Master in Design Engineering @Harvard
Team: Taylor Greenberg Goldy, Jake Schonberger, Mengxi Tan
Feb 2019 - May 2019
Streamline form filling
PROVIDING SOCIO-ECONOMIC MOBILITY BY MAKING BENEFIT ADOPTION EASIER
The US benefit system is there to support the lives of thousands who struggle to provide for their families. The application process should be easy, but in reality, it requires a massive amount of research into what you qualify for, filling out of complex forms, scheduling appointments, and locating the correct offices to go to. 20 websites, 126 pages and 67 hours later, you might end up receiving the benefits you qualify for. But, it doesn’t need to be this way. inForm is a website that serves as a platform for low-income benefit enrollment. Using our platform, users will be able to learn what benefits they might be eligible for that they aren’t already involved in and pre-fill all documents needed for the pre-enrollment stages, interview prep guides, and maps of where to go to finish the in-person enrollment.
PEOPLE QUALIFY FOR BENEFITS BUT ARE NOT ENROLLED
According to the Mass Legal Services, roughly 740,000 households in Massachusetts qualify for low income benefits, but are not fully enrolled. In dollar amount, it means $93,000,000 unclaimed potential benefits per month in MA alone.
THE PROCESS IS COMPLICATED, SCARY AND LONG
After weeks of research, we came to the conclusion that this benefit gap in enrollment is due to 1 the scary and complex process to figure out the potential benefits one may qualify for and 2 the complicated benefit enrollment procedure (form-filling, interview, providing proofs, renew enrollment) one need to go through after the eligibility screening.
STAKEHOLDERS INPUT FROM INTERVIEWS AND MEETINGS
During the research phase of this project we talked to economic stakeholders such as the office of the mayor of Boston, the mayor's office of women's advancement, and the economic mobility lab. We interviewed low income community stakeholders such as food pantries, low income housing groups and startups fighting food insecurities. We also interacted with community engagement groups such as the Cambridge Public Library, the Boston Children's hospital and the Harvard Equitable Tax Initiative. Through these meetings, we gathered these takeaways:
> Building outside the current system is faster and much more impactful
> The low income community is generally open to collaboration
> The benefit process is so confusing people rarely understand what is available
> The low income community is a tight network with high smartphone usage rate
> Libraries and community centers are seen as safe places
> Hospitals are a good touchpoint since 85% of families bring kids for checkups
inForm is divided into two easy steps, Eligibility & Pre-Enrollment, which provides you with all the information you’ll need to enroll in benefits.
> During the Eligibility step, we’ll ask you 8 simple questions & show you which benefits you’re eligible for and how much you might receive.
> During the Pre-Enrollment, we will guide you through our Masterform, which is a single form that is generated based on the benefits you are eligible for. We’ll use this consolidated information to pre-fill all the forms you might need. We will THEN generate an enrollment packet that contains personalized pre-filled forms, planners, & maps to your benefit centers. We’ll also send automated reminders to alert you when benefit providers need more information or it’s time to re apply.
In the information design stage, we mapped out the information flow and modules of our system. For the first step of eligibility determination, the user will input answers to the eight eligibility questions, which then goes through the eligibility calculator decision module to create an output list of eligible benefits and potential amounts. The system will then generate a custom application masterform based on the specific benefits this user is qualified for. For the second step of pre-enrollment, the application information will recycle information from the eligibility round, and ask the user for additional data to fill in the application masterform. Finally, the masterform will auto-populate the actual benefit application pdf form to provide users with a filled package in the end. See below the system flow chart.
To build out the information system, we collected eligibility and enrollment information from the official websites  of the seven major Massachusetts benefits and built out a custom database. We were able to narrowed from 22 to 8 questions in the eligibility process, and from 58 to 17 in the pre-enrollment form filling process, reducing 2/3 of the redundant process.
STEP 1 : ELIGIBILITY
Auto-generate based on eligibility
Database auto-populate to PDF
STEP 2 : PRE-ENROLLMENT
Based on the information design structure, we implemented the user interface design in Figma. In a repeated process of prototyping, wireframing and checking-in with stakeholders, we iterated and improved our interface design for more than ten rounds. We transitioned from mobile to website as a platform to make our tool more accessible to the users, adjusted specific language usage (e.g. from “your form” to “my form”, from “encryption” to “protect”) to give a sense of inclusion and reflect the habit of our users, and updated UX flow to guide users more clearly throughout the process.
It was important to us to incorporate behavioral insights in the development of the platform to nudge the applicants at the right moments and ensure retention when they feel like they want to give up. We mapped out what the process would look like for the user and at what moments we intervene for a nudge, for data collection or data sharing and what the outcomes are.
Functional prototype is available for testing at: https://bit.ly/2XASqht
Complete code can be viewed at: https://github.com/mengxit/nudgeplus
The final technical build reflects the exact decision making steps of the seven major benefits to reflect the eligibility and qualification value status. The eight input variables will be passed onto the seven benefit functions once user finishes submitting, and calculation will be finished and displayed within 0.5 second. For user testing, we added encrypted logging as a feature to record the inputs, outputs, behaviors and access platforms of our users in real-time safely.
EVALUATING THE IMPACT
When thinking about how to measure impact, we looked at the potential for life & societal improvement from our users perspective, and from the government & economic perspective. The goal of our product is to a) provide a nudge to the low income users who might be eligible for benefits and b) provide those same users a way to act on their nudge and follow through with the application. We were able to implement a version of the nudge portion of the product, in the form of a potential benefit calculator.
Average Current Income
$708 / month
Average Benefit Estimated
$856 / month
Monthly Budget Increases
Average Food Beneft
$188 / month
Food Assistant National I/O Multiplier
1.79 / dollar
$337 / person
Only 1% of our target user
$2.49M / month
According to our user data, the typical users report about $708/mo in income with an average of 1.5 children. On the benefit side, we were able to estimate that these people are generally eligible for $856 in government benefits, which breaks down to $188 in food related benefits, $518 in housing, $73 in childcare and $71 in utilities related benefits.
Additionally, combining this information with economic benefit indicators, like those used by the USDA, we can gain an understanding of the potential impact of the collection of any single dollar driven by our platform. From an economic perspective, we used the USDA Food Aid National Input/Output Multiplier (FANIOM) to estimate what the potential impact on the local economy we could drive if our users were awarded the benefits we estimated them eligible for.
It suggested that for every $1 in food related benefits adopted by citizens, $1.79 are put back into the economy.
Thus, we could say that for each of our users, if their benefits were fully adopted, they would contribute over $300 in economic stimulus. Furthermore, if 1% of our target population (around 7,400 people) were to use our product, that number could rise to around $2.5M in economic stimulus per month.