The most important part of any ride-sharing app’s interface is the map. You use it to pinpoint your location, to see your planned trajectory, and to watch as your driver’s little car makes its way to you.
This interface—pioneered by ride sharing company Uber when it launched in San Francisco in 2010—works well on high-end phones in grid-based cities where the internet and GPS always works. But it starts to break down in some places around the globe.
That’s what Uber’s design team found as the company expanded to Southeast Asia, Latin America, and the Middle East (the company now operates in 600 cities in 63 countries).
“As we were entering these markets, a lot of the original hypotheses that led to the core flagship app weren’t holding true,” says Shirish Andhare, head of product for Uber India. If the internet isn’t fast, GPS isn’t accurate, and phones don’t have a large amount of storage, a complicated app like Uber doesn’t work very well.
Andhare and his team, who are based in India, started noticing that users weren’t successfully booking rides, particularly in countries dominated by Android phone users. To understand why, his team traveled to a handful of cities in India and in Latin America to meet users face to face and understand what their issues were with using Uber. Their insights led to a complete revamp of the core Uber app, called Uber Lite, which launched as a pilot in India in summer 2018. Now, a year into the launch, Uber Lite is available in 30 countries, in English, Spanish, Portuguese, and Arabic (Hindi is coming soon). Since the end of 2018, the number of Uber Lite downloads has increased 300%.
It’s a success story for a company that has run into problems as it has expanded globally. In some cases, Uber has bowed to local ride-share apps, as it did in Southeast Asia in 2018, Russia in 2017, and China in 2016, where a local competitor acquired Uber’s China operation. Earlier this year, Uber bought its biggest competitor in the Middle East, Careem, for $3.1 billion. In others, the regulatory landscape has proved too challenging and it’s simply shut down, like in Denmark.
Uber Lite seems designed to directly combat some of the problems that Uber has faced internationally by making the experience of booking a car easier for riders. By actually talking to its users, the company realized that in Android-dominated countries, which tend to be non-Western, many phones were not designed to run an app as large as Uber’s. In many of these places, internet networks and GPS aren’t as reliable. And less tech-savvy users were getting overwhelmed by the sheer amount of text, which was harder to read outside on lower-quality screens.
“We found that as much as people had intent to take an Uber, the performance of the app wasn’t optimal,” Andhare says. “They were also operating in network conditions that are not like Market Street in San Francisco. I joke, 4G is so oversubscribed it feels like a 2G connection. Close to a third [of riders] were operating in these spotty conditions. They were more likely to cancel rides because they were frustrated.”
For Sri Jalasutram, a product design lead at Uber who led the design process for Uber Lite, the biggest decision was to ditch Uber’s map interface entirely. Some of this insight came from a user named Maria who Jalasutram spoke to when he visited Brazil. She ran her own business and was a savvy user of Whatsapp and Gmail, but she was confused by the Uber app and if she needed a ride she would ask her daughter for help. One of the biggest problems was the map.
“When she opened the app she had to figure out where she was—entering pickup location and expecting GPS to resolve it was not optimal,” Jalasutram says. “She was on a lower-end device and the GPS chips in these devices are not as precise as what you’d expect. The blue dot would jump around.” The search field didn’t help either, since she was used to thinking about her city in terms of landmarks, not addresses.
“We realized that we were building this entire app around a map. It was really not helping here,” Jalasutram says. “There was so much information on the map, and all she cared about was, ‘can I just get a ride, I don’t want to type so much.'”
As a result, Jalasutram decided to ditch the map completely and start by asking about current location instead of relying on a GPS to determine it. The Uber Lite home page has a half-blue, half-white screen, with large text saying, “Select your pickup point.” There’s a list of frequented locations and other local landmarks, including nearby hospitals, train stations, bus stops, airports, and malls, so that the user doesn’t need to type at all.
Instead of using the map to indicate the status of someone’s ride, the team decided to use bright, bold colors with high contrast ratios and larger text sizes that would make it easier to read in harsh daylight on lower quality screens. “While you are requesting in Lite, you see a big slash of blue to indicate comfort and safety,” Jalasutram says. “As soon as you are matched, the whole screen turns green. When people saw that, they understood green, success, I got a car. That gave them more confidence in using the app.”
From the testing the team did, the results were positive. “Nobody missed the map,” Jalasutram says. “Nobody was asking us, ‘Where did the map go?'”
Another challenge the team learned about was around affordability. Previously, the Uber app listed its different options, like Pool, through a horizontal swipe interface. But when they spoke to people in their homes, everyone was asking for more affordable options. The team realized that many of these users weren’t accustomed to swiping horizontally, and didn’t realize that there were more types of rides they could take.
As a result, the Uber Lite app lists all of the ride options in a vertical format, from least expensive to most expensive. The change was so significant that the design team decided to bring it over to the core Uber app as well, which started rolling out in November 2018.
Since launching Uber Lite, Andhare says the team has noticed that Lite riders tend to opt into these more affordable options twice as much as core app riders. “What we’ve learned is there seems to be a correlation between the cost of phone and affordability,” he says.
Andhare’s team is already using this insight to think about new affordable services Uber can offer. It’s part of the thinking behind Uber Bus, which is currently piloting in Cairo, Egypt, with its own separate app. Uber Bus works just like a public bus, with regular pickup and drop-off points, and Andhare thinks it’s a kind of “premium economy” experience for people who want a small step up from public transportation.
The Uber Bus app is built entirely on top of the Uber Lite platform, utilizing much of the same design, like the block colors and the lack of a map. Andhare says it took the team only a fifth of the time to launch this new app because they were able to build on Lite’s backbone. “The investment in Uber Lite over the last year resulted in amazing velocity for a completely new experience that we could provide for a new modality in an emerging market,” Andhare says.
Another goal in redesigning the app was purely technical: to strip down the app so that it would only take up five megabytes of space (compared to the 40 megabytes of the U.S. app). That meant cutting out any unnecessary elements of the code—a process that Andhare jokingly calls going to the “Uber Lite gym,” where the engineers would negotiate over every kilobyte they needed. Instead of making a connection with the network every time a user does something in the Uber app, the engineers found a way to package many actions together, bundling them so that they could make fewer calls to the network. The decision, Andhare says, led to the app being 1/20th less network-heavy.
Some design elements didn’t make the cut, like the font, which Andhare says cost about 200 kilobytes. “In another app that would be nothing,” Andhare says. “But in our world it was a lot of currency. We pushed back and negotiated with the brand team and the result was that we basically adopted the native font that was available within Android.” That meant the team didn’t need to spend any precious storage space on fonts—the result was an app that was 80% lighter than the core app, with most of the same functionality.
Of course, building a single app for billions of people located across multiple continents can’t take into account all of the cultural differences at play. The team tried to accommodate that too, particularly when it came to social norms around calling and texting drivers. Jalasutram says that users in India would always call their driver immediately to let them know where they were, but users in Brazil and El Salvador didn’t feel comfortable having their phones out on the street, so they’d send text messages to their driver instead. He tried to accommodate these subtle differences in the app. “Depending on where Lite is being launched, the teams there and users there have the flexibility of having options they prefer culturally,” he say. “If you want to text, Lite has option of just letting you text.”
The process of building Lite was an exercise in abandoning assumptions about how people use technology and asking them directly. Andhare says that cultural differences will continue to inform Uber’s product decisions going forward. “The notion of inconvenience is very different in different parts of the world,” he says. “Where we as citizens of Silicon Valley may shudder to walk a certain amount, people in India and other parts of the world are very much okay with it. But they value the certainty of the ride and comfort and they’re willing to deal with what we’d consider inconvenience. We’re experimenting with understanding those ideas.”