Q & A with Baltimore CitiStat’s Chad Kenney
Socrata Director of GovStat Beth Blauer recently hosted a group of PerformanceStat leaders from across the country to discuss “3 Lessons in Data-Driven Government Performance.” They covered the topics of leadership, program impact, and citizen engagement tips.
We received a number of questions during the hangout that we didn’t have time to answer. Fortunately, Chad Kenney, Director of CitiStat in Baltimore, agreed to answer some of them from his perspective. Chad leads one of the longest-standing, most successful data-driven performance initiatives in the nation.
The following are the questions we received and Chad’s answers.
1. What is the “next level” of performance reporting? Who or what is “best in class”?
Of course, we think Baltimore is the best in class!
For us, there are a couple different pieces to reaching the “next level.” The first is trying to automate the collection of as many of the performance metrics as possible, while still maintaining the flexibility and ease of adding, removing, or changing metrics. As described below, the current process relies on the agencies to provide summary data to us and it takes a considerable amount of energy.
It’s way, way better to spend that energy on measuring progress than not measuring progress at all. However, if we can devise some methods to automate this collection and reporting, it will free up more time for analysis and problem solving.
The second piece is to continue to look for focus areas where we can have meetings centered around issues and not individual agencies. I think it’s really important that a baseline of agency-based meetings never go away as they provide a foundation that keeps everyone in a routine of rigorous performance management.
That said, there are opportunities to pick certain issues and bring outside stakeholders into the discussion. We started with CleanStat, which includes multiple city agencies that play a role in keeping the city clean. We then added GunStat, which brings in non-city criminal justice partners focused on coordinating efforts to reduce gun violence.
Mayor Stephanie Rawlings-Blake has overseen the addition of DVStat, which focuses on reducing domestic violence, as well as the recent addition of WatershedStat, which focuses on improving the water quality of the Inner Harbor and surrounding watershed. We’ve also had “V2VStat” sessions focused on Mayor Rawlings-Blake’s Vacants to Value Initiative to eliminate blight.
These meetings are a further evolution of the stat model in that we bring in strong non-profit partners to help address the issue. In his November 2012 Performance Leadership Report, Bob Behn refers to this model as PerformanceStat 3.0 or CollaborationStat (http://www.hks.harvard.edu/thebehnreport/All%20Issues/November2012.pdf)
2. When implementing a CitiStat program, do you recommend a phased-in approach, working a couple or few departments at a time?
I would recommend a phased-in approach based on the resources you have to dedicate to the operation. The CitiStat analyst is the most important “tool” in the whole process, so it will depend on the size of the team you have. Regardless, I think it’s smart to start with a few priority departments, get people used to the process, show agencies that this is a process that you are committed to, and build from there.
3. What methods do you use to collect data? What resources are dedicated to this process? Is it a side-project of existing staff or do you have a dedicated office? What does that look like, thinking of the nuts and bolts like data collection, analysis and reporting? Who does this work?
For data collection, we still heavily rely on Excel to collect data across the agencies just like we did 13 years ago when Governor O’Malley (then Mayor) started the program. The resources that are used are based within the agencies, typically clerical staff where this is a side duty that is a natural extension of the data entry they already do to track operations.
They will pull data from the systems they use to manage their operations and enter summary data into an Excel spreadsheet. This is then sent to one point person in the agency who gathers data from the various services that the agency provides, aggregates it into one excel workbook we call the “template,” and then sends it to an analyst in the CitiStat Office.
For analysis and reporting, we have a dedicated team of five analysts in the Mayor’s Office who take the data from the agencies and use it as a baseline to produce a memo that includes charts, graphs, and tables, highlighting various aspects of the agencies’ performance the previous period. This memo serves as the agenda for our CitiStat meeting that occurs every four weeks for each department.
Each analyst is assigned a portfolio of agencies that they work closely with in order to learn about operations and prepare the memorandum that the Mayor’s Office receives.
If you go to our OpenBaltimore site at data.baltimorecity.gov and search “CitiStat” you can see sample templates we use to track performance. The templates are dynamic documents. We may decide that measures need to be added, changed, or removed from one CitiStat meeting to the next.
I think one key to our long-term success has been that we kept the operation low-tech. This has allowed us to remain very flexible with what we look at to evaluate performance. We spend a large portion of our time on CitiStat Tenet 1, accurate and timely intelligence, because all the others fall in line if you get the first one right.
If we collectively decide a set of metrics misrepresents an operation we change them or get rid of them. If we need to temporarily track a set of operational metrics that are “in the weeds,” we add it to the template to monitor until we see consistently improved performance.
At the center of all of this is the CitiStat analyst. They are the most critical piece to making the meetings productive and the performance measures relevant. They put an incredible amount of thought and energy into preparing for the meeting and without their hard work the process would suffer tremendously.