The devops handbook how to Create World-Class Agility, Reliability, & Security in Technology Organizations By Gene Kim, Jez Humble, Patrick Debois, and John Willis

• 
CSG (2013): In 2013, CSG International had $747 million in 
revenue and over 3,500 employees, enabling over ninety thou-
sand customer service agents to provide billing operations and 
customer care to over fifty million video, voice, and data cus-
tomers, executing over six billion transactions, and printing 
and mailing over seventy million paper bill statements every 
month. Their initial scope of improvement was bill printing, 
one of their primary businesses, and involved a COBOL main-
frame application and the twenty surrounding technology 
platforms. As part of their transformation, they started per-
forming daily deployments into a production-like environment, 
and doubled the frequency of customer releases from twice 
annually to four times annually. As a result, they significantly 
increased the reliability of the application and reduced code 
deployment lead times from two weeks to less than one day.
• 
Etsy (2009): In 2009, Etsy had thirty-five employees and was 
generating $87 million in revenue, but after they “barely survived 
the holiday retail season,” they started transforming virtually 
every aspect of how the organization worked, eventually turning 
the company into one of the most admired DevOps organizations 
and set the stage for a successful 2015 IPO.
CONSIDER BOTH SYSTEMS OF RECORD AND SYSTEMS 
OF ENGAGEMENT
The Gartner research firm has recently popularized the notion of 
bimodal IT
, 
referring to the wide spectrum of services that typical enterprises support. 
Within bimodal IT there are 
systems of record
, the ERP-like systems that run 
our business (e.g., MRP, HR, financial reporting systems), where the correctness 
of the transactions and data are paramount; and 
systems of engagement
, which 
are customer-facing or employee-facing systems, such as e-commerce systems 
and productivity applications. 
Systems of record typically have a slower pace of change and often have 
regulatory and compliance requirements (e.g., SOX). Gartner calls these types 
of systems “Type 1,” where the organization focuses on “doing it right.” 
Systems of engagement typically have a much higher pace of change to support 
rapid feedback loops that enable them to conduct experimentation to discover 
Promo 
- Not 
for 
distribution 
or 
sale

---

Chapter 5 • 57
how to best meet customer needs. Gartner calls these types of systems “Type 
2,” where the organization focuses on “doing it fast.”
It may be convenient to divide up our systems into these categories; however, 
we know that the core, chronic conflict between “doing it right” and “doing 
it fast” can be broken with DevOps. The data from Puppet Labs’ State of DevOps 
Reports—following the lessons of Lean manufacturing—shows that high 
performing organizations are able to simultaneously deliver higher levels of 
throughput and reliability.
 
Furthermore, because of how interdependent our systems are, our ability to 
make changes to any of these systems is limited by the system that is most 
difficult to safely change, which is almost always a system of record. 
Scott Prugh, VP of Product Development at CSG, observed, “We’ve adopted a 
philosophy that rejects bi-modal IT, because every one of our customers 
deserve speed and quality. This means that we need technical excellence, 
whether the team is supporting a 30 year old mainframe application, a Java 
application, or a mobile application.”
Consequently, when we improve brownfield systems, we should not only strive 
to reduce their complexity and improve their reliability and stability, we should 
also make them faster, safer, and easier to change. Even when new functionality 
is added just to greenfield systems of engagement, they often cause reliability 
problems in the brownfield systems of record they rely on. By making these 
downstream systems safer to change, we help the entire organization more 
quickly and safely achieve its goals.
START WITH THE MOST SYMPATHETIC AND 
INNOVATIVE GROUPS
Within every organization, there will be teams and individuals with a 
wide range of attitudes toward the adoption of new ideas. Geoffrey A. 
Moore first depicted this spectrum in the form of the technology adoption 
life cycle in 
Crossing The Chasm
, where the chasm represents the classic 
difficulty of reaching groups beyond the 
innovators 
and 
early adopters
(see figure 9).
In other words, new ideas are often quickly embraced by innovators and early 
adopters, while others with more conservative attitudes resist them (the 
early 
Promo 
- Not 
for 
distribution 
or 
sale

---

58 • Part II
majority
, 
late majority
, and 
laggards
). Our goal is to find those teams that already 
believe in the need for DevOps principles and practices, and who possess a 
desire and demonstrated ability to innovate and improve their own processes. 
Ideally, these groups will be enthusiastic supporters of the DevOps journey.
M
ark
et G
rowth
Time
Innovators
(Techies)
“Just try it.”
Early adopters
(Visionaries)
“Get ahead
of the herd.”
Early majority
(Pragmatists)
“Stick with
the herd.”
Late majority
(Conservatives)
“Stick with
what’s proven.”
Laggards
(Skeptics)
“Just say no.”
The Chasm
2.5%
13.5%
34%
34%
16%