Building Digital Services in a Shifting Policy Environment

Our work with California on Pandemic EBT food assistance shows the challenges and opportunities of building policy and implementation tools at the same time

Meeting people’s needs through benefits programs is challenging in the best of times, and in moments of crisis, it’s even harder.  The COVID-19 pandemic caused a historic child hunger emergency, and at Code for America, we’ve been determined to help states build and maintain the systems they need to distribute food assistance to families who need it—rapidly and accurately. 

In March 2023, the California Department of Social Services (CDSS) approached Code for America for a partnership on their Pandemic Electronic Benefits Transfer, or P-EBT, food assistance program. P-EBT 4.0 was the final round of this pandemic-era benefit, and was designed to include assistance for families with children who were homeschooled or attended virtual school during the 2022–2023 school year (and began this method of schooling specifically due to the COVID-19 pandemic). The majority of benefits had to be distributed before September 30, following the expiration of the federal Public Health Emergency (PHE)—which meant we had to do rapid work to ensure that all those who qualified for this benefit received it. 

What made the work more complicated? The fact that during the P-EBT 4.0 implementation process, the policy was being written as we were building the application. Here’s how we did it, and what we learned. 

Working quickly to stand up an application

For students who learned from home, and therefore didn’t have access to free and reduced-price school meals, California provided money for food on an EBT card. Each state participating in this program could design their application process differently, subject to approval from the U.S. Department of Agriculture (USDA) Federal Nutrition Service (FNS), which funded the program. We originally hoped to have an application designed by the end of May to allow a three-month application window, but shifting federal requirements meant we had to re-scope the project several times.

Digital whiteboard with stickies identifying the biggest risks for clients using the application.
We talked a lot about the risks of this application—meaning the difficult places that might stop someone from getting all the way through.

We spent a lot of time debating how we could adjust income questions to reduce barriers (since we know that income questions are often where people get stumped and give up), and how to use information we already had to make sure the application was as easy to fill out as possible. Because the policy that determined P-EBT eligibility and rules distribution was still being written, we had to work around changing requirements—but this also left us with the chance to provide feedback on policy language and surface opportunities for improvement. We were able to bring learnings and best practices from other products like GetCalFresh and MNbenefits to improve income questions, particularly for households with more than one job or people who are self-employed. 

We found that the spaces between the hard rules of the policy—where things were undefined or negotiable—allowed us the opportunity to improve the experience for clients at certain points in the application.

Building the plane while flying it

The process of spec’ing out application requirements, talking about different use cases and applicant personas, mocking up screens, and designing flows actually surfaced policy nuances that weren’t yet considered when the policy was being written. We found that the spaces between the hard rules of the policy—where things were undefined or negotiable—allowed us the opportunity to improve the experience for clients at certain points in the application. Here are some things we were able to accomplish by connecting policy to its implementation, in partnership with CDSS:

  1. More client-centered interpretation of future income: We recommended an interpretation of income rules that allows the state to assess eligibility based on an estimate of future income if it’s lower than what the client made last month, based on existing rules for CalFresh (SNAP).
  2. Giving self-employed clients a default standard deduction to save them time: We advocated to allow people the choice to take an automatic 40% standard deduction from their gross pay if they are self-employed to save them from having to calculate their taxes and operating expenses. 
  3. Increased maximum number of students in household: The number of students allowed on the application was initially limited to six. However, by working with CDSS to understand that this was based only on the number of students allowed on a registered homeschool application, we realized the number of students didn’t need to be constrained by this. As a result, we were able to allow up to nine students to be put on the application to accommodate households with more children. 
  4. Providing helpline numbers: There are two different phone numbers for the state P-EBT helpline, one that recipients can call into for assistance and one that only places outgoing calls to recipients regarding questions and eligibility determinations. We worked with the state to provide clients with the phone number they’d be contacted from to mitigate the risk of these calls being perceived as scams, and to support applicants in effectively screening out other calls. 
Income screens on the application
We used learnings and research from other products to inform our income flow designs. Pictured: Mock-ups of the self-employment standard deduction (left), future income estimate (center), income summary (right) screens.

Partnership brought us through

The process of building and debuting P-EBT 4.0 taught us a lot about working in a rapidly changing policy environment—both in terms of the challenges it surfaced and the opportunities it exposed. By working closely with CDSS, we were able to pivot where necessary, spot places where we could make things easier for our clients, and build capacity within government to take these processes forward to new projects. Every design review on Figma and Zoom call to parse through policy contributed to the collaborative environment necessary to build a tool that served applicants equitably and met their needs.

Building towards the future

While P-EBT functioned throughout the Public Health Emergency to keep children fed, we know that child hunger is an enduring challenge. We also know there’s huge potential for helping more families if we can take the lessons learned here about accessibility and user experience in safety net programs and apply them to work at scale. That’s why we’re taking what we know from working on P-EBT and applying it to our Summer EBT work.

A new federal safety net program launching in 2024, Summer EBT has the potential to distribute $3.6 billion in food benefits to families across the country and represents an enormous investment in the well-being of the almost 30 million children who will qualify for the program. Summer EBT can do for children what P-EBT did—but on a permanent basis. We’re looking forward to meeting the moment, using the lessons we’ve learned from P-EBT along the way. 

Want to learn more about our continued work on Summer EBT? Check out the Summer EBT Playbook.

Related stories