2012-10-19

Want Proof That Market Fit Is Everything? Test Your App In The Slums Of Sao Paulo

For the Stanford-educated founders of Emprego Ligado, creating a successful app in Brazil required dismantling every assumption about the target audience.



Connect a smart team with a huge market, and the results can be explosive growth. But fitting the two together can mean months of validation and product testing--sometimes leading to results you'd never expect. “The temptation is to do it as easily as possible--put up a solution and get Google AdWords out there,” says Derek Fears, the American cofounder of Emprego Ligado, a job-search platform for Brazil's underclass. “It may be an elegant solution that makes sense to you, but isn't what people want or need.”

Dee, Fears, and Rosenbloom

Fears is talking about Sao Paulo's blue-collar labor market, which he and his cofounders believed could make leaps in efficiency if it were brought online. The problem: People here get jobs by word-of-mouth, so neither the job seekers nor the jobs ever make it to the web. Building a solution would take months of on-the-ground research like the kind ethnographer Tricia Wang does for Nokia or Microsoft.

Emprego Ligado, which translates to "connected job," launched in Sao Paulo this summer with the aim of connecting unskilled laborers to jobs close to home via SMS: Workers text the system when they need a job, and they system texts back with jobs in the area that match their preferences. It sounds simple enough, but arriving at a working model required dismantling every assumption the founders had about their target market.

“The password field, when you’re signing up--I would assume most people would think this means they should create a password,” says Fears. “But watching people completing forms who've only ever signed up for email, that's not evident for them. So they'd assume it was asking for the password to their Hotmail account and they’d get confused.”

Ask someone working in a Sao Paulo supermarket where they live, and they’ll always give you an answer you’d never expect--up to 20 or 30 kilometers away, explains Jacob Rosenbloom, another cofounder. “There are cities within cities here, and while there are plenty of blue-collar jobs, the distribution of labor here isn’t exactly optimal.” Poor job fit means the churn rate can be anywhere from 60% to 90% for customer service jobs, and workers are frequently late because of their commutes. Sao Paulo’s air quality suffers as much as its workers. “We knew this would be difficult, but if we can help the bottom of the pyramid, and there's a green element--that we really like,” says Fears.




He and his two cofounders, Rosenbloom and Nathan Dee, decided to tackle the problem with good old-fashioned sociological research, which they used as a basis for a simple working prototype. “We are an engineering team with degrees in statistics,” says Fears, “but really the key was following the MVP model.”

The first challenge was choosing a medium. Web penetration is growing in Brazil’s underclass thanks to cheap Android devices, but it's still not utilized consistently. Mobile penetration is over 100%, but when Fears and his team began doing market research, they found that 85% of the operational workforce here uses prepaid talk and text with no data. “Once we realized SMS was comfortable for everyone, then we hired developers in Chicago to build a prototype in Drupal,” says Fears. But connecting their site to an SMS gateway required partnerships with the mobile carriers here--not exactly something you can register for online.

They reached out for help to Marcelo Sales, the founder of an incubator in Rio de Janeiro who founded a mobile services company. "It's such a relationship-based business culture--this is how things get done here," Fears says. "You take people out to lunch, you get it done. That's what's cool about the Brazilian guys in the technology ecosystem--they're so willing to help you. Everyone huddles together because there are no resources down here.”

With carrier partnerships in place, they needed to figure out how to match applicants with employers. They did intensive focus groups and one-on-one interviews with 40 human resources managers in the operational sector. After 30 days of ethnographic research inside employment agencies, they still needed to learn about the other half of their audience: the applicants. But Sao Paulo is huge--counting the outlying suburbs it approaches 20 million people. They contracted local market research companies, which pointed them to the Zona Norte, or north zone, a populous area with high congestion. “We think of this as a flow problem. We had to do the calculus of what's the right neighborhood to start with--that's how we mapped out the market.”

The team passed out fliers asking people if they were looking for jobs, and used their limited web framework to enter basic details about each worker’s CV. They spent time in metro stations, going to schools, talking to people about to graduate and enter the job force, and tried to find patterns in their search for employment. In all, they interviewed 1,000 blue-collar workers, asking them dozens of questions about how they searched for jobs. “It was a qualitative on-the-ground search for information at a grassroots level,” says Fears, pictured at right. “We were focused on being in the market, with constant user and customer feedback.”




Once they had a pool of test users, they began prototyping a site where employment agencies could list jobs, which would be sent out via text message to candidates that fit the profile. “We showed wireframes to HR managers at companies, and we watched how they hired. We did focus groups where we ask them zillions of questions to find out their pain points. We showed them slides about the concept for the product. It was constant feedback,” says Fears.

Once their prototype was working, they connected to an SMS gateway and started texting some of the applicants they had recruited as testers on the streets. They ran multiple versions of products simultaneously to see which one got the highest engagement. The job seekers surprised them again.

“On the SMS side, one thing we did notice is that despite our attempts to automate everything for these users, a lot of time they were treating it as if there was a person on the other end answering their questions,” says Nathan Dee. “In the beginning we had to monitor SMS responses and constantly check for questions that users sent via SMS that our system wouldn't be able to automatically pick up. We realized we had to build in automatic responses that could identify patterns in messages and find trigger words.”

End users in HR companies also reacted unexpectedly to Emprego Ligado’s web-based dashboard for employers. “The terms and conditions were a big deal,” says Fears. “Most people in the United States click through and spend no time reading, but people here, for whatever reason--perhaps they're actually interested--would read it. They thought they were supposed to. They'd scroll down through pages and pages of legal mumbo jumbo--I can only imagine what they thought about it.” It was costing time in the signup process and killing their conversion rates. “Now the terms are just a link that you have to click on, which opens up a different window. We had to make it just a little more seamless.”

The reaction from users on both sides of the labor market has been overwhelmingly positive, which helped Emprego Ligado land funding from venture heavyweights like 500 Startups, Initial Capital, Rising Ventures, and Fortify. Fears says the lesson is simple: “The only way to figure out if there's a product-market fit is to get local.”

[Brazil Flag Image: Flickr user Gustavo Minas]


Article Tags: emerging