Scrum

Jeff Sutherland, a former fighter pilot, cancer researcher, and software executive, co-created a project management framework called Scrum with Ken Schwaber in the early 1990s. In this book, Sutherland argues that Scrum, which became the dominant method for building software in the technology industry, can transform how any organization works. He traces Scrum's origins, explains its core mechanics, and presents case studies from business, education, government, and poverty alleviation to make the case that the framework can help teams in any field dramatically increase their productivity.

Sutherland opens with the FBI's decade-long failure to modernize its computer systems. In 2010, the Bureau still filed most reports on paper using a system from the 1980s, a deficiency the 9/11 Commission identified as a key reason the Bureau failed to detect the September 11 attacks. Two successive modernization attempts collapsed: the Virtual Case File system consumed $170 million and produced nothing usable, and the Sentinel project, contracted to Lockheed Martin for $451 million, had spent $405 million by 2010 to complete only half the work. Both relied on the Waterfall method, a sequential approach in which every step is mapped out in advance on detailed Gantt charts. Sutherland argues that these plans invariably prove wrong when they meet reality. Assistant Director Jeff Johnson and CIO Chad Fulgham brought development in-house, cut staff from hundreds to under fifty, and adopted Scrum. Working in two-week Sprints, the team measured velocity, the amount of work completed per cycle, and tripled their productivity. Sentinel was deployed bureau-wide in July 2012.

Sutherland then traces his path to creating Scrum. As an unarmed reconnaissance pilot in Vietnam, he internalized the OODA loop, a decision-making cycle of Observe, Orient, Decide, and Act. After the war he studied statistics, artificial intelligence, and biometrics, researching how cancer cells transition between stable states. He recognized that organizations behave like complex adaptive systems, groups whose behavior emerges from simple rules and environmental interactions, and began searching for rules that could shift teams into more productive states. At MidContinent Computer Services in 1983, he restructured a failing division into a self-contained unit with team-based bonuses; within six months it became the company's most profitable. In 1993, at a software company called Easel, he formally created Scrum. A team member introduced a 1986 Harvard Business Review paper by Hirotaka Takeuchi and Ikujiro Nonaka, "The New New Product Development Game," which described how top companies used overlapping, cross-functional, autonomous teams rather than sequential hand-offs, comparing the best teams to a rugby scrum. Sutherland implemented these ideas and delivered the product on time and under budget, later codifying the framework with Schwaber in a 1995 paper. He also credits W. Edwards Deming, whose PDCA cycle of Plan, Do, Check, Act became foundational to the Toyota Production System and, by extension, to Scrum.

Sutherland argues that teams, not individuals, are the fundamental unit of performance, citing research across roughly 3,800 projects showing that team-level performance differences dwarf individual ones. Great teams share three traits identified by Takeuchi and Nonaka: a transcendent purpose, autonomy to decide how to achieve their goals, and cross-functionality, meaning the team possesses all the skills needed to complete its mission without hand-offs. He illustrates these traits with examples including his experience turning the lowest-ranked cadet company at West Point into the top-performing one through transparent performance ratings, NPR's self-organized coverage of the 2011 Egyptian revolution guided by Scrum's daily check-in, and U.S. Army Special Forces A-teams. Teams should be small, ideally seven members plus or minus two, since research shows groups of three to seven required about 25 percent of the effort of groups of nine to twenty. He introduces the Scrum Master, a servant-leader who facilitates the process and helps the team remove impediments. He also warns against the Fundamental Attribution Error, the tendency to blame individuals rather than examine systems, citing the NUMMI auto plant where Toyota rehired the workforce General Motors had called the worst in America and produced cars with Japanese-level quality simply by changing the system.

Time, Sutherland argues, must be structured into short, fixed cycles. He borrowed the Sprint concept from the MIT Media Lab, which required teams to demonstrate working projects every three weeks or face cancellation. Sprints create a consistent rhythm: the team commits to a set of tasks, works without outside interference, and delivers a usable increment at the end. He pairs Sprints with the Daily Stand-up, a fifteen-minute meeting where each member answers three questions: What did you do yesterday? What will you do today? What is in your way? This practice derives from Jim Coplien's study of communication saturation, a measure of how effectively information flows among team members, at Borland Software. When the first Scrum team at Easel adopted the Daily Stand-up, it finished a month's planned work in one week. To illustrate Scrum's broader applicability, Sutherland describes Eelco Rustenburg, who used Scrum to remodel his house in six weeks on budget; Rustenburg's neighbor, doing identical work on an identical house with the same contractors but without Scrum, took three months and paid nearly twice as much.

A chapter on waste draws on Taiichi Ohno, the architect of the Toyota Production System, who classified waste into three types: Muri (unreasonableness), Mura (inconsistency), and Muda (outcomes). Sutherland attacks multitasking, citing research showing that working on five simultaneous projects results in 75 percent loss to context switching. He argues that half-finished work creates no value, and presents data from Palm showing that fixing a software bug three weeks after its creation took twenty-four times longer than fixing it the same day. He challenges the culture of overwork, presenting Scott Maxwell's finding at OpenView Venture Partners that peak productivity fell at slightly under 40 hours per week. Research on Israeli judges showed that parole decisions correlated most strongly with how recently the judges had eaten, illustrating how decision-making degrades as energy is depleted.

On planning, Sutherland introduces the Cone of Uncertainty, which shows that initial project estimates can be off by a factor of 16. He advocates estimating work in relative terms using the Fibonacci sequence (1, 3, 5, 8, 13), since humans are poor at absolute estimates but skilled at comparing sizes. To avoid groupthink, he recommends Planning Poker, where team members simultaneously reveal numbered cards representing their estimates, discuss outliers, and converge. Work items should be framed as user stories following the format "As an [X], I want [Y], so that [Z]," capturing the character, the desired outcome, and the motivation. He returns to the Medco case: A pharmaceutical company facing a Wall Street deadline used Scrum to cut a two-foot stack of requirements down to prioritized sticky notes. After three Sprints, team velocity rose from 20 to 60 points per Sprint. When Sutherland presented management with a list of 12 impediments, a Senior Vice President eliminated all of them within days, boosting velocity by another 50 percent. The team delivered on time, and Medco's stock price roughly doubled within the year.

Sutherland argues that happiness drives success rather than following from it, citing a meta-analysis of 225 papers concluding that happiness precedes important outcomes. He introduces the Happiness Metric: four questions asked at the end of each Sprint about how team members feel and what one change would make them happier. The top improvement becomes the highest priority for the next Sprint. At Scrum, Inc., this practice tripled velocity within weeks. He identifies autonomy, mastery, and purpose as foundations of workplace happiness, and warns against "happy bubbles," complacent teams that stop improving.

The chapter on priorities introduces the Product Owner, the person responsible for deciding what to build and in what order. Modeled on Toyota's Chief Engineer, a role known in Japanese as Shusa that leads through persuasion rather than authority, the Product Owner must have domain knowledge, decision-making power, availability to the team, and accountability for value delivered. Sutherland advocates releasing a Minimum Viable Product (MVP) as early as possible to obtain real feedback, and introduces the "Change for Free" contract model, where customers can reorder Backlog items each Sprint without change fees.

The final chapter extends Scrum beyond business. At Ashram College in the Netherlands, chemistry teacher Willy Wijnands applies eduScrum, an adaptation of Scrum for education, having students form cross-functional teams and run multi-week Sprints; test scores jumped more than 10 percent. In Uganda, the Grameen Foundation equipped roughly 1,200 Community Knowledge Workers, local people who share agricultural data and real-time market prices with rural farmers, with smartphones; one farmer reported doubling both her crop yield and her prices. Iceland attempted to draft a new constitution using Scrum after its 2008 financial crisis, publishing sections weekly and collecting public feedback via social media, though entrenched political interests blocked adoption. At Valve, the self-funded videogame company behind Half-Life and Portal, there is no management hierarchy: employees choose their own projects and recruit teammates by persuasion. Sutherland closes by urging readers to reject cynicism, arguing that Scrum provides a practical framework for removing the impediments that prevent people and organizations from achieving greatness.