Setting Product Teams Up for Systems Change in Colorado

Impact

• We equipped the state team to improve the application experience by building client-centered product development and data practices, including how to leverage Qualtrics client survey data and Salesforce application metrics.
• We worked with the state team to develop both a data action plan and a product development action plan, including a customized an ‘Impact and Effort Calculator’ tool to help them make decisions for project prioritization.
• We launched a pilot for automated text message reminders for SNAP interviews in Boulder County. Based on our pilot design, Boulder County successfully adopted automated text messaging and has seen a 10% reduction in missed interviews.

The challenge

Colorado uses a system called PEAK that allows families and individuals to apply for multiple benefits programs that they qualify for all at once. There are a lot of great things about the application, including the fact that it’s mobile-first and that recent modernization efforts have led to a consistent user interface, both of which are important accessibility measures. But when we conducted user research, we heard that some people felt confused when they applied through PEAK and worried they would be denied benefits.

At the time, the PEAK team had no way of regularly surveying clients to hear about their pain points, and instead had to ask a third party vendor for ad hoc data. The team received a lot of suggestions for changes to the application, but had a difficult time forming strategies to prioritize those that would have the highest impact.

Our challenge has been to build internal capacity on the PEAK team, empower them to use data to make informed and human-centered design decisions, improve the experience of using PEAK for clients and rebuild trust in government systems.

Our approach

Our goal in all our work is to enact systems change—to not only improve the delivery of digital services, but also to empower government to continue innovation and modernization work long after our partnership concludes. In Colorado, we wanted to ensure that the state was getting feedback about their application from the proper sources, including clients in various counties and caseworkers on the frontlines. 

When we reviewed the flow of the PEAK application, we highlighted examples of application improvements they could make that would have a significant impact on the client experience. These included:

• Increasing trust in the application by ensuring PEAK follows plain language guidelines and trauma-informed best practices, refining the UX design of the flow to align with user goals, and aligning the application experience to reflect our best practices for a human-centered safety net. 
• Making the PEAK application more concise and transparent by reducing duplicative questions , pre-populating fields with available data, making sure clients are only answering questions directly relevant to their situation, and adding a completion progress bar to give clients reassurance about where they are in the process.

We worked together to identify pain points in the PEAK application process, find areas for improvement, and shared best practices for testing solutions on a small scale before deploying them. Our focus involved building frameworks that allow the team in Colorado to integrate data inputs from various sources and make informed choices about what to prioritize for both short- and long-term impact. 

Our main goals of this work included:

• Getting the PEAK team to a place where they could actively track client pain points and use this data to define priorities for PEAK
• Making sure that there is a smaller, more manageable backlog of PEAK changes that are prioritized based on client need and outcomes

We also worked with the state to begin some small scale experiments—starting with Boulder County, where we worked to demonstrate how automated text messages could reduce missed interview rates and improve communication with clients, while minimally burdening staff.  Boulder County had one of the highest rates of missed SNAP interviews in the state, and had experienced many operational challenges launching automated text messaging themselves. We aligned our texting approach with Code for America’s Texting Playbook to ensure our strategy reflected tested best practices in UX design, plain language, and trauma-informed care. We also launched an in-text survey for those who received our messages and conducted qualitative research interviews with select participants to gather feedback on the impact of the text messages. With this valuable feedback, Boulder County improved the text messaging pilot, ensuring that future text communications are informed by Colorado clients for their long-term texting plan.

Outcomes

We expect many of the outcomes to unfold over time as the state team implements these strategies and continues building internal capacity. Early indicators of impact include:

• The PEAK team is using the ‘Impact and Effort Calculator’ framework we introduced to evaluate and make decisions about the backlog of changes they want to make. 
• We trained the PEAK team to use Qualtrics (an experience management tool)  for analyzing client feedback for their end-of-application survey, and demonstrated how to use that data to prioritize future improvements. This represents the first time they’ll be strategically using client feedback to make decisions about their product. 
• As a result of our partnership, Boulder County implemented automated texting reminders for SNAP interviews and has seen a 10% reduction in missed interviews since. Clients who received text messages were three times more likely to say the process of applying for benefits was “very easy” compared to those who did not receive texts. Based on our five year projection, we believe that 2,700 more clients will be approved for benefits because of texting in Boulder.

We believe that improving government’s capacity to innovate and try small-scale experiments represents one of the best ways to craft a better experience for people applying for benefits. When government teams have the tools they need to try new things, prioritize the most meaningful projects, and hear directly from the people they serve, the solutions they craft work better for everyone.

Read more