IT MANAGEMENT BEST PRACTICE SERIES
Agile Development Adoption and
Best Practices
January 2017
After more than two decades of maturation, agile development methodologies promise to deliver
new systems more quickly, improve software quality, and increase user satisfaction. While adoption
remains only moderate, agile development continues to gain adherents as organizations seek to
respond to a continuously changing environment. In this study, we report how widely and deeply
agile development is being embraced by IT organizations today, by organization size and sector. We
conclude with our recommendations on how to best use, and not misuse, agile development
methodologies.
Inside this Report
TABLE OF CONTENTS
Agile Is a Collection of Approaches 3
Agile Earns a Low, but Growing Maturity Rating 6
Practice Rate Expands to Half of Organizations 7
Few Companies Fully Practicing Agile 8
Larger Organizations Lead in Adoption 9
High-Tech and Discrete Manufacturing Practice Agile to Greatest Extent 10
Computer Economics Viewpoint 11
Methodology and Sample 13
About Computer Economics 14
TABLE OF FIGURES
Fig. 1: Maturity Profile: Agile Software Development 6
Fig. 2: Adoption Rate Trend: Agile Software Development 7
Fig. 3: Practice Adoption Stages: Agile Software Development 8
Fig. 4: Practice Rate and Level by Organization Size: Agile Software Development 9
Fig. 5: Adoption Rate by Sector: Agile Software Development 10
Fig. A-1: Key Demographics of Survey Sample 13
RELATED STUDIES
IT Management Best Practices
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Agile Development
Adoption and Best Practices
The software development function remains as important as ever, even though
packaged applications and software as a service are reducing the level of custom
software in today’s enterprise. Freed from the need to develop and maintain
custom systems for back-office applications, such as payroll, accounting, ERP,
and CRM, the application development function can give greater focus to more
strategic business applications that deliver real competitive advantage. But,
being more strategic, the requirements behind such applications are much more
volatile. Business conditions are constantly changing, and business users are
often not able to clearly articulate their needs up front. Older “waterfall”
Older ‘w aterfall’ methodologies, with long up-front phases for requirements specification and
m ethodologies, w ith design prior to development, are simply not up to the task. A more agile
long up-front phases approach is needed.
for requirem ents
specification and Agile Is a Collection of Approaches
design, are sim ply Agile development is an umbrella term for a set of software development
not up to the task. methods that emphasize collaboration within tightly knit teams, iterative
development, early delivery, continuous improvement, and the ability to
respond rapidly to changing requirements.
The major approaches to software development that fall under the agile
umbrella include the following.
Extreme Programming. This was the first manifestation of agile
development, in the 1990s, which promoted a philosophy of building
system features incrementally and then simplifying the code in
subsequent iterations. Developers work in small empowered teams with
customers actively involved during the development process. Instead of
a long up-front attempt at detailed requirements specification and
functional design, Extreme Programming promotes “user stories” and
“use cases,” which give examples of the kinds of things that users will
want the new system to do. An interesting element of this approach is
its use of pair programming, which takes peer reviews to the extreme:
two developers work on the same piece of code at the same time, one
© 2017, COMPUTER ECONOMICS INC. PAGE 3
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
narrating what he or she is doing, while the other gives continual
feedback.
Lean Software. This approach to agile development borrows from
concepts of Lean Manufacturing to eliminate waste, which is anything
that does not deliver value to the customer. Detailed documentation of
user requirements, status reports, and unnecessary management
activities are examples. Lean Software also adopts the concept of just-
in-time production to delay design decisions or development of
features until customers actually need them. Some Lean Software teams
also employ Kanban techniques to minimize work in progress: new
tasks are only put into development when other tasks are completed,
thus encouraging developers to work on one thing at a time. Consistent
with Lean Manufacturing, Lean Software promotes empowered teams
of developers, quality designed into the development process, and a
Scrum em phasizes
holistic view of the project, focusing on the end goal.
effective team ing as
an alternative to Scrum. Borrowing the term “scrum” from the sport of rugby, this
traditional top-dow n approach focuses on the organizational aspects needed to make agile
project m anagem ent development effective. However, scrum terminology can take some
techniques. getting used to. For example, a “sprint” is a fixed duration development
cycle, typically a week or two, to deliver a pre-defined set of
functionality requirements. A “daily scrum” is a short daily project
team meeting to plan the day’s work. A “product owner” is the
project’s key stakeholder, who conveys the project’s objectives to the
scrum team. A “scrum master” is the facilitator of the agile team.
Scrum emphasizes effective teaming as an alternative to traditional top-
down project management techniques. The approach also includes
period reviews (“sprint retrospectives”) so that the project team can
learn from its experience and apply lessons learned to the next sprint.
As far as the development process itself, scrum adds the concept of
freezing requirements during each sprint. Organizations interested in
adopting scrum can take advantage of a variety of training programs, as
well as certification programs offered by the Scrum Alliance.
Crystal. Less well-known than scrum, crystal provides a series of
methods to deal with projects of varying sizes and criticality. This
reflects the reality that a team responsible for a large mission-critical
project, for example, should take a different approach than the
approach recommended for a small noncritical project. Crystal’s
primary focus is on effective communications within the project team.
Like other agile methods, it also promotes iterative development and
delivery of small pieces of functionality and involvement of the
customer during development. Like lean software, it encourages
© 2017, COMPUTER ECONOMICS INC. PAGE 4
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
developers to work at one thing at a time. Like scrum, it requires the
team to periodically look back and learn from their experiences.
These are four complementary approaches to software development, with many
overlapping principles and techniques. They can be used individually or in
combination to meet the needs of the organization and the project at hand.
In addition to these four approaches, there are several other methodologies that
can be grouped under the agile umbrella. These include:
The Dynamic System Development Method (DSDM), which dates
from 1994. It is an iterative development methodology. With its focus
on working within overall budget and schedule constraints, it may be
appropriate in projects where developers must meet contractual
commitments.
The Rational Unified Process (RUP), which provides a greater
number of artifacts than are commonly delivered by agile
methodologies. As such, it is useful in large, more complex
development projects.
The Distributed System Development (DSD) methodology, which
provides agile best practices where developers are in remote locations.
As such, it is useful in conjunction with one of the other agile
methodologies discussed in this report.
While adoption remains only moderate, agile development continues to gain
adherents as organizations seek to respond to continuously changing
requirements. In this study, we report how widely and deeply agile development
is being embraced by IT organizations today, by organization size and sector.
We conclude with our recommendations on how to best use, and not misuse,
agile development methodologies.
© 2017, COMPUTER ECONOMICS INC. PAGE 5
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Agile Earns a Low, but Growing Maturity Rating
Among commercial software providers, agile development in one form or
another is far and away the most dominant approach to software development.
Microsoft, Google, Facebook, Twitter and virtually all cloud services providers
have been using agile methods and principles for many years. Although agile
development among corporate IT organizations is much less well established,
its use is growing.
Our maturity profile, as shown in Figure 1, compares agile development with
the other 32 practices in our annual study, IT Management Best Practices. The
Although agile ratings are based on a relative scale, defined by the lowest and highest values in
developm ent am ong the study, providing a context for understanding the adoption of agile
corporate IT development. Our findings are as follows:
organizations is
m uch less w ell Practice Rate: Relative to the other IT best practices in our study, the
percentage of organizations practicing agile development is moderate.
established, its use is
While adoption is growing, it is far from a standard practice in business
grow ing.
IT organizations.
Practice Level: The practice level is low. Among companies that have
embraced agile development, the percentage that claim to be “fully
practicing” falls within the bottom third of our scale.
Maturity Rating: This practice earns a low maturity rating based on its
combined practice rate and level. Many of the practitioners are just at
the beginning stages of adoption.
Fig. 1: Maturity Profile: Agile Software Development
© 2017, COMPUTER ECONOMICS INC. PAGE 6
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Practice Rate Expands to Half of Organizations
In Figure 2, we see the pace of adoption for agile development. From 2015 to
2016, adoption rates rose from 49% to 55%, continuing a steady increase since
we added it to our survey. For the first time, agile development methods are
embraced by more than half of all IT organizations. Still, many organizations
For the first tim e, either prefer traditional methods or are not doing in-house development and
agile developm ent customization of their applications.
m ethods are
em braced by m ore
than half of all IT
organizations.
Fig. 2: Adoption Rate Trend: Agile Software Development
© 2017, COMPUTER ECONOMICS INC. PAGE 7
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Few Companies Fully Practicing Agile
Moreover, many adopters are only partially practicing agile development. Figure
3 shows where organizations are at each level of adoption. Among the adopters,
as shown by the green bars, more than half are fully or partially using agile
development methodologies. (Our survey defines partial practice to include
situations where organizations are attempting to practice agile development but
have not provided formal training in the methodologies.) While 18% of all
organizations claim to be fully practicing agile development, another 38% are
partially engaged with the practice. This accounts for the low practice-level
rating in the maturity profile.
About 18% of all
organizations claim
to be fully practicing
agile developm ent.
Fig. 3: Practice Adoption Stages: Agile Software Development
The low rates of organizations fully practicing agile development shouldn’t
come as a major surprise. Not only is agile development rather new, but it is
perceived by many as not appropriate for all projects. Some IT leaders believe
that traditional development practices work best for projects with known,
complex requirements or for projects where contracting with third-party
developers lead buyers, rightly or wrongly, to insist on a traditional waterfall
approach. Also, agile works best when developers are dedicated to specific
projects, one at a time, which is not always possible. Many organizations believe
applying a mix of development practices depending on specific needs is the best
strategy.
Among the remaining organizations, 23% report no activity when it comes to
even considering adopting agile development. That leaves 10% that are in the
process of adopting this practice and 11% that are considering doing so. Most
of the growth in adoption will come from companies already partially engaged
with it, but the practice is also attracting a fair number of new adherents.
© 2017, COMPUTER ECONOMICS INC. PAGE 8
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Larger Organizations Lead in Adoption
Organization size has a significant influence on adoption trends for agile
development. As shown in Figure 4, larger organizations are far more likely to
engage in this practice than smaller organizations. Among large organizations,
70% practice agile development, compared with 55% of midsize organizations
and 44% of small organizations.
Organization size has
a significant influence
on adoption trends
for agile
developm ent.
Fig. 4: Practice Rate and Level by Organization Size: Agile Software Development
Enthusiasm for the adoption of agile development doesn’t translate into any
discernible pattern in practice levels, however: 33% of small organizations are
fully engaged with the practice compared with 22% of midsize and 40% of large
organizations.
In this study, we define small organizations as those with IT operating budgets
of less than $5 million, midsize as those with budgets of from $5 million to less
than $20 million, and large as those with IT operating budgets of at least $20
million.
© 2017, COMPUTER ECONOMICS INC. PAGE 9
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
High-Tech and Discrete Manufacturing Practice
Agile to Greatest Extent
Among industry sectors, organizations in the high-tech and discrete
manufacturing sectors are more likely to use agile development methods. Both
sectors have a 74% adoption rate, Figure 5 shows.
The hospitals/ health
system s and energy
and utilities sectors
are am ong the least
likely to adopt agile
developm ent.
Fig. 5: Adoption Rate by Sector: Agile Software Development
The hospitals/health systems and energy and utilities sectors are among the
least likely to adopt agile development. Hospitals/health systems report 27%
adoption rates, and the energy and utilities sector report a 15% adoption rate.
© 2017, COMPUTER ECONOMICS INC. PAGE 10
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Computer Economics Viewpoint
Agile development may have moved into the mainstream, but it still has a long
way to go to attain widespread adoption in enterprise IT. The low practice level
indicates that there are barriers to implementation or that individual developers
are adopting agile development without formal training by the organization as a
whole. And even when they are fully ready to adopt agile development, many
perceive that it is not right for every project. IT organizations should pay
particular attention to the best practices developed over many years.
Agile development has already proven itself as capable of delivering business
value. In fact, major buyers such as the U.S. Department of Defense now
Agile enthusiasts mandate some sort of iterative development in their software procurement
som etim es do contracts.
dam age to the cause
by prom ising Unfortunately, agile enthusiasts sometimes do damage to the cause by
promising unrealistic benefits and, at the same time, unnecessarily rejecting
unrealistic benefits
other best practices as “not ‘agile.’ ” Business leaders, therefore, should embrace
and, on the other agile development, while developers should not overpromise what it can deliver,
hand, unnecessarily nor reject well-established good business practices out of hand.
rejecting other best
practices as ‘not Our recommendations are to find the broad middle ground where agile
agile.’ development can be best used, and not misused, within the context of best
business practices.
Lay a foundation with formal training. IT organizations sometimes
get into agile development through the efforts of a handful of
developers who either read about it or who have experience with a
previous employer. Although such enthusiasm is commendable, it is
not enough to build the knowledge and skills needed across the IT
organization. Start with formal training in agile development from a
reputable training organization. Give stakeholders and business users
an introduction to agile development along with more in-depth training
for developers.
Start with a few pilot projects. Obviously, the first attempts at Agile
development in an IT organization should not be for mission-critical
projects. Find two or three small development projects that can serve
as pilots for applying agile methods learned in the training sessions. If
there are two or three experienced agile developers already on staff,
assign them as scrum masters. The scrum retrospectives are a great way
to learn from these early experiences and to decide how agile
development should be applied in future projects. This will give the IT
organization the beginnings of a standard set of agile practices
specifically tailored to the needs of the business.
© 2017, COMPUTER ECONOMICS INC. PAGE 11
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Do not reject estimating. As mentioned earlier, some proponents of
agile tend to take things to an extreme, rejecting other best business
practices as “not ‘agile.’ ” A prime example of this is in estimating,
where some vocal advocates of agile reject project estimating, believing
Another area w here that developers should only be required to commit to a set of
som e agile deliverables, or a budget, but not both. Needless to say, such an attitude
enthusiasts get is not generally shared by business stakeholders, who have to plan
carried aw ay is in the within constraints of time and money. Agile development has been
necessity of successfully applied to large-scale, mission-critical software projects in
docum entation. aerospace and defense, where contractual estimates are the rule. If agile
can be applied in such environments, do not accept claims that
estimating is incompatible with agile development.
Insist on a proper level of documentation. Another area where
some agile enthusiasts get carried away is in the necessity of
documentation. Some reject any form of documentation whatsoever as
“waste,” believing that well-written code is sufficient. Although that
may be true for the original development team, it is not true for users
of the new system, who need at least some level of user documentation
and training. Likewise, future developers assigned to the system will be
thankful for at least some basic documentation that explains what the
application does, its major components, and the business needs that it
fulfills. In policies and procedures, organizations should include a
minimum level of documentation for systems developed through agile.
The above recommendations apply to most agile development projects, but it is
far from a complete list. Start small, and learn from experience. In this way,
agile development, properly applied, can become part of the IT organization’s
culture and deliver real business value.
For further information on various agile methods, along with an assessment of their
effectiveness, see Agile! The Good, the Hype and the Ugly, by Bertrand Meyer. Thanks
to Glen Alleman, Niwot Ridge, LLC, for reviewing and providing feedback on this report.
© 2017, COMPUTER ECONOMICS INC. PAGE 12
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
Appendix
Methodology and Sample
The survey was conducted from January to April 2016. We identified and selected participants by making
solicitations to specific organizations in the U.S. and Canada that met our criteria for organization size and
industry sector. As the survey progressed, we monitored response volume by industry and organization size
and adjusted our survey solicitation activities accordingly to ensure that the stratification of the survey sample
was within acceptable bounds. Surveys were conducted using an online survey tool or by electronic form.
A total of 149 IT organizations in the U.S. and Canada participated in the IT Management Best Practices study.
Figure A-1 displays the key demographics for organizations in the sample.
Key Demographics of Survey Sample
25th 75th
Metric Percentile Median Percentile
Revenue $160M $560M $1.69B
IT Spending $3.9M $12.0M $39.0M
Outsourcing Budget $45,000 $294,000 $1.7M
Source: Computer Economics, 2017 Figure A-1
Fig. A-1: Key Demographics of Survey Sample
The sample is balanced by organization size. Large organizations made up 33% of the sample, midsize
organizations 34%, and small organizations, 33%. Large organizations have IT operating budgets of $20
million or greater, midsize organizations have IT operating budgets of from $5 million up to $20 million, and
small organizations have IT operational budgets of less than $5 million. Organizations must have at least $50
million in revenue or $2 million in IT operational spending to be included.
By sector, the sample is composed of process manufacturers (10%), discrete manufacturers (13%),
professional/technical services (16%), banking and finance (15%), government/nonprofit organizations
(12%), hospitals/health systems (8%) and other sectors (26%).
© 2017, COMPUTER ECONOMICS INC. PAGE 13
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
AGILE DEVELOPMENT ADOPTION AND BEST PRACTICES
About Computer Economics
Computer Economics provides research and advisory services on the strategic and financial management of
information technology. Our clients include IT end-user organizations and major consulting firms in North
America. Our IT Spending and Staffing Benchmarks study, published annually since 1990, is the definitive source
of IT benchmarking data.
Other annual studies include Technology Trends, an assessment of technology adoption, spending, and
economic experience; IT Outsourcing Statistics, which provides data on the use of and experience with IT
outsourcing; IT Management Best Practices, which measure adoption trends of strategic IT practices; and IT
Staffing Ratios, a series of benchmarking studies with metrics for 15 IT job functions.
In addition to these major studies, we publish IT management advisories on various issues of concern to IT
managers. These reports are made available through our website. For further information on our custom
benchmarking services, website subscriptions, advisory reports, and other services, please contact our office
or visit our website at www.computereconomics.com.
U U
Contact Information:
Address: 2082 Business Center Drive, Suite 240, Irvine, CA 92612, USA
Telephone: +1 (949) 831-8700; Fax: +1 (949) 442-7688
www.computereconomics.com
© 2017, Computer Economics, Inc., All Rights Reserved
Unauthorized reproduction or distribution in whole or in part in any form, including photocopying, faxing,
image scanning, e-mailing, or making available for electronic downloading is prohibited without written
permission from Computer Economics. Prior to photocopying items for internal or personal use, please
contact Computer Economics, Inc. All trade names, trademarks, or registered trademarks are trade names,
trademarks, or registered trademarks of their respective owners. Information contained in this publication has
been compiled from sources believed to be reliable, but the accuracy of this information is not guaranteed.
Computer Economics disclaims all warranties and conditions with regard to the content, express or implied,
including warranties of merchantability and fitness for a particular purpose, nor assumes any legal liability for
the accuracy, completeness, or usefulness of any information contained herein. Any reference to a
commercial product, process, or service does not imply or constitute an endorsement of the same by
Computer Economics. This publication is designed to provide accurate and authoritative information in
regard to the subject matter covered. It is sold or distributed with the understanding that Computer
Economics is not engaged in rendering legal, accounting, or other professional service. If legal advice or other
expert assistance is required, the services of a competent professional person should be sought.
© 2017, COMPUTER ECONOMICS INC. PAGE 14
UNAUTHORIZED DISTRIBUTION OR REPRODUCTION PROHIBITED
