Professional Documents
Culture Documents
13 INSIGHT - 2023 - Rosser
13 INSIGHT - 2023 - Rosser
FEATURE
SPECIAL
Agile Reviews
JUNE 2O23
VOLUME 26/ ISSUE 2
Larri Rosser, larri.rosser@rtx.com
Copyright © 2023 by Larri Rosser. Published and used by INCOSE with permission.
ABSTRACT
Current technical oversight approaches used for government programs (for example, stage-gate reviews) are not agile — their
expectations are not aligned with agile development cadences, and they are not adequately responsive to continuous unpredictable
change. This article explores ways to provide insight and responsive forward looking actionable guidance for agile projects in the
context of government and defense programs. It proposes a general oversight approach that produces minimal drag and disruption
and keeps pace with agile product development.
F
INTRODUCTION
or at least a decade, the US gov- ile Assessment Guide: Best Practices for Agile describes agile systems engineering as
ernment has been exploring the Adoption and Implementation (US Govern- “effective system engineering in the face
use of agile methods on defense ment Accountability Office 2020), Design of uncontrolled change” and identifies “an
and other federal acquisitions in and Acquisition of Software for Defense asynchronous/concurrent life cycle model
order to reduce costs while more effectively Systems (Defense Science Board 2018), and framework” as one of the hallmarks of agile
supporting missions. Compared to com- The TechFAR Handbook (TechFAR Hub systems engineering. (Schindel and Dove
mercial application software development 2022). In addition, they have worked with 2019) This approach to systems engineering
efforts from which agile practices initially industry to provide instructions for apply- does not fit comfortably with current
emerged, these acquisitions tend to be com- ing specific agile practices to government prevalent gate review practices, which tend
plex, completion-based efforts to realize programs, including earned value (National to assume that product realization activities
systems, not just software. These challenges Defense Industrial Association 2019) and occur sequentially and that reviews evaluate
encourage evolution in agile execution, agile contracting (Office of Management final outputs of one phase of engineering
leading to the definition of numerous and Budget 2012). activities that has recently completed to
approaches to scaling and expanding agile In the area of technical oversight, determine if the project can proceed to
practices. Descriptions of these advanced however, the standard recommended by the next (sequential) phase (Lapham et
agile approaches can be found in commer- the Department of Defense (Office of the al. 2016 p. 19). Mismatches between the
cial scaling frameworks such as scaled agile Under Secretary of Defense for Acquisi- expectations of agile engineering and
framework (SAFe) (Scaled Agile Frame- tion, Technology, and Logistics 2017) is traditional gate reviews can lead to several
work 2021), disciplined agile (DA) (Pro- ISO/IEC/IEEE 15288.2, IEEE Standard for undesirable effects, including incorrect
gram Management Institute 2020), and en- Technical Reviews and Audits on Defense evaluation of solution maturity, unhelpful
terprise unified process (EUP) (Enterprise Programs (IEEE Computer Society 2014). or unactionable findings, and inaccurate
Unified Process 2020) as well as execution Although tailoring of criteria is permit- assessment of risk.
case studies from Lockheed Martin (Dove, ted and even encouraged, tailoring this This observation does not imply that
Schindel, and Garlington 2018), Northrup standard appropriately for agile programs either agile product development or
Grumman (Dove and Schindel 2017), and presents challenges, as the standard is traditional stage-gate reviews are wrong,
Rockwell Collins (Dove, Schindel, and organized around the phases in the tradi- but rather to point out that, as currently
Hartley 2017), among other sources. tional waterfall lifecycle. implemented, they are ineffective together,
During this time, various agencies of and to propose alternatives that preserve
the US government have studied agile AGILE DEVELOPMENT VS. STAGE-GATE the intended value of both concepts.
execution and released several documents REVIEWS At a conceptual level, the intent of gate
discussing the application of agile practices Agile Systems Engineering Life Cycle reviews is to assess progress and determine
to government contracts. These include Ag- Model for Mixed Discipline Engineering if the program is able to proceed to their
53
21564868, 2023, 2, Downloaded from https://incose.onlinelibrary.wiley.com/doi/10.1002/inst.12446 by libffg rrtyu - Beijing Normal University , Wiley Online Library on [09/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Agile values application to gate reviews ploy agile methods, it’s necessary to focus
FEATURE
SPECIAL
54
21564868, 2023, 2, Downloaded from https://incose.onlinelibrary.wiley.com/doi/10.1002/inst.12446 by libffg rrtyu - Beijing Normal University , Wiley Online Library on [09/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
program execution model that is employed
FEATURE
SPECIAL
Inception Realization Transition should be a part of any agile review. This is
particularly important since most of these
“Deployment”
“Delivery” practices are fairly new and potentially
Program Startup unfamiliar to team members or at odds
“Increment Zero” or “Construction” “Sell-Off”
“Sprint Zero” “Development” “Initial Operating with established expectations or adjacent
“Define/Design/Build/Integrate/Test” Capability” processes.
Activities needed to “Transition to Production”
enable the program to Activities needed to create valuable Review Content
JUNE 2O23
VOLUME 26/ ISSUE 2
create value in an capabilities in short cycles of learning Movement of The content of agile reviews should in-
agile manner completed capabilities
to the next step clude artifacts that contribute to or indicate
the health of program execution. Some
Figure 1. Agile phases and boundaries items to evaluate are described below.
Working plans – This may include the
development. This phase includes both authorization to operate (ATO). There may master phasing schedule and integrated
traditional program startup activities and be multiple transition events in the context master schedule but should also include
those engineering and technical activities of a single program if multiple phases or any agile planning artifacts such as back-
required to make the program ready to effectivities have been contracted. logs and roadmaps.
commence agile development. Systems Integrated test plan and results – most
Engineering for Software Intensive Projects Review Focus agile practices encourage early and frequent
Using Agile Methods refers to these activities All reviews should focus on concerns integration and test, and both the approach
as “pre-planning” activities and describes that have high impact on the project or and results are valuable objects of review.
them in some detail (Rosser et al. 2014 p. product being reviewed, and agile project Traceability is exceptionally important
6). From the perspective of review plan- reviews are no different in this regard. on agile projects where changes occur
ning, the end of this period provides a good However, the specific elements that have frequently and anywhere in the chain from
point to evaluate the program’s readiness to high impact may not be the same as in re- customer requirements to operational sys-
set the agile realization process in motion. views for other project types. The following tems. Bi-directional traceability from end to
Plans, requirements, staffing, tooling, and paragraphs highlight some areas worth the end of realization should be evaluated.
artifacts can all be evaluated, not to deter- evaluating on agile projects. Tools and automation is a force multi-
mine if they are finished, but rather if they Working capabilities are the primary plier for agile projects. Effective integrated
are adequate to start. yardstick of progress on agile projects. A toolsets can increase both quality and
Realization is the phase in which ca- critical part of an agile review should be productivity when properly configured and
pabilities are designed, implemented, and focused on what’s working. Newly work- employed.
tested using agile methods. During this ing capabilities should be assessed for Baseline maturation and quality – as
time, work is done in small batches, often fulfillment of customer needs, previously the central value output of agile product
repeated many times. Concurrent activities implemented elements should be checked realization, baselines should be monitored
occur, and reprioritization of work and to ensure that they are still working, still for current state of quality, maturity, and
pivoting in direction is common. During meet needs and are still aligned with overall performance as well as trends that may
this phase, large formal reviews are costly program direction. Progress to date can need attention.
and ineffective and should be eliminated also be evaluated compared to the overall Continuous improvement based on
in favor of frequent, tightly scoped light mission requirement and remaining time short cycles of learning is an intrinsic
weight course correction touch points. The and budget. advantage of agile development, but only
focus of these touchpoints should primarily Integration and synchronization are when improvement opportunities are iden-
be capabilities just completed and those critical to success on agile projects. Disci- tified, and improvements made.
that are selected for completion in the pline in continually identifying and man-
upcoming iteration. Requirements, design, aging dependencies at all levels is critical, Review Approach
and test plans/results for the capabilities as there is typically not a lengthy “get well” The general approach to agile reviews
being reviewed are all considered. integration period planned near the end of should reflect agile values and principles.
Transition – at some point, agile capabil- the program. Reviews should be cross functional and
ity development ceases, or at least pauses. Health of baselines must be evaluat- collaborative, not stove-piped. The focus
In some cases, the contract may end, and ed regularly. The possibility of multiple should be on delivery of value not comple-
the program shuts down. In other cases, significant changes occurring close together tion of intermediate artifacts. They should
an initial operating capability may have means that baselines must be up to date, be collaborative, not confrontational, and
been achieved, or a first article completed working, and configuration controlled at all seek to course correct without limiting
and transition to operations or production times. Deficiencies such as technical debt velocity, with goals of continuous improve-
required. This is another useful time for and performance shortfalls must be tracked ment rather than striving for immediate
an extensive review, this time to determine and kept visible due to their impact on perfection.
if the deliverable items are ready for the decisions and prioritization of work.
pending transition. For government and Working execution assets such as Review Outcomes
defense contracts, transition reviews are backlog, models, regression test results, and The outcomes of an agile review should
a good point to evaluate against required demonstration outcomes provide excellent align with the principles and characteristics
standards such as technology readiness lev- insight into program progress, process, and of agile project realization. Some desirable
el (TRL) or manufacturing readiness level risk without requiring program personnel outcomes are described below.
(MRL) or to assess the solutions readiness to construct presentations. Enhanced value delivery – agile projects
to request operational permissions such as Best practices for whatever agile measure success in valuable capabilities
55
21564868, 2023, 2, Downloaded from https://incose.onlinelibrary.wiley.com/doi/10.1002/inst.12446 by libffg rrtyu - Beijing Normal University , Wiley Online Library on [09/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
delivered to the customer. Reviews should to make actionable recommendations for to transform into eagerness for inquiry and
FEATURE
SPECIAL
strive to make actionable recommendations improvements not only to the program improvement. And we must make these
that enhance value delivery in some way. under review, but to the review team, changes in an environment of constantly
This can include increased quality, faster the review process, and the organization accelerating change. However daunting
delivery, lower cost, decreased risk, or overall. that sounds, it’s the best option before us.
better fulfillment of mission needs. If the world is changing rapidly, we need
Predictable performance – given Summary agile practices, including agile oversight, to
that agile projects encourage ongoing The key to agile reviews is to review as respond effectively to that change.
modifications to the solution intent based you develop — agilely. Encourage inline Fortunately, there are nuggets in new
JUNE 2O23
VOLUME 26/ ISSUE 2
on changes in the mission, technical and quality improvements rather than waiting practices and changing technologies
business environment, it’s critical that for outsiders to initiate corrections. Focus that promise to help us provide effective
the team performance can be reliably on valuable outcomes, fast feedback, and oversight that keeps pace with change.
predicted. This enables effective replanning continuous learning. Integrated digital engineering and the
and reprioritization. Reviews strive to make use of models at all levels of the system
actionable recommendations that enhance CHALLENGES AND OPPORTUNITIES of interest provide direct insight into the
performance transparency through Implementing an agile approach to current state of baseline capability and
metrics, retrospectives, management of reviews is not without its challenges. maturity. Increasing automation offers
dependencies, and risks. Existing norms and processes may need several opportunities. Automated checks of
Ongoing improvements – successful to change to admit new ways of working. concerns like traceability and test coverage
agile projects leverage the short cycles Standards and processes that assume can reduce human effort in reviews as they
of learning inherent in the agile practice sequential execution of engineering also enhance quality. The use of end-to-end
to continually improve in a wide range activities need to be refactored to address DevSecOps pipelines offers the possibly
of areas throughout their period of end to end completion of small units of in-line zero preparation reviews, in
performance. Reviews strive to make of work. The perception that reviews which data collection and presentation
actionable recommendations for are “grades” and the program’s goal is become an ongoing automated function
improvement. Prioritize those areas where to “pass” must evolve to see reviews as of the pipeline and meetings can focus on
the program may be struggling, but don’t checkpoints that identify opportunities discussing concerns and ways to overcome
ignore those areas that are “good enough. to improve. Criticism needs to give way them. The data and information gathered
Continuous learning for all – for to collaboration, where all participants from these reviews can be fed back to ev-
programs, review teams, and organizations are focused on enhancing value delivery. ery level of our organization to ensure that
“the best we can do” last time may not be Cultural habits of covering our backsides or we continue to improve on pace with the
the best we can do this time. Reviews strive hiding our challenges and shortfalls needs challenges we face. ¡
REFERENCES
■■ Beck, K., et al. 2001. “Manifesto for Agile Software Develop- Engineering. Published by John Wiley & Sons, Inc.
ment.” viewed 21 November 2021. https://agilemanifesto.org/ . ■■ Lapham, M. et al. 2016. RFP Patterns and Techniques for
■■ Defense Science Board. 2018. “Design and Acquisition of Successful Agile Contracting. Carnegie Mellon University,
Software for Defense Systems.” 14 February. Software Engineering Institute Special Report, November.
■■ Dove, R., W. Schindel, and K. Garlington. 2018. “Case Study: ■■ Marbach, P., L. Rosser, and G. Osvalds. 2015. “Principles for
Agile Systems Engineering at Lockheed Martin Aeronautics Agile Development.” Proceedings International Symposium.
Integrated Fighter Group.” Proceedings International International Council on Systems Engineering, Virtual Event,
Symposium, International Council on Systems Engineering, 17-22 July.
Washington, US-DC, 7-12 July. ■■ National Defense Industrial Association. 2019. An Industry
■■ Dove, R., and W. Schindel. 2017. “Case Study: Agile SE Process Practice Guide for Agile on Earned Value Management
for Centralized SoS Sustainment at Northrop Grumman.” Programs.
Proceedings International Symposium. International Council ■■ Office of Management and Budget. 2012. ”Contracting
on Systems Engineering, Adelaide, AU, 15-20 July. Guidance to Support Modular IT Development.” Viewed 20
■■ Dove, R., W. Schindel, and R. Hartley. 2017. “Case Study: Agile March 2022.https://obamawhitehouse.archives.gov/sites/default/
Hardware/Firmware/Software Product Line Engineering at files/omb/procurement/guidance/modular-approaches-for-
Rockwell Collins.” Proceedings 11th Annual IEEE International information-technology.pdf .
Systems Conference, Montréal, Québec, CA, 24-27 April. ■■ Office of the Under Secretary of Defense for Acquisition,
■■ Eby, K. 2016. ”Understanding the Agile Software Development Technology, and Logistics. 2017. Best Practices for Using
Lifecycle and Process Workflow.” SmartSheet. Viewed 4 March Systems Engineering Standards (ISO/IEC/IEEE 15288, IEEE
2022. https://www.smartsheet.com/understanding-agile-soft- 15288.1, and IEEE 15288.2) on Contracts for Department of
ware-development-lifecycle-and-process-workflow . Defense Acquisition Programs.
■■ Enterprise Unified Process c. 2020. ”Life Cycle Phases.” ■■ Program Management Institute c. 2022. ”Full Delivery Life
Viewed 4 March 2022. http://www.enterpriseunifiedprocess.com/ Cycles.” viewed 3 March 2022. https://www.pmi.org/disciplined-
essays/phases.html . agile/process/introduction-to-dad/full-delivery-lifecycles-
■■ IEEE Computer Society. 2014. ISO/IEC/IEEE 15288.2, IEEE introduction .
Standard for Technical Reviews and Audits on Defense ■■ Rosser, L., G. Osvalds, P. Marbach, and D. Lempia. 2014.
Programs. “Systems Engineering for Software Intensive Projects Using
■■ INCOSE. 2015. Systems Engineering Handbook: A Guide for Agile Methods.” Proceedings International Symposium.
System Life Cycle Process and Activities (4th ed.). D. D. Walden, International Council on Systems Engineering, Las Vegas US-
G. J. Roedler, K. J. Forsberg, R. D. Hamelin, and T. M. Shortell NV,30 June – 3 July.
(Eds.). San Diego, US-CA: International Council on Systems > continued on page 65
56