How Data Can Drive Performance for Social Services
A Colorado agency is using the approach in an unusual way: to orchestrate the efforts of service-delivery partners.
Most government agencies don't deliver services directly, by themselves. Typically, they deliver services jointly with, or through, other entities. The challenge is aligning a range of partners -- often with different objectives, timeframes and interests - to achieve a common goal.
The Colorado Department of Human Services (CDHS) seems to have accomplished just that via a data-driven performance-management framework it created back in 2012 called "C-Stat." It is based on many of the principles pioneered a couple decades earlier by the New York City police department's award-winning "CompStat" approach and replicated since then by many government agencies to manage their direct operations and services.
What makes C-Stat different is that it is used to orchestrate the state's 64 counties and multiple contractors in delivering more than two dozen social-services programs such as food assistance, income support and mental health services. One-third of these programs are delivered by counties, one-third by contractors and the remainder by state employees.
A new report for the IBM Center for the Business of Government by Melissa Wavelet, the former head of performance and strategic outcomes for CDHS, describes the approach taken to align these various partners to achieve a common goal, using C-Stat as the organizing hub.
After Reggie Bicha was appointed to lead CDHS in 2011, he put the C-Stat framework in place to help achieve an elusive goal: bringing the department out from under a long-running court settlement agreement by processing recipient applications for its food- and cash-assistance programs in a timely manner at least 95 percent of the time in every county, and sustaining that level of performance for 12 consecutive months.
CDHS' leadership team undertook a multi-pronged approach to engage the agency's partners as part of the C-Stat management framework. C-Stat, writes Wavelet, creates an expectation that everyone is "engaged in dialog that is based in the data and focused on how to get better and improve performance." But collecting and sharing data can create a sense of vulnerability for those affected. The key was to build relationships that focused on learning from the data, not using it to punish or belittle partners.
The first step was to create shared ownership over the data being collected, and the commitment to use it to make informed decisions came from Bicha himself. County and contractor partners becoming engaged. "They became consumers of their own performance data," Wavelet writes, "and asked themselves and CDHS staff about their data and their own performance compared to others."
Much of this was enabled by the creation of monthly C-Stat reports and, subsequently, an online dashboard, that combined data from nine separate systems and made it easier to analyze. These one-stop, easy-to-use tools allowed counties and contractors to shift from the position of spending a lot of time compiling and analyzing data to a point where they could use it to learn what worked, what didn't, and what needed some fine-tuning.
As important as data and tools for analyzing it are, there was more to the CDHS performance effort. The agency created a peer-to-peer learning network so counties could share successful practices. This not only generated goodwill between the state and counties but also increased attention to performance in small and medium-sized counties. The C-Stat framework "has certainly changed how we work with our program partners at CDHS," said Dan Makelky, director of the Douglas County Department of Human Services. "While we haven't always agreed with every C-Stat measure or the goals, we are learning together and performance is improving."
As the performance system matured, CDHS also created a number of recognition awards for individuals, teams and counties. There is a distinguished performers award for counties that achieve 75 percent of their goals every month for an entire year, for example. And there's another award for counties that are most improved in their population category. As of the end of 2018, about 450 county staff had been recognized in their own communities with an award.
And the court settlement agreement that helped trigger the C-Stat approach? As of April 2016, statewide performance exceeded the 95 percent goal across all five types of recipient applications for the first time. In January 2017, the Denver District Court terminated the long-running settlement agreement. Families across the state were getting access to benefits as quickly as ought to be expected. There's little doubt that C-Stat helped make that difference.