Digitizing the EBT process at Farmers Markets to make it more seamless and efficient
Client: Community Farmers Market
Timeframe: August 2018 - December 2018
Teammates: Kelsie Belan, Rachel Feinberg, Bang Tran, myself
My roles: UX researcher and designer
Platform: Mobile (Android, iOS)
Electronic Benefit Transfer (EBT) refers to the debit card that people below national poverty level use to redeem their benefits at farmers markets (Wholesome Wave Georgia doubles the amount of EBT funds spent). To narrow down the scope of our project, we chose to focus on Community Farmers Markets in Atlanta. The current EBT system processes funds at the farmers markets by using tokens. Tokens make the system very inefficient due to a number of problems associated with their use, including:
difficulty of tracking
proneness to human error
time spent counting
bulkiness of carrying
Also, the current reporting system used by managers is redundant and inefficient.
How can we redesign the current EBT system to provide a better experience for all the user groups involved and enable them to spend more time on things that they love doing?
During the research phase, I conducted three interviews and formulated the questions for them, created the personas and task analysis diagram. I also came up with one of the three solutions, illustrated a storyboard as well as fabricated low-fidelity wireframes for it. I led the key design decisions for all the three prototypes. For the testing, I switched between being a notetaker and facilitator/interviewer for different sessions. I redesigned the main app according to the Material Design Guidelines while incorporating the feedback we received.
What I learned
I realized that designing for multiple user groups at the same time limits the room for brainstorming if I’m to satisfy everyone’s needs. Additionally, designing for simplicity is challenging and requires a lot of effort and thought. Collaborative brainstorming with individual designs diverging, and then converging can be time-consuming, but it allows us to create much stronger overall designs by taking into account the strengths of different designs.
Even though I thought we had executed our user research well and invested a lot of energy in designing our prototypes, I realized we could always identify areas that needed improvement. Testing the prototypes with users and experts helped us validate our design decisions and identify those areas.
Secondary research, market visits and user interviews
We conducted secondary research about farmers markets to gain background information and know the surrounding regulations to understand the constraints. We visited Community Farmers Markets to learn about the context first-hand and conducted structured and semi-structured user interviews with the three user groups involved in the EBT process - Market Managers, the Vendors, and the EBT Customers to understand the tasks performed by them, their needs and pain points. We did five interviews with each user group. By analyzing the research data, we were able to determine a significant process constraint, draft personas for managers and vendors, sketch a user profile for EBT Customers and identify critical user needs.
Process constraint - We learned that all our digital system would be used AFTER the EBT cards are initially processed by MarketLink, the ONLY app approved by USDA
Based on the data we collected from user interviews, we performed a task analysis of the overall existing EBT system at a Community Farmers Market to help us understand the different steps it encompasses and where the pain points lie.
With our user groups’ needs in mind, we sat together and brainstormed potential solutions. We realized that having a connected digital application for Vendors and Market Managers was the only way we could satisfy all of their needs (easy to use and efficient reporting). This meant that we could brainstorm different solutions for EBT customers while the digital application for the other two user groups stays constant. We came up with three solutions.
QR Card: A reloadable card that carries the EBT user’s transferred funds and gets scanned at Vendor booths.
Basket Wallet: A basket that can be loaded with funds and scanned by Vendors; it is also used to carry purchased items.
Customer Application: A customer-facing app that lets EBT users view, manage and complete transactions.
For each solution, every team member presented their storyboard and design ideas. We converged on everyone’s ideas by outlining a step-by-step narrative for a solution, weighing the pros and cons of steps that we differed on. This activity allowed us to create an enhanced overall solution. To map the users’ journeys and explain the experience to users, we sketched storyboards for our final solutions.
For each solution, we discussed the high-level design concepts and followed that up by quickly designing low-fidelity wireframes to test them with users for features and information architecture.
We held feedback sessions with two managers, both with experience at the farmers market as EBT users and vendors. Thus, they could be considered as experts within the EBT system. Due to time constraints, we kept our sessions focused on concepts of our solutions for the EBT customers only. We walked them through our concepts using storyboards.
QR Card: Both managers felt that it was the best solution. They liked that it was very simple and straightforward. They thought it would be the easiest to explain to new EBT customers and vendors as the idea matched up with existing mental models.
Basket Wallet: They didn’t like this idea because of the constraints that a physical basket imposed such as lack of storage space and their long-term durability.
Customer Application: They felt that adding a new step of signing up for an account and downloading the app made the EBT system less accessible.
High-fidelity interactive prototypes
Based on our feedback from the Market Managers, we decided to move forward with the QR Card idea. However, we also decided to create an optional app for the EBT customers as it would provide them with an increased access to checking their current balances. As the QR Card number is the identifier, there is no need for a sign-up process.
Key interactions and user flows -
1. Market manager app
Add Market Bucks to a customer’s card
By digitizing the process of providing funds to EBT customers, we mainly managed to eliminate all the problems associated with the use of tokens mentioned earlier. Managers would load funds onto the EBT customer’s Market Bucks card by scanning it via their app.
View, edit and export reports
The app automatically records all EBT transactions, saving a great deal of time on counting and entering token totals. At the end of the market, managers can view the auto-generated reports, make any necessary edits, and export them for sharing.
2. Vendor app
Charge Market Bucks to a customer’s card
The flow of this interaction is similar to adding market bucks to a customer’s card in the manager app. Along with eliminating problems associated with the use of tokens, the app also gets rid of the vendors’ problem of having to round down on sales totals and thus, they wouldn’t have to lose any money.
View sales history and track reimbursements
The app allows the vendors to track their EBT sales in real time as well as their past sales which they can refer to when they get reimbursed.
3. Customer app
Check card balance
EBT customers can use this feature at their discretion as frequently as necessary without having to sign up. The main advantage offered is reduced cognitive load for the customers. They don’t need the app to view their balance though. They can do so via manager’s or vendor’s app whenever they add funds to their card or have it charged respectively.
View purchase history
EBT customers can use this feature if they link their card with the app. This feature enables the customers to view and track their purchases. These serve as digital receipts which aren’t provided by vendors at the moment.
Check out the interactive prototypes by clicking on the buttons below!
All of our tests were moderated and conducted in-person.
We used this usability tool to evaluate the feelings and thought processes of users as they performed tasks on our prototypes. We also noted what troubles users are having using the prototypes by watching them interact with those. Due to time constraints, we conducted the test with two managers, who as mentioned earlier, can be considered as experts within the system. We used a within-subjects testing method and presented the three apps in different orders to the managers to avoid order effects. Managers were asked to perform a set of tasks for all the three apps while thinking aloud.
To make up for the lack of the number of end users we could gain access to, we decided to leverage the expertise of evaluators instead by conducting heuristic evaluations. We asked three experts to evaluate all the three apps for a similar set of tasks as the think-aloud while referring to a list of Nielsen’s heuristics that they were knowledgeable about.
Mock Farmers Market
We built a fake farmers market environment to simulate the context of our system and understand how it would operate under conditions of a real farmers market without impeding on it. We asked three HCI students to act as EBT customers and perform a couple of tasks for the current as well as new EBT system. We followed the tasks with a series of questions to evaluate the qualitative measurements of efficiency and comfort in using both the systems and compare those results to understand how our system improves the way that EBT Users would shop at farmers markets. Two team members served as the manager and vendor.
Issues identified and proposed solutions -
1. Market Manager
Issue 1 - The process for selection of a report is inefficient as there are repeated entries for the first column. Solution - This page has the potential for reducing the manager’s cognitive load by breaking the process into two distinct steps - select market name first and then date from a shorter list of dates corresponding to the market selected.
Issue 2 - For inexperienced managers, it wouldn’t be immediately recognizable what the columns are about. Additionally, not having column names makes the format inconsistent with that of the report the managers would eventually export. Solution - Naming the columns would facilitate recognition on one level or the other for all managers.
Issue 3 - Tapping outside the pop-up closes the pop-up. This could lead the managers to lose the changes they made inside the pop-up if they were to tap outside accidentally. Solution - Tapping outside shouldn’t close the pop-up. Issue 4 - Our system is supposed not to use “EBT tokens.” Nevertheless, somehow we managed to slip up. This could lead to potential confusion among the managers given the transition from tokens to cards. Solution - Replace the term with “EBT Funds” or “Market Bucks.”
Issue 5 - We added the colored dots to indicate the addition of their corresponding tokens against a particular vendor. However, these dots could be misinterpreted as the number of tokens added. Solution - To avoid this issue and also help managers confirm the number of tokens they’ve added, we could display their number inside the corresponding colored dot.
Issue 6 - EBT customers would be viewing these numbers from the Manager’s phone by leaning over the table from a certain distance away. The current size of numbers highlighted would make it difficult for the EBT customers to read them. Solution - Increasing their size would make the numbers easier to read.
Issue 7 - Our system doesn’t enable the managers to refund customers in case they err while adding market bucks to a customer’s card. Solution - A functionality of reimbursing EBT customers would also allow the managers to repay EBT customers for the amount they’ve left unused at the end of the market. This would ensure that the managers and customers have something to fall back onto.
2. Vendor app
Issue 1 - The prototype violates one of the critical usability specifications we had identified earlier - accuracy. Tokens require the vendors to round down on transaction amounts. Yet another slip-up, our prototype doesn’t allow the vendors to enter the exact amount of their transaction with a customer. Solution - Allowing the vendors to enter amounts accurately, down to the right decimal point would solve this problem.
Issue 2 - Displaying the EBT customer’s balance to the vendor could be a major violation of the customer’s privacy. While we intended this as a way of showing an EBT user their balance, they might not want the vendor to know this information. Solution - Text the EBT user about their updated balance after they’ve completed a purchase with any vendor.
3. Customer app
Issue 1 - When checking the card balance, an account is not necessary for checking the balance. Having the button at the bottom below the sign-in might imply that the user should create an account. Solution - Move the “check card balance” button above the form.
Issue 2 - The system provides no feedback when the card information is saved other than a small checkmark at the bottom of the page. This could make it easy for users to save or delete their information accidentally. Solution - Providing a confirmation popup would make the EBT customer ensure that they did intend to save their card information on the app.
Issue 3 - Displaying entries in the form of a list doesn’t make it evident that the entries can be tapped on to view more information about them. Solution - Displaying the entries in the form of cards would make it apparent that they can be tapped on.
Redesigning the prototype
I redesigned the Market Manager app according to the Material Design guidelines. I also made certain changes to the original design to solve the issues identified during testing.
We can test the redesigned prototype to check if it solves the issues identified during the testing phase.
Photo Courtesy for market and tokens in overview - Jenna Shea Photography