Zhou, Y., Deng, S., & Zhang, P. (2018). Text analytics to support sense-making in social media:
A language-action perspective. Management Information Systems Quarterly, 42(2), 427–464. doi:10.25300/
MISQ/2018/13239
Abbasi, R. A., Maqbool, O., Mushtaq, M., Aljohani, N. R., Daud, A., Alowibdi, J. S., & Shahzad, B. (2018). Saving
lives using social media: Analysis of the role of Twitter for personal blood donation requests and dissemination.
Telematics and Informatics, 35(4), 892–912. doi:10.1016/j.tele.2017.01.010
Agnihotri, R. (2020). Social Media, customer engagement, and sales organizations: A research agenda. Industrial
Marketing Management, 90, 291–299. doi:10.1016/j.indmarman.2020.07.017
Ali, M. A., Ting, D. H., Isha, A. S., Ahmad-Ur-Rehman, M., & Ali, S. (2022). Does service recovery matter?
relationships among perceived recovery justice, recovery satisfaction and customer affection and repurchase
intentions: The moderating role of gender. Journal of Asia Business Studies. doi:10.1108/JABS-02-2021-0060
Aphinyanaphongs, Y., Fu, L. D., Li, Z., Peskin, E. R., Efstathiadis, E., Aliferis, C. F., & Statnikov, A. (2014).
A comprehensive empirical comparison of modern supervised classification and feature selection methods for
text categorization. Journal of the Association for Information Science and Technology, 65(10), 1964–1987.
doi:10.1002/asi.23110
Appel, G., Grewal, L., Hadi, R., & Stephen, A. T. (2019). The future of social media in marketing. Journal of
the Academy of Marketing Science, 48(1), 79–95. doi:10.1007/s11747-019-00695-1 PMID:32431463
Badawi, B., Hartati, W., & Muslichah, I. (2021). Service Recovery Process: The Effects of Distributive and
Informational Justice on Satisfaction over Complaint Handling. The Journal of Asian Finance. Economics and
Business, 8(1), 375–383. doi:10.13106/JAFEB.2021.VOL8.NO1.375
Benedicktus, R. L., Brady, M. K., Darke, P. R., & Voorhees, C. M. (2010). Conveying trustworthiness to online
consumers: Reactions to consensus, physical store presence, brand familiarity, and generalized suspicion. Journal
of Retailing, 86(4), 322–335. doi:10.1016/j.jretai.2010.04.002
Brodie, R. J., Ilic, A., Juric, B., & Hollebeek, L. (2013). Consumer engagement in a Virtual Brand Community:
An exploratory analysis. Journal of Business Research, 66(1), 105–114. doi:10.1016/j.jbusres.2011.07.029
Carlson, J., Rahman, M., Voola, R., & De Vries, N. (2018). Customer engagement behaviours in social media:
Capturing innovation opportunities. Journal of Services Marketing, 32(1), 83–94. doi:10.1108/JSM-02-2017-0059
Chakravarty, A., Kumar, A., & Grewal, R. (2014). Customer Orientation Structure for Internet-Based Business-
to-Business Platform Firms. Journal of Marketing, 78(5), 1–23. doi:10.1509/jm.12.0442
Chang, K. C., Jackson, J., & Grover, V. (2003). E-commerce and corporate strategy: An executive perspective.
Information & Management, 40(7), 663–675. doi:10.1016/S0378-7206(02)00095-2
Chen, P.-Y., Dhanasobhon, S., & Smith, M. D. (2008). All reviews are not created equal: The disaggregate impact
of reviews and reviewers at Amazon.com. SSRN Electronic Journal. 10.2139/ssrn.918083
Chowdhary, K. R. (2020). Natural language processing. Fundamentals of Artificial Intelligence, 603–649.
10.1007/978-81-322-3972-7_19
Clemons, E. K. (2007). An Empirical Investigation of Third-Party Seller Rating Systems in E-Commerce:
The Case of buySAFE. Journal of Management Information Systems, 24(2), 43–71. doi:10.2753/MIS0742-
1222240203
Connell, C., Marciniak, R., Carey, L. I., & McColl, J. (2019). Customer engagement with websites: A transactional
retail perspective. European Journal of Marketing, 53(9), 1882–1904. doi:10.1108/EJM-10-2017-0649
Cuong, P. H. (2020). Review Consensus Effects on e-Wom and Consumer Goods e-Purchase Satisfaction: Evidence
from Vietnam. Journal of Advanced Research in Law and Economics, 11(2), 321. doi:10.14505/jarle.v11.2(48).05
de Oliveira Santini, F., Ladeira, W. J., Pinto, D. C., Herter, M. M., Sampaio, C. H., & Babin, B. J. (2020).
Customer engagement in Social Media: A Framework and meta-analysis. Journal of the Academy of Marketing
Science, 48(6), 1211–1228. doi:10.1007/s11747-020-00731-5
16
Deshpande, V., & Pendem, P. K. (2022). Logistics performance, ratings, and its impact on customer
purchasing behavior and sales in e-commerce platforms. Manufacturing & Service Operations Management,
msom.2021.1045. doi:10.1287/msom.2021.1045
Dewnarain, S., Ramkissoon, H., & Mavondo, F. (2021). Social Customer Relationship Management: A customer
perspective. Journal of Hospitality Marketing & Management, 30(6), 673–698. doi:10.1080/19368623.2021.
1884162
Dospinescu, O., Dospinescu, N., & Bostan, I. (2021). Determinants of e-commerce satisfaction: A comparative
study between Romania and Moldova. Kybernetes, 51(13), 1–17. doi:10.1108/K-03-2021-0197
Eisenstein, J. (2019). Introduction to natural language processing. The MIT Press.
Emmert-Streib, F., & Dehmer, M. (2019). Understanding statistical hypothesis testing: The logic of statistical
inference. Machine Learning and Knowledge Extraction, 1(3), 945–961. doi:10.3390/make1030054
ETech (Ed.). (2022). Amazon India marketplace clocks 49% revenue growth in FY21. The Economic Times.
https://economictimes.indiatimes.com/tech/technology/amazon-india-marketplace-clocks-49-revenue-growth-
in-fy21/articleshow/88700485.cms
France, C., Merrilees, B., & Miller, D. (2016). An integrated model of customer-brand engagement: Drivers and
consequences. Journal of Brand Management, 23(2), 119–136. doi:10.1057/bm.2016.4
Fu, J.-R., Ju, P.-H., & Hsu, C.-W. (2015). Understanding why consumers engage in electronic word-of-mouth
communication: Perspectives from theory of planned behavior and justice theory. Electronic Commerce Research
and Applications, 14(6), 616–630. doi:10.1016/j.elerap.2015.09.003
Gensler, S., Völckner, F., Liu-Thompkins, Y., & Wiertz, C. (2013). Managing brands in the Social Media
Environment. Journal of Interactive Marketing, 27(4), 242–256. doi:10.1016/j.intmar.2013.09.004
Ghobadi, S., Campbell, J., & Clegg, S. (2015). Pair programming teams and high-quality knowledge sharing: A
comparative study of coopetitive reward structures. Information Systems Frontiers, 19(2), 397–409. doi:10.1007/
s10796-015-9603-0
Giachanou, A., & Crestani, F. (2016). Like It or Not. ACM Computing Surveys, 49(2), 1–41. doi:10.1145/2938640
Gligor, D., Bozkurt, S., & Russo, I. (2019). Achieving customer engagement with social media: A qualitative
comparative analysis approach. Journal of Business Research, 101, 59–69. doi:10.1016/j.jbusres.2019.04.006
Goad, E. A., & Jaramillo, F. (2014). The good, the bad and the effective: A meta-analytic examination of selling
orientation and customer orientation on sales performance. Journal of Personal Selling & Sales Management,
34(4), 285–301. doi:10.1080/08853134.2014.899471
Grover, P., & Kar, A. K. (2020). User engagement for mobile payment service providers–introducing the
social media engagement model. Journal of Retailing and Consumer Services, 53, 101718. doi:10.1016/j.
jretconser.2018.12.002
Gupta, S., & Ramachandran, D. (2021). Emerging market retail: Transitioning from a product-centric to a
customer-centric approach. Journal of Retailing, 97(4), 597–620. doi:10.1016/j.jretai.2021.01.008
Harrigan, P., Evers, U., Miles, M., & Daly, T. (2017). Customer engagement with tourismsocial media brands.
Tourism Management, 59, 597–609. doi:10.1016/j.tourman.2016.09.015
Hennig-Thurau, T., Hofacker, C. F., & Bloching, B. (2013). Marketing the pinball way: Understanding how
social media change the generation of value for consumers and companies. Journal of Interactive Marketing,
27(4), 237–241. doi:10.1016/j.intmar.2013.09.005
Henseler, J., Hubona, G., & Ray, P. A. (2016). Using PLS path modeling in new technology research: Updated
guidelines. Industrial Management & Data Systems, 116(1), 2–20. doi:10.1108/IMDS-09-2015-0382
Hollebeek, L. D., Sharma, T. G., Pandey, R., Sanyal, P., & Clark, M. K. (2021). Fifteen Years of customer
engagement research: A Bibliometric and network analysis. Journal of Product and Brand Management, 31(2),
293–309. doi:10.1108/JPBM-01-2021-3301
17
Izogo, E. E., & Mpinganjira, M. (2020). Behavioral consequences of customer inspiration: The role of social
media inspirational content and cultural orientation. Journal of Research in Interactive Marketing, 14(4), 431–459.
doi:10.1108/JRIM-09-2019-0145
Jain, M., Dixit, S., & Shukla, A. (2022). Role of E-service quality, Brand Commitment and E-WOM Trust on
E-wom intentions of millennials. International Review of Retail, Distribution and Consumer Research, 33(1),
23–43. doi:10.1080/09593969.2022.2070860
Jiménez, F. R., & Mendoza, N. A. (2013). Too popular to ignore: The influence of online reviews on purchase
intentions of Search and Experience Products. Journal of Interactive Marketing, 27(3), 226–235. doi:10.1016/j.
intmar.2013.04.004
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! the challenges and opportunities of social
media. Business Horizons, 53(1), 59–68. doi:10.1016/j.bushor.2009.09.003
Kar, A. K., & Dwivedi, Y. K. (2020). Theory building with big data-driven research–Moving away from the
“What” towards the “Why”. International Journal of Information Management, 54, 102205. doi:10.1016/j.
ijinfomgt.2020.102205
Karatepe, O. M. (2006). Customer complaints and organizational responses: The effects of complainants’
perceptions of justice on satisfaction and Loyalty. International Journal of Hospitality Management, 25(1),
69–90. doi:10.1016/j.ijhm.2004.12.008
Kaur, H., Paruthi, M., Islam, J. U., & Hollebeek, L. D. (2020). The role of Brand Community Identification
and reward on consumer brand engagement and brand loyalty in Virtual Brand Communities. Telematics and
Informatics, 46, 101321. doi:10.1016/j.tele.2019.101321
Krippendorff, K. (2004). Reliability in content analysis: Some common misconceptions and recommendations.
Human Communication Research, 30(3), 411–433. doi:10.1111/j.1468-2958.2004.tb00738.x
Kumar, P., Hollebeek, L. D., Kar, A. K., & Kukk, J. (2022). Charting the intellectual structure of customer
experience research. Marketing Intelligence & Planning, 41(1), 31–47. doi:10.1108/MIP-05-2022-0185
Kumar, S., Kar, A. K., & Ilavarasan, P. V. (2021). Applications of text mining in services management: A
systematic literature review. International Journal of Information Management Data Insights, 1(1), 100008.
doi:10.1016/j.jjimei.2021.100008
Kushwaha, A. K., Kumar, P., & Kar, A. K. (2021). What impacts customer experience for B2B enterprises on
using AI-enabled chatbots? Insights from Big data analytics. Industrial Marketing Management, 98, 207–221.
doi:10.1016/j.indmarman.2021.08.011
Lal, B., Ismagilova, E., Dwivedi, Y. K., & Kwayu, S. (2019). Return on investment in Social Media Marketing:
Literature Review and suggestions for future research. Digital and Social Media Marketing, 3–17. 10.1007/978-
3-030-24374-6_1
Lapidus, R. S., & Pinkerton, L. (1995). Customer complaint situations: An equity theory perspective. Psychology
and Marketing, 12(2), 105–122. doi:10.1002/mar.4220120203
Laroche, M., Habibi, M. R., & Richard, M.-O. (2013). To be or not to be in social media: How brand loyalty
is affected by social media? International Journal of Information Management, 33(1), 76–82. doi:10.1016/j.
ijinfomgt.2012.07.003
Larose, D. T., & Larose, C. D. (2015). The Silhouette Method. In Data Mining and Predictive Analytics,
583–589. Wiley.
Liu, Y., Chen, D. Q., & Gao, W. (2020). How does customer orientation (in)congruence affect B2B electronic
commerce platform firms’ performance? Industrial Marketing Management, 87, 18–30. doi:10.1016/j.
indmarman.2020.02.027
Liu, Y., Huang, Y., Luo, Y., & Zhao, Y. (2012). How does justice matter in achieving buyer-supplier relationship
performance? Journal of Operations Management, 30(5), 355–367. doi:10.1016/j.jom.2012.03.003
Liu, Y., & Tang, X. (2018). The effects of online trust-building mechanisms on trust and repurchase intentions.
Information Technology & People, 31(3), 666–687. doi:10.1108/ITP-10-2016-0242
18
Luca, M. (2017). Designing online marketplaces: Trust and reputation mechanisms. Innovation Policy and the
Economy, 17, 77–93. doi:10.1086/688845
Maecker, O., Barrot, C., & Becker, J. U. (2016). The effect of social media interactions on customer relationship
management. Business Research, 9(1), 133–155. doi:10.1007/s40685-016-0027-6
Malthouse, E. C., Haenlein, M., Skiera, B., Wege, E., & Zhang, M. (2013). Managing customer relationships
in the social media era: Introducing the social CRM House. Journal of Interactive Marketing, 27(4), 270–280.
doi:10.1016/j.intmar.2013.09.008
Manyanga, W., Makanyeza, C., & Muranda, Z. (2022). The effect of customer experience, customer satisfaction
and word of mouth intention on customer loyalty: The moderating role of Consumer Demographics. Cogent
Business & Management, 9(1), 2082015. doi:10.1080/23311975.2022.2082015
Martínez-Tur, V., Peiró, J. M., Ramos, J., & Moliner, C. (2006). Justice perceptions as predictors of customer
satisfaction: The impact of distributive, procedural, and Interactional Justice1. Journal of Applied Social
Psychology, 36(1), 100–119. doi:10.1111/j.0021-9029.2006.00005.x
McAlexander, J. H., Schouten, J. W., & Koenig, H. F. (2002). Building Brand Community. Journal of Marketing,
66(1), 38–54. doi:10.1509/jmkg.66.1.38.18451
Mccoll-Kennedy, J. R., & Sparks, B. A. (2003). Application of fairness theory to service failures and service
recovery. Journal of Service Research, 5(3), 251–266. doi:10.1177/1094670502238918
Meilatinova, N. (2021). Social Commerce: Factors affecting customer repurchase and word-of-mouth intentions.
International Journal of Information Management, 57, 102300. doi:10.1016/j.ijinfomgt.2020.102300
Mosteller, J., & Mathwick, C. (2016). Consumers helping consumers; the role of psychological need fulfillment
in an online reviewer community. Let’s Get Engaged! Crossing the Threshold of Marketing’s Engagement Era,
377–378, 377–378. doi:10.1007/978-3-319-11815-4_106
Mu, J., & Zhang, J. Z. (2021). Seller marketing capability, Brand Reputation, and consumer journeys on
e-commerce platforms. Journal of the Academy of Marketing Science, 49(5), 994–1020. doi:10.1007/s11747-
021-00773-3
Mukerjee, K., & Shaikh, A. (2019). Impact of customer orientation on word-of-mouth and cross-buying.
Marketing Intelligence & Planning, 37(1), 97–110. doi:10.1108/MIP-01-2018-0030
Muniz, A. M. Jr, & O’Guinn, T. C. (2001). Brand Community. The Journal of Consumer Research, 27(4),
412–432. https://doi.orz/10.1086/319618. doi:10.1086/319618
Munzel, A. (2016). Assisting consumers in detecting fake reviews: The role of Identity Information Disclosure
and consensus. Journal of Retailing and Consumer Services, 32, 96–108. doi:10.1016/j.jretconser.2016.06.002
Muralidharan, E., Guo, W., Fazel, H., & Wei, W. (2019). Organizational response to goods failure complaints:
The role of culture on perceptions of interactional justice and customer satisfaction. Global Business Review,
22(6), 1327–1344. doi:10.1177/0972150919861783
Nam, K., Baker, J., Ahmad, N., & Goo, J. (2020). Determinants of writing positive and negative electronic
word-of-mouth: Empirical evidence for two types of expectation confirmation. Decision Support Systems, 129,
113168. doi:10.1016/j.dss.2019.113168
Neslin, S. A., Grewal, D., Leghorn, R., Shankar, V., Teerling, M. L., Thomas, J. S., & Verhoef, P. C. (2006).
Challenges and opportunities in Multichannel Customer Management. Journal of Service Research, 9(2), 95–112.
doi:10.1177/1094670506293559
Neslin, S. A., & Shankar, V. (2009). Key issues in Multichannel Customer Management: Current Knowledge
and Future Directions. Journal of Interactive Marketing, 23(1), 70–81. doi:10.1016/j.intmar.2008.10.005
Otter, D. W., Medina, J. R., & Kalita, J. K. (2021). A survey of the usages of deep learning for natural language
processing. IEEE Transactions on Neural Networks and Learning Systems, 32(2), 604–624. doi:10.1109/
TNNLS.2020.2979670 PMID:32324570
19
Palazón, M., López, M., Sicilia, M., & López, I. (2022). The customer journey: A proposal of indicators to
evaluate integration and customer orientation. Journal of Marketing Communications, 28(5), 1–32. doi:10.10
80/13527266.2022.2051584
Pansari, A., & Kumar, V. (2017). Customer engagement marketing. Customer Engagement Marketing, 1–27.
10.1007/978-3-319-61985-9_1
Pavlou, P. A., & Dimoka, A. (2006). The nature and role of feedback text comments in online marketplaces:
Implications for Trust Building, price premiums, and seller differentiation. Information Systems Research, 17(4),
392–414. doi:10.1287/isre.1060.0106
Qin, L., Qu, Q., Zhang, L., & Wu, H. (2021). Platform trust in C2C e-commerce platform: The sellers’ cultural
perspective. Information Technology Management. doi:10.1007/s10799-021-00349-1
Quynh, N. H. (2019). The moderating influence of brand image on the relationship between customer engagement
and customer loyalty. Indian Journal of Marketing, 49(9), 42. doi:10.17010/ijom/2019/v49/i9/146939
Rodriguez, J., Watson, R. T., & Piccoli, G. (2021). Seller labs: Democratizing E-merchants competitiveness.
Communications of the Association for Information Systems, 49(1), 800–815. doi:10.17705/1CAIS.04941
Rosado-Pinto, F., & Loureiro, S. M. (2020). The growing complexity of customer engagement: A systematic
review. EuroMed Journal of Business, 15(2), 167–203. doi:10.1108/EMJB-10-2019-0126
Rouibah, K., Al-Qirim, N., Hwang, Y., & Pouri, S. G. (2021). The determinants of eWoM in Social Commerce.
Journal of Global Information Management, 29(3), 75–102. doi:10.4018/JGIM.2021050104
Rouibah, K., Al-Qirim, N., Hwang, Y., & Pouri, S. G. (2021). The determinants of eWoM in Social Commerce.
Shahapure, K. R., & Nicholas, C. (2020). Cluster Quality Analysis Using Silhouette Score. 2020 IEEE
7th International Conference on Data Science and Advanced Analytics (DSAA). IEEE. doi:10.1109/
DSAA49011.2020.00096
Shawky, S., Kubacki, K., Dietrich, T., & Weaven, S. (2020). A Dynamic Framework for managing customer
engagement on Social Media. Journal of Business Research, 121, 567–577. doi:10.1016/j.jbusres.2020.03.030
So, K. K., Li, X. R., & Kim, H. (2019). A decade of Customer Engagement Research in hospitality and tourism:
A systematic review and Research Agenda. Journal of Hospitality & Tourism Research (Washington, D.C.),
44(2), 178–200. doi:10.1177/1096348019895562
Sohaib, O. (2021). Social Networking Services and Social Trust in Social Commerce. Journal of Global
Information Management, 29(2), 23–44. doi:10.4018/JGIM.2021030102
Tandon, A. (2017, September 25). Seller churn in ecommerce: The silent killer of your business. Praxisga.
https://www.praxisga.com/insights/technology-and-internet/seller-churn-in-ecommerce-the-silent-killer-of-
your-business
Tang, M., Liao, H., Xu, J., Streimikiene, D., & Zheng, X. (2020). Adaptive consensus reaching process with
hybrid strategies for large-scale group decision making. European Journal of Operational Research, 282(3),
957–971. doi:10.1016/j.ejor.2019.10.006
Team, P. R. (2022). Why do consumers provide product ratings & reviews? PowerReviews. https://www.
powerreviews.com/insights/why-do-consumers-provide-product-ratings-reviews/
Thakur, R. (2018). Customer engagement and online reviews. Journal of Retailing and Consumer Services, 41,
48–59. doi:10.1016/j.jretconser.2017.11.002
Trstenjak, B., Mikac, S., & Donko, D. (2014). KNN with TF-IDF based framework for text categorization.
Procedia Engineering, 69, 1356–1364. doi:10.1016/j.proeng.2014.03.129
Tu, Y. J., Neuhofer, B., & Viglia, G. (2018). When co-creation pays: Stimulating engagement to increase
revenues. International Journal of Contemporary Hospitality Management, 30(4), 2093–2111. doi:10.1108/
IJCHM-09-2016-0494
20
Wang, H., Du, R., & Olsen, T. (2018). Feedback mechanisms and consumer satisfaction, trust and repurchase
intention in online retail. Information Systems Management, 35(3), 201–219. doi:10.1080/10580530.2018.14
77301
Xiao, L., & Li, Y. (2019). Examining the effect of positive online reviews on consumers’ decision making.
Ye, H., & Tripathi, A. (2016). Using social media to manage customer complaints: A preliminary study. 2016
49th Hawaii International Conference on System Sciences (HICSS). doi:10.1109/HICSS.2016.478
Yen, C.-H., Teng, H.-Y., & Tzeng, J.-C. (2020). Innovativeness and customer value co-creation behaviors:
Mediating role of customer engagement. International Journal of Hospitality Management, 88, 102514.
doi:10.1016/j.ijhm.2020.102514
Zimmerman, D. W. (1987). Comparative Power of StudentTTest and Mann-WhitneyUTest for Unequal Sample
Sizes and Variances. Journal of Experimental Education, 55(3), 171–174. doi:10.1080/00220973.1987.10806451
Anuj is the founder and CEO of Quanolytics, which is a fast growing data science consulting company, headquartered
in India. Before Quanolytics, he has worked for companies like Adobe, Sapient, HCL technologies etc. in leadership
roles in data science and strategy. He also teaches at premier universities and institutions as a visiting and part
time professor in India, Europe and USA. Anuj has double post-graduation and is an IIM-Calcutta alumnus. He’s
doing his research at IIT Delhi.
Arpan Kar is Chair Professor in Data Science and Decision Science in Indian Institute of Technology Delhi, India.
Within IIT Delhi, he shares a joint appointment in the Department of Management Studies and Yardi School of
Artificial Intelligence. Administratively he is the Chair of Corporate Relations and is a member of Board of Academic
Programme and Institute of Eminence Committee, among others. He also supports multiple ministries and other
universities also in different advisory boards/roles. He works in the space of data science, digital platforms and
apps, social media and ICT-based public policy. He has authored over 175 peer reviewed articles and edited 10
books and has over 8000 citations with an H-Index of 47. Over 60 publications are in ABDC A, ABS 3 or WoS
Q1 journals including 12 publications in ABDC A* journals. He is the Editor in Chief of International Journal of
Information Management Data Insights, published by Elsevier, which is ranked as a Q1 journal in Scopus. He has
been a Guest Editor for ABDC A*/A journals like Decision Support Systems, Industrial Marketing Management,
International Journal of Information Management, Information Systems Frontiers, Australasian Journal of Information
Systems, etc. He has recieved over 20 prestigious awards for his research.
Satpathy has over 16 years of mixed experience in corporate and academics. Pursuing her passion in teaching, she
is competent to bring together wide array of information into the classroom that is effective at driving performance.
She holds a PhD in Management and Masters in Business Administration. Se has presented papers in national
and international conferences and published research in various journals of high repute.
21

Cross Platform Analysis of Seller Performance and Churn For Ecommerce Using Artificial Intelligence

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cross Platform Analysis of Seller Performance and Churn For Ecommerce Using Artificial Intelligence

Uploaded by

Copyright:

Available Formats

Journal of Global Information Management

Cross-Platform Analysis of Seller

Arpan Kumar Kar, Indian Institute of Technology, Delhi, India*

DOI: 10.4018/JGIM.322439 *Corresponding Author

Customer Engagement Through Social Media

Seller Ratings and Reviews on E-Commerce Platforms

Customer Intention of Rating and Reviews

THEORETICAL FRAMEWORK AND HYPOTHESES DEVELOPMENT

Reviews, Ratings, and Justice Theory

Degree of Consensus in Reviews

Customer Orientation, Content Type, and Seller Ratings

Seller Disengagement and Churn

Example of marketing tweet: “1083306762227204096|2019-01-10 10:17:13|b’New Styles Added

Data Pre-Processing and Data Preparation

Degree of Consensus in Reviews

1. “Packagingwas safe n secure.”

Customer Orientation, Content Type, and Seller Ratings

At α = 0.05 (assumption), n number of observations at 260, and k = number of sample at 2,

Seller Rating Score on Amazon = f (Count of marketing tweets on Tweeter, Count of

Seller Disengagement and Churn

Table 1. Results of the bootstrap algorithm for path significance testing

Implications for Practice

Limitations and Future Work

All authors of this article declare there are no competing interest.

This research and publication were supported by Quanolytics LLP, India.

You might also like