Decoding Text Messages to Deliver Food Assistance

For anyone who has sent a text message on the go, it should come as no surprise that those texts are often riddled with typos, amusing autocorrects, and other errors that could affect readability.

Normally, all it takes to resolve this miscommunication is a clarifying text or phone call.

But, what if that text wasn’t just letting Mom know you’d be late to dinner or your boss know that you’re about to hit send on that important email? What if that text was standing between you and feeding your family?

From a user experience perspective, those innocuous typos could be the difference between being able to deliver a much needed service to our clients and being unable to do so.

The GetCalFresh team at Code for America has made applying for the state of California’s CalFresh food assistance program (known nationally as SNAP) easier. Using tools from consumer tech, the team has worked to make the process more user centered, data-driven, and iterative. From shortening a 45–55 minute application process to under 10 minutes, to making the application available in Spanish, and providing ongoing guidance (via email, SMS and live chat) throughout, the team has worked to address many potential pitfalls along the way to applying for food assistance. Already, we’ve made the process of applying for food benefits easier, more dignified, and less burdensome for more than 12,000 Californians.

As the client success manager for the GetCalFresh team, I help clients navigate the CalFresh application and enrollment process by answering client communication (emails, texts, and live chats in English and Spanish). Before coming to Code for America, I worked as a senior case manager with primarily Spanish-speaking immigrants. That meant I was meeting clients in person to assist with food and clothing assistance, medical care, and interpreting and translating needs, among many other things.

Now, through GetCalFresh.org, services previously delivered in person are being done over text messaging and online. Some of the challenges that I came across in person were taken to a whole other level when it came to providing those same services online.

Meet people where they are

GetCalFresh.org is the first fully functioning CalFresh application assistance tool designed for mobile devices. Spanish-speaking users filling out their applications using GetCalFresh can text with me, which is important, as over a quarter of U.S. households (27%) do not have access to a computer with broadband internet at home. The majority of low-income clients who do have internet access only have it through their smartphones.

But, for people who do not write fluently in Spanish, texting can be a challenge. Major spelling errors abound, to the point of near unreadability. To understand what something means, I have to read the sentence out loud (or quietly to myself) to get the phonetic gist of it, in order to piece together what the client is actually trying to communicate.

You may or may not recall from high school Spanish class, in Latin American Spanish the Hs are silent, and Ss,Cs, and Zs sound the same, as do Bs and Vs. Without the English glottal stop (or the pause between words, for those less familiar with their linguistic terminology), vowels blend from one word into the next. Add to that the informal language used in texting and typos common to SMS.

What that means is that a text like this… …may appear like this.


The former translates as, “Yes, thank you. It’s just that, I had made a mistake. But, now we explained everything. Thank you. Take care!”

The latter, if you put it through Google Translate, an imperfect tool to begin with, translates as, “If you like it, I would like to bake it, but it’s already pleasing.”


Code for America aims to make government services easier to access, through technology, for populations who need it most. And nowhere is that need more apparent than among those applying for food assistance. But sometimes, to truly reach people where they are, a human touch (and a not-insufficient understanding of linguistics) is what is needed.


And that’s where I come in.

Where do we go from here?

Decoding text messages to help people get the food assistance they need is just one step towards making government services work better for the people who need them most. Are you doing something innovative to meet people where they are?