Harvard University’s Bob Behn has been working on his latest book about the “Stat” movement for more than a decade. I’ve been eagerly awaiting the release of The PerformanceStat Potential and told him I would read it and share my impressions with others. In his inimitable response, he just asked that I spell “Behn” and “PerformanceStat” correctly.
Since 2001, Bob Behn has visited dozens of governmental organizations that are using the PerformanceStat model—some well and some poorly. His bold objective was to answer the research question of whether PerformanceStat really makes a difference in improving performance and how it works. In short, his answer is: It depends, and it’s complicated.
Behn has written a definitive book about the PerformanceStat phenomena, and apologizes that it is so long. The challenge, he says, was converting the tacit knowledge he developed during years of research into explicit written knowledge in a book for government executives.
“PerformanceStat” is Behn’s shorthand for a concept developed 20 years ago by New York City deputy police commissioner Jack Maple as a crime reduction strategy, which Maple dubbed “CompStat.” This approach was so successful in reducing crime, it quickly spread to other government functions and was called similar names—CitiStat, StateStat, ParkStat—and sometimes completely different names, but with the same core elements.
None of these various “Stats” were exactly the same, so Behn culled some common characteristics from his observations:
A leadership strategy that is designed to achieve specific public purposes, where the leadership team persists in holding ongoing series of regular, frequent, integrated meetings.
At these meetings, the leadership team:
- Uses current data to analyze specific, previously defined aspects of recent performance.
- Provides feedback on performance versus targets.
- Follows up on previous decisions and commitments to produce results and learn from efforts to improve.
- Identifies and solves performance-deficit problems, and sets the next performance targets.
This approach has been adopted in hundreds of government organizations at all levels. In the process, it has attracted a wide range of champions and promoters who are creating a movement to use it more broadly. In fact, there is implicit encouragement within the federal government to use this approach. It is embedded in the 2010 GPRA Modernization Act and guidance from the Office of Management and Budget. The mandate requires each agency’s chief operating officer to conduct quarterly performance reviews for priority goals set by their agency. But the law and guidance don’t say how to conduct meaningful reviews. They offer no help on which strategies and behaviors are ineffective or destructive.
How PerformanceStat Works
About the characteristics of PerformanceStat, Behn says “you can find these principles and components in any management book.” There are many process-oriented descriptions of how PerformanceStat works. In fact, in 2007, Behn wrote a guide showing mayors how to conduct a CitiStat, based on Baltimore’s version. That guide describes the meeting room, the attendees, how the meetings are conducted, and the resulting improvements in performance. But Behn cautions that these descriptions don’t get at the hard problem—understanding what makes a PerformanceStat work and why. That led him to visit dozens of PerformanceStat operations over the years to figure it out, and then determine how to explain.
He says there is too much of a focus on the visible features of PerformanceStat (data, projectors, meetings) versus sorting out the underlying cause-and-effect of how it works. He concludes that it is a leadership strategy, not a standardized, mechanistic system. In fact, he says “one of the worst things that could happen to the PerformanceStat strategy would be for it to be kidnapped by the ISO cabal.”
Behn says the success of PerformanceStat is that it is a leadership strategy, not a policy or a reform; it is a discipline, not a model; it is a way of thinking and behaving.
Essentially, PerformanceStat shifts responsibility for action from an organization to an individual. “They need to get specific individuals to accept responsibility for specific results,” Behn says. And, in terms of using this approach to enforce accountability versus improving performance, he adds: “PerformanceStat is not about them [the accountability imposers]. PerformanceStat is about us . . . It is a leadership strategy that we are using to improve our own performance . . . an internal effort to achieve public purposes by producing better results.”
But Behn says you can start only after members of the leadership team have made some “basic choices—purpose; performance deficit; target—[so they] can they craft the rest of their performance strategy. They can’t start with data. They can’t start with meetings. They have to start with purpose.”
A Complex Adaptive System
Behn says PerformanceStat is best viewed as a complex, adaptive system, not a mechanistic system or model. “The challenge facing the leadership team is not to build some new, fabulous machine,” he says. “Rather, it is to work with the existing people, relationships, and structures . . . all while analyzing data, asking questions, scrutinizing reports, in an effort to learn whether their leadership behaviors are introducing new feedback loops that foster adaptations that help to achieve the purposes.” Behn identifies and describes 16 leadership behaviors that help explain what he means. These behaviors include clarifying public purposes to focus upon, analyzing data, creating targets and timeframes, and assigning responsibility for action.
So Where Do You Start?
Behn’s research concludes that the PerformanceStat leadership strategy does improve performance, when adapted thoughtfully—not mechanistically—to an organization’s environment. He also concludes that we pretty much know what to do, technically, to get it to work. The challenge is getting people to do it.
Launching a PerformanceStat initiative requires a personal commitment from leaders. But explicitly and publicly committing to observable, measurable results with deadlines is often seen by politicians as dangerous. Behn concludes: “Superficially, PerformanceStat appears to be about data and accountability—a system with data and computers. At its core, however, PerformanceStat is about the responsibility to achieve specific public purposes. For public executives, embracing this responsibility is dangerous . . . [Nevertheless,] to realize the PerformanceStat potential, public executives have to make an explicit—and thus very dangerous—commitment.”
What is remarkable is that there are so many public leaders willing to do this anyway.