How we approached designing a truly accessible service.

Product design, Code for America

In 2018, Jerry Brown, then out-going governor of California, signed legislation that opened up CalFresh (food stamps) to over 500,000 low-income seniors and people with disabilities.

When the State of California asked us to provide the official online service for CalFresh enrollment, we wanted to ensure that our service, GetCalFresh, would be ready to receive the influx of newly eligible applicants. We knew our clients’ access to food depended on our ability to get it right.

The services don’t come to you. You gotta go to them.

- Sacramento resident

All too often, it’s easy to be caught up in “accessibility theater” — that is, meeting a general checklist of accessibility criteria without really understanding if it holds up for the people in real life. To ensure GetCalFresh was truly accessible, we took an approach that we think is key to doing anything well:

  1. Take time to develop true care and understanding
  2. Embed learnings in both your work and culture
  3. Never consider the work done

Taking time to develop true care and understanding

For a value as important as accessibility, we did not believe any single method could help us to develop a full understanding. We approached our learning through three separate efforts: learning for ourselves, learning from others, and early prototyping.

My biggest expense is rent. I pay $700 in rent. I get $950 in income.

- Culver City resident

Learning for ourselves

We approached this project as we do for all our projects at Code for America — start with your user and their needs. We sought to answer some basic questions. Who were the people we would be helping? What are their lives actually like? What would it mean for them to successfully apply for CalFresh?

We had a team of three talented user researchers and two data scientists who led a discovery research effort.

A report put together by our data science team that helped us to make sure we were talking to the right balance of people during our discovery research.

Testing an early paper prototype with research participants.

The research team synthesizing findings from their trips throughout California.

Learning from others

It’s also important to recognize when to build on work of others. In addition to conducing primary research, we sought guidance from the established community of practice around accessibility.

A report put together by Knowbility documenting accessibility issues found in our design and code.

An SF Design Week event about Accessible Design Systems with presenters from Google, Lyft, and LinkedIn.

Early testing

Accessibility wasn’t a new concept for many of us at Code for America. Even before we got deep into research, we already had many ideas of what we could do to improve our service.

As we conducted early research, we took that opportunity to understand their ability to use our application. We had ideation sessions early and brought prototypes out with us.

I help her financially. Shopping, heavy lifting, I cook for her… I started to help her at a young age. She’s my mom so…

- Los Angeles resident

Embedding learnings in both our work and culture

What we learned

By the end of our research phase, we felt like like we had a reasonable grasp what we were working with.

A lot of our work was catching up to speed with the larger community of practice. We learned the difference between situational, temporary, or permanent disabilities and how accessibility benefits everyone. We learned the well-understood norms of how to make a product that accessible to this with visual or physical impairments.

We started by refining our color system both by reducing the colors used and making sure each color passed WCAG Level AA contrast guidelines.

Inspired by golden proportions, we defined a sizing system that would be used to define everything else in our design system including typography, patterns, and layout. This provided an easy to understand system for designers and engineers alike to ensure a clear and consistent hierarchy.

We simplified and strengthened the type hierarchy used across our service based on our spacing system. This helped to ease cognitive load by improving readability and made our service more accessible to screen readers. We also darkened the color of detail text to fix readability issues found in usability testing.

We more clearly outlined, increased the size, and bumped the contrast of interactive elements. We also used rounded corners and depth to distinguish buttons from inputs. Sizing, spacing, and rounding of elements are all based on golden proportions.

As a bonus, we also evolved our logo to be slightly easier on the eyes. Admittedly, this was for the designers more than anyone else =)

We also developed some insights that very few people talk about. For example, we saw first-hand the challenges that people with cognitive disabilities face — and how little we as a design community knew about how to help them. We also learned not to assume familiarity with basic mobile norms such as expanding accordions and scrolling, especially for older adults.

We simplified our layout to be fully left aligned. This helped to reduce cognitive load on people who already had high anxiety, stress, or mental incapacities.

We adjusted the language we used in order to reduce anxiety. For instance, we clarified why we were asking certain “trigger” questions. In another example, instead of asking for how much money the applicant had saved up, we simply asked if they had below a certain threshold (and why it matters).

Many of our older clients did not intuitively know to scroll. We added a conditional scrolling indicator that only showed up if we did not detect scrolling after a few seconds (and would disappear once they did). While many of us had advocated for a dedicated back button in the past, research evidence helped us to push this feature to the top of the queue.

A simple yet valuable insight we had was that that seniors and people with disabilities do not exist in a vacuum. More often than not, you’ll find a wide network of helpers who assist them with their needs. This could be an adult child or sibling who cares for and represents them, an IHSS (In Home Supportive Services) worker who helps their perform daily activities, or the lobbyist at the senior home that connect them with services.

True accessibility means making services accessible to the people who care for them.

We added screens that acknowledged the presence of “helpers.” By passing this information through to the county, we were empowering the people who aided seniors and people with disabilities to play their part.

Making accessibility a part of our culture

We had clear improvements to make on GetCalFresh but we also wanted to make sure that our learnings were not lost to the broader organization. In additional to improvements to our product, we made sure to embed the learnings into our operations and culture so that future features and products could benefit from our work.

Our researchers put up a walkthrough of research insights for everyone in the office to see. By seeing quotes,

Our team partnered with Meals on Wheels to conduct usability testing. We made sure to provide opportunities for everyone on the team to come along.

Never considering our work done.

On June 1st, 2019, CalFresh opened up to the 1.3 million people who lived on Supplementary Security Income in California. Our service experienced a five-fold increase in the number of applications coming in that week. As of August 2019, we have helped over 200,000 SSI recipients apply for CalFresh. Our rate of completion and satisfaction remain steady.

Over the last few months, we have developed a much deeper understanding of accessibility at Code for America. We have helped others in the organization understand why accessibility is important in the real world and made it easier incorporate accessibility into their work.

Of course, our work remains unfinished. The last key point to doing accessibility well is to remember that your work is never done. Even at this moment, we are continuing to evaluate the effectiveness of our work through surveys and continued usability testing. We have a long way to go to get our copywriting down to a 5th grade reading level. We plan on deepening our understanding of true accessibility by partnering with Lighthouse Labs and working directly with more people who are vision-impaired. Our understanding of how to better serve and collect feedback from people with cognitive disabilities is still painfully low.

As is often repeated, we must remember that design is a verb, not a noun. There will always be new additional challenges to tackle and biases to uncover. We are excited to keep our users at the center and push the envelope for what it means to be truly accessible.