The lesson of the high-risk team’s success seems to be: Surround the problem with the right people; give them early notice of that problem; and align their efforts toward preventing specific instances of that problem. To clarify that last point, this was not a group that was organized to discuss “policy issues around domestic violence.” This was a group assembled to stop particular women from being killed. Note the similarity to the Chicago Public Schools story earlier in the book. Remember this quote from Paige Ponder, who led the district’s Freshman On-Track efforts: “The beautiful thing about teachers—you can have whatever philosophy you want, but if you’re engaged in a conversation about Michael, you care about Michael. It all boils down to something real that people actually care about.… ‘What are we going to do about Michael next week?’ ” That same motivation led the work in Newburyport. The cops and DAs and advocates and health workers all had different institutional priorities. But what they shared was a desire not to see one of their neighbors murdered by her abusive husband. And that shared aim became the fuel for their coordination. The other point of connection between the two stories is the primacy of data, which was a theme I observed repeatedly in my research, and one that surprised me. I knew data would be important for generating insights and measuring progress, but I didn’t anticipate that it would be the centerpiece of many upstream efforts. I mean this even in a literal sense—what the teachers and counselors in Chicago were doing, and what the high-risk team members in Newburyport were doing, was sitting around a table together and looking at data. Discussing how the fresh data in front of them would inform the next week’s work. In Chicago, the data was: Has Michael been coming to school since we last met? How are his grades in all of his courses? How can we help him this week? In Newburyport, the data was: Where was Nicole’s abuser? What has he been doing? How can we help her this week? This kind of system is what Joe McCannon calls “data for the purpose of learning.” McCannon is an expert in scaling up efforts in the social sector—a former nonprofit and government leader, he has advised movements in many countries. McCannon distinguishes “data for the purpose of learning” from “data for the purpose of inspection.” When data is used for inspection, it sounds like this: Smith, you didn’t meet your sales targets last quarter—what happened? Williams, your customer satisfaction numbers are going down—that’s unacceptable. Using data for inspection is so common that leaders are sometimes oblivious to any other model. McCannon said that when he consults with social sector leaders, he’ll ask them, What are your priorities when it comes to data and measurement? “And I never hear back ‘It’s important to set up data systems that are useful for people on the front lines.’ Never,” he said. “But that’s the first principle! When you design the system, you should be thinking: How will this data be used by teachers to improve their classrooms? How will this data be used by doctors and nurses to improve patient care? How can the local community use the information? But that’s rarely how the systems are designed.” McCannon believes that groups do their best work when they are given a clear, compelling aim and a useful, real-time stream of data to measure their progress, and then… left alone. The situation at Expedia, with its millions of unnecessary calls, provides a model. A cross-functional group is presented with a goal: Help millions of our customers avoid the nuisance of calling us. That’s a valuable and challenging target. And then the group is basically locked in a room together, armed with regularly updated data to see if the number of calls is going up or down. The team members come up with theories and then they test them. They watch what works. That’s the “data for learning” part. They don’t need a boss standing over them, hollering out specific targets: “We need to cut four percent call volume by tomorrow!”
A common theme I’ve observed is that in most organizations, the frontline staff are the ones with least access to the data. Most data is used for “introspection” as in this case, and most information is usually descriptive rather than prescriptive.
Knowing for example, your number of visitors everyday, is a descriptive statistic. There is usually no clear actionable or revealing of problem from it. However, these statistics and metrics are usually the easiest to generate.
Having to dive deep into data to reveal certain insights or actions you should take is using data for “learning”, but these are generally less straightforward. Furthermore, the ones whom need to use these data (people closer to the frontline rather than high level managers), are usually not super data driven. Even if they were, they probably wouldn’t get any the attention from the data team, who would quite rightly prioritize the optics from working with high level managers.
In such cases, the organization then slowly shifts towards being more reactive towards customer trends and behaviours, focusing on “downstream” activities, as this book would put it.
