What I Learned From Building An App For Low-Income Americans

The brief was simple: build a tech product for low-income American households. It wasn’t simple at all.

What I Learned From Building An App For Low-Income Americans
[George Washington Bridge: pisaphotography via Shutterstock]

I was lost in the Bronx. It was my first week as a Significance Labs Fellow, where my job was to create a tech product for some of the American households who earn less than $25,000 a year. In 2013, 45.3 million Americans lived at or below the poverty line, which for a family of four is $23,834.


Another fellow and I spent an hour on the subway from Brooklyn to do a user interview. This area of the Bronx has no cafes or shops, only the odd cluster of fast food joints around a subway station. Yellow cabs don’t drive here. We couldn’t find the address. Google Maps kept moving the location further away. Coming from the technology world, we were confounded when that technology failed us. Eventually, a car service dispatcher told us that the address didn’t exist.

To some extent technology has failed low-income Americans too. Developers don’t build apps for them. Growth hackers ignore them. At Significance Labs, I learned a lot about how low-income Americans live and use technology but also about its limitations, and my own.

It’s The Scarcity, Stupid

Our first week was spent in some of NYC’s poorest neighborhoods interviewing all kinds of people. We talked to a devoted teenage dad in Washington Heights, a school aide in Brownsville, and an undocumented Mexican immigrant who had built a good life for her family on 25 years of babysitting and cleaning jobs. When I asked what her ideal job would be she said “computer programmer.”

The product eventually built by my team was for housecleaners. My colleagues targeted the underbanked, elderly Android users, first-generation college students, and food stamp applicants.

Every person we met had an intensely individual story, but common themes emerged. Like most New Yorkers, our interviewees were busy. Many juggled multiple jobs, and sometimes school, with family responsibilities. Like other Americans, they often traded off time and convenience against cost.


It also became clear that inequality isn’t purely about income. It’s about information and status and opportunity. If you look at the dollar amount, my own income as a freelance writer probably wasn’t much higher than some of my interviewees, but I still had resources like educational credentials and social capital which many of them lacked. One of the reasons that graduation rates among low-income first generation college students hover at around 10% is that they don’t have the “college knowledge” taken for granted by their peers.

Living on a low income translates into other forms of scarcity: of power, information, respect, opportunity, time, health, security, and even of sleep. Our job was to build a piece of technology which could increase our users’ stock of at least one of those resources.

Your Users Won’t Trust You

A few years ago I interviewed a Mexican impact investor named Álvaro Rodríguez Arregui. He explained that impact investors need to be very clear about their motives. “Do you want to do good, or do you want to feel good?” he said. “It’s much easier to feel good by giving away meals to starving kids in Sudan, but you are not going to solve any systemic problem in the world by doing that. This is business, and business is messy and you have to make hard decisions.”

I often didn’t feel good. Sometimes that was because I had to make hard decisions, sometimes because people didn’t understand why we were building mobile tech for domestic workers. All the teams at Significance Labs worked for significantly below market rates. They left jobs, or in my case, even countries. That makes it all the more unsettling when your potential users misunderstand or mistrust you. They have good reasons.

When doing nonprofit or volunteer work it’s all too easy to congratulate yourself for taking on the work at all.


My team worked closely with housecleaners to build Neatstreak. We assembled a panel of “superusers” who tested multiple versions and suggested features we implemented. Cleaners were often delighted just to have someone ask them about their work. People rarely do.

Nevertheless we often had trouble persuading housecleaners and other domestic workers to come to interviews, even though we paid $25 per hour, which was higher than their regular hourly rate. They didn’t know us and it looked too good to be true. Low-income Americans are often the targets of scams which advertise fake education credentials or applications for government benefits.

In the last week of the program one of Neatstreak’s superusers mailed me to say that he felt slighted. “I’m starting to feel like when corporate America uses the little guy for ideas and then forgets about them,” he said. “I was excited and ready to be hands on but instead feel used for my ideas.” During a user testing session with a group of Spanish-speaking cleaners, one of the testers gave a speech to the others about how we were a company (Significance Labs is a nonprofit) trying to take advantage of them. When building for low-income users you have to work harder to win their trust and to demonstrate your product’s value.

There Isn’t Always A Technical Solution

Technologists are problem solvers. It’s tempting to either jump in too quickly with a technical solution to an intractable systemic or human problem or to be discouraged by its difficulty. One of my first interviews was with a 21-year-old father of two, Angel. Angel was in foster care before he turned 1 and had been in trouble with the law as a teenager. What he really needed was a steady job which would provide him with an income for his family. No mobile app I could build in three months was going to deliver that.

Another issue was impact versus scale. Should we try to solve a smaller problem for a large number of people or have a bigger impact on a smaller group? Angel had attended Green City Force (GCF), an impressive program in Brooklyn where low-income young people do six months of national service related to the environment and are prepared for sustainable careers. GCF had a huge impact on the graduates we met but this kind of “high-touch” program is not where technology excels. Our best bet might be to create a little more breathing space for a large number of people.


My colleague Jimmy Chen, for example, built a mobile app called EasyFoodStamps to do the first stage of the application for food stamps, saving people hours of standing in queues at the food stamp office. When you lose a day’s work or have to get a babysitter to watch your kids in order to apply, that really makes a difference.

Furnishing a naive technical fix is the software equivalent of building a well in a developing country which the locals have neither the motivation nor the skill to maintain. You have to understand the whole context. For example, housecleaners prefer to be paid in cash (so mobile payments were out), mainly use text messaging, and sometimes don’t want to reveal professional information online, especially if they were undocumented.

What Next?

The U.S. has a serious inequality problem. The top 0.1% of Americans own more than the bottom 90%. Technology in many cases has made that inequality worse by eliminating jobs or replacing them with more insecure ones. Disruption is all very well when you are one of the beneficiaries. The tech business has a moral obligation to see what it can do to help.

But I am also convinced that there are sustainable, if not wildly profitable, businesses to be built on providing valuable services to low-income Americans. At Significance labs, essentially we made bottom-of-the-pyramid products for the developed world. Nearly one in four New Yorkers reply on food stamps and 40,000 more apply every week. Multiple companies chase their dollars outside the food stamp office. We estimated that housecleaners work in 20 million homes in the U.S. These are big numbers.

One of impact investor Rodríguez Arregui’s investments in Mexico is Finestrella, a successful startup which developed a set of algorithms to assess the creditworthiness of people who don’t have an official employment history, bank account, or credit rating, in order to offer them a mobile phone plan which costs much less than pre-paid. Two Silicon Valley VC funds with no impact investing mandate, Storm Ventures and Bay Partners, have also invested in Finestrella.


Maybe the best long-term solution is to train a new generation of developers and designers from a low-income background to build their own solutions, but that’s easier said than done. People on low incomes already lead a precarious life juggling multiple, low-paying, no-benefits jobs or government support. The last thing they want is more risk of the kind that is involved in launching a startup.

On the other hand, it’s striking that all six Significance Labs Fellows are zero or first-generation immigrants. Jimmy was born in China. Margo grew up in Ghana. They went on to Ivy league schools and jobs in companies like Facebook and LinkedIn but their families know what it was like to live a very different life.

Many of the housecleaners I met were already entrepreneurs. Our office cleaner at Significance Labs, Jason, employed five or six people in his cleaning business while also holding down another full-time job. The best thing about my time at Significance Labs was meeting incredible people like Jason and Angel. The most fun I had last summer was sitting in a room chatting to housecleaners.

About the author

Lapsed software developer, tech journalist, wannabe data scientist. Ciara has a B.Sc. in Computer Science and and M.Sc in Artificial Intelligence