top of page
Louisiana Benefits Digital Assister cover.png

Louisiana Benefits Digital Assister

Code for America

Building an online application for SNAP (Supplemental Nutrition Assistance Program), WIC (Women, Infants, and Children), and ECE (Early Childhood Education) for residents in the state of Louisiana.

My project role:

Lead UX Designer

Duration:

1.5 years

Platform:

Mobile and web

1. What's the problem space?

Before I joined the Louisiana team at Code for America in January 2023, the project team had already completed several months worth of discovery and prioritization work in partnership with our primary government partner, Louisiana's Department of Children and Family Services (DCFS). One of the opportunity areas they had identified was improving the experience of applying for social safety net benefits.

​

For clients (how we refer to the families we are trying to serve), the team found they largely struggled with:

​

  • Creating an account and logging in through LA CAFÉ, Louisiana's current (and pretty old school) benefits application website

  • Getting confused by questions that are written in complicated ways, leading to incorrect answers or drop-off

  • Wanting to use a mobile device to apply, but having trouble navigating the flows

 

For caseworkers (the government workers who process applications and act as the "middle people" between clients and state agencies), their challenges looked like:

​

  • Additional time and energy needed to support clients in filling out their applications

  • More back-and-forth time when a client submits an incomplete or inaccurate application

  • Higher volumes of help center calls or foot traffic into in-person offices

​

This is how our team landed on the decision to build and pilot a new online benefits application website, internally referred to as the LA (Louisiana) Digital Assister.​

Screenshot 2024-10-12 at 12.23.08 AM.png

A screenshot of the home page of LA CAFÉ, Louisiana's current online benefits application website. Daunting, isn't it?

2. Understanding the landscape

The Digital Assister project officially kicked off in May 2023, starting with me and our team's 2 researchers doing a full assessment of all the questions are asked in LA CAFÉ.

 

Our lead researcher built an impressive Airtable database, where she documented each question along with tags, such as which benefits programs the question was being asked for.

 

Meanwhile, I gathered design references, patterns, and best practices from other benefits applications we've built at Code for America, including Minnesota's MNbenefits and California's GetCalFresh (GCF). 

Screenshot 2024-10-12 at 12.49.02 AM.png

A screenshot of the Airtable database of the whopping 260 questions in LA CAFÉ'.

Screenshot 2024-10-12 at 8.24.13 PM.png

A screenshot of the Figma file where I gathered other benefits application designs to reference.

3. Designs and research: Take 1 (Original scope)

After our lead researcher completed the Airtable database and I finished gathering design resources, I cross-referenced both to build the first version of the full Digital Assister flow.

​

For v1, I kept the designs mid-fidelity because it was easiest to copy paste existing designs from MNbenefits and update them with the appropriate Louisiana Department of Children and Family Services (DCFS) branding. The goal of these v1 designs was to illustrate a general idea of the overall flow, so we could start having feedback conversations internally on our team but also with our external partners.

A screenshot of a section of my v1 designs - my designs are on the top row, compared across the questions in LA CAFÉ screenshots on the bottom row.

image (20) 1.png

A side-by-side comparison of LA CAFÉ's screen where they ask for people in your household, compared with the simplified redesign I created using Code for America's design patterns and component library.

After creating these v1 designs, I went through 4 rounds of reviews:

  1. An internal review with our policy advisor

  2. An internal review with our core project team, including our Product Manager, Researchers, and Programs Manager (this person was our main liaison with our government partners)

  3. An external review with our primary state partners at DCFS

  4. An internal review with our engineers

​

Post-review feedback, I prepared interactive Figma prototypes for usability testing with clients, as well as supported application process shadowing and concept testing with caseworkers.

 

One important aspect of our design and research was to consider multilingual access - for the particular clients in Louisiana, we prioritized designing in both English and Vietnamese. We were lucky to have a bilingual Vietnamese researcher on our team, who helped translate the v1 design prototype and led research interviews in Vietnamese with both clients and social workers. 

The Vietnamese prototype we used for the first round of usability testing.

Screenshot 2024-10-28 at 11.54.28 PM.png
Screenshot 2024-10-28 at 11.56.39 PM.png

4. Introducing a new scope: WIC and ECE

Initially, we had planned to include all 4 benefits that currently exist in LA CAFÉ: SNAP (Supplemental Nutrition Assistance Program), FITAP (Family Independence Temporary Assistance), KCSP (Kinship Care Subsidy Program), and LaCAP (Louisiana Combined Application Program). 

 

However, we had to shift gears after Code for America partnered up with New Orleans Collaborative for Early Childhood Research (CECR) and Tulane University under a National Science Foundation grant. This new partnership meant we now had to incorporate WIC (Women, Infants, and Children) and ECE (Early Childhood Education) programs into the Digital Assister, something both Louisiana and Code for America have never done before!

​

Because of our new scope of work, we de-prioritized FITAP, KCSP, and LaCAP programs (which made up for a minimal number of applications on LA CAFÉ at the time, anyways) in order to have time and resources to include WIC and ECE.

​

Our team quickly regrouped to flesh out the project phasing rollout:

​

  • Phase 1: Multiple person household, SNAP only

  • Phase 2: Multiple person household, WIC and ECE included

  • Phase 3: Experiment treatment logic for WIC and ECE built

  • Phase 4: Single person applicant, SNAP only​​​

5. Designs and research: Take 2 (Adding WIC and ECE)

After a series of kickoff meetings with our new external partners including folks at Tulane University, New Orleans Public Schools (NOLA P.S.), and the Women, Infants, and Children (WIC) department, I got back into design gear.

 

I started by cross-referencing the list of WIC and ECE questions against the existing designed SNAP questions to see which we were already gathering data for and which we needed to add to the designs. I worked closely with our external partners and our backend engineer to ensure the data mapping made sense.

Screenshot 2024-10-22 at 12.40.57 AM.png

A spreadsheet I created and collaborated on with our backend engineer to cross-reference data.

After figuring out which additional questions I needed to add, I moved forward with creating screens. 

​

A key part of our collaboration with Tulane University, NOLA P.S., and WIC partners was to test different UI/UX approaches of allowing families to apply for WIC and ECE. The goal of this experiment was to figure out which way, if any, was more effective at increasing WIC and ECE enrollment. Together, we designed and documented an experiment treatment plan that boiled down to different treatments depending on:

​

1. Parish location: if the family that's applying is located in Orleans Parish or elsewhere

​

2. "Links" V. "Apply" treatment: the families that qualified for WIC and ECE were randomly sorted into "Links" or "Apply" experiences -​

  • Families that got the "Link" experience were asked in the middle of their SNAP application if they'd be interested in WIC and ECE, and if they said yes, they got a link to start their WIC and ECE applications at the end of their SNAP application​​

  • Families that got the "Apply" experience could answer a few additional questions in their SNAP application to get their WIC and ECE applications started

​

I drafted several design iterations of this experiment design in close partnership with our Tulane research partners, being especially conscious of content and interactions in order to encourage families to proceed as much as possible.

​

We also completed a second round of research with the new WIC and ECE designs, to get feedback from clients about usability and content.

Screenshot 2024-10-27 at 9.52.07 PM.png

A high-level overview from a presentation deck of the most distinct design differences between the 2 "Links" V. "Apply" experiment designs.

6. Iterating, collaborating, and more iterating

Over the course of 8 months (v1 designs in August 2023 to pilot launch in March 2024), I led the design of the entire Digital Assister including the WIC and ECE experiment designs. This was definitely not without the help of many, many collaborators!

​

  • I had weekly syncs and working sessions with the 2 researchers on our team

  • I had weekly 1:1s with our Product Manager and Programs Manager

  • I initiated weekly project check-ins with the core team (myself, researchers, Product, Programs) and created a spreadsheet to track project timeline + milestones

  • I met with both our frontend and backend engineers to hand off designs and collaborate on data mapping

  • Our team had weekly meetings with DCFS partners

  • And also weekly meetings with Tulane partners

​

Through these many, many meetings - I was able to keep the rest of my internal team and our external partners informed on my design progress, work closely with researchers to identify research opportunities early on, and make sure we were fulfilling all of our project goals especially with the Tulane work.

Screenshot 2024-10-27 at 10.07.44 PM.png

A screenshot of the project planning spreadsheet I created to keep our team aligned in our different workstreams.

7. Prepping for the big launch

Another big chunk of design work I led was working closely with our Programs Manager to prepare for our pilot launch.​ What this specifically looked like meant I was creating and helping present...

​

  • Slide decks to inform caseworkers and other DCFS staff about the upcoming launch

  • Flyers for DCFS staff to print and post at their in-person SNAP offices

  • Trainings for caseworkers and help desk workers on what info they might need to know to help families who are applying

8. Metrics and outcomes

The Digital Assister website ended up being live from March 2024 to September 2024 (around 6 months). In July 2024, we had a total of ~11,000 SNAP applications, with 115 applicants interested in ECE and 595 applicants interested in WIC. On average, people took 23 minutes to complete their application and 89% of people who provided feedback in our short survey at the end of the application rated the experience as "easy" or "very easy".

 

At the time of me writing this case study (October 2024), the team is still working on wrapping up the WIC and ECE experiment treatment results and handing off our final deliverables. In the meantime, you can browse through my final designs in this publicly shareable Figma file, meant to encourage anyone to view and access our design knowledge!

bottom of page