When Uber launched in China, access to computers, the Great Firewall and slow mobile connections all made it difficult to onboard new drivers online. One Uber engineer built a WeChat app. The messaging app WeChat basically is the internet for many people there, Ed Baker, the former lead of Uber’s growth team explained. Drivers used it to quickly answer onboarding questions through a dialogue on their mobiles and were ready to roll.
“Which was amazing, until WeChat started blocking us, because they were owned by Tencent, which is a competitor,” Baker explained. He told the story in a new video from “The Scaleup Offsite” published by the venture firm, Greylock Partners.
Baker shared the stage with former Airbnb growth lead, Gustaf Alströmer (now at Y Combinator) and Greylock’s Josh Elman.
Here are five tips the two former growth chieftains offered the crowd:
Figure out a product’s north star metric
Baker worked at Facebook before Uber. There, the north star metric was monthly active users. That one number is a way for a member of the growth team to decide if they are working on the right thing or not. At Uber, it was weekly trips. This worked well for Uber, because it’s a two-sided marketplace. They needed both drivers and riders, but weekly trips ultimately measured both at once.
Alströmer remembers a time when not everyone at startups thought about growth. Growth was someone’s job, but not everyone’s job. He found that this had changed by the time that Airbnb was new. “I think really great companies, they really have everyone really thinking about metrics,” he said.
Calculate the lifetime value of each user and then spend it to get them
“On those big platforms like Google and Facebook, the opportunity to do free growth is getting smaller and opportunity to do paid growth is getting larger,” Alströmer said.
Baker agreed. “Uber would have kept growing if we’d never done any paid acqusition,” he said, but the more people are using it the more valuable it becomes for everybody. Therefore, it makes sense to spend the money to get people on.
Uber actually had a data team working to figure out the lifetime value of each new driver and rider in each city. That number changes depending on the ratio of the two, so by calculating it from place to place they were able to figure out what to spend money on and how much to invest.
Elman, by the way, said he’s heard that Instagram ads are seriously underpriced for their results right now (#hottip).
Run experiments, complete with a control group
“Every two weeks we get in a room and we show experiments that we’ve shipped and gotten significant results on in the last month,” Alströmer said. An engineer would get on stage and describe the control group and the intervention for the experiment. Each time, he would let the audience weigh in on which side should show the most change. There was always a lot of debate, which proved that a team can’t know until they try it.
“There’s not necessarily a correlation between how much work you put in and how much of a step change you make,” Baker concurred. At Facebook, he said, they changed Japan from the least viral country to one of the most by simply changing one word on the website. When a new person joined, Facebook always prompted new users to “invite” new friends. When it changed the language from “invite new friends” to “announce you’ve joined” (in Japanese, obviously), Japan started to blow up.
Translate your product
We wrote about this recently. Google and Udacity have been putting out guidelines for internationalizing mobile apps. “It’s a huge growth driver,” Alströmer said.
Walk through the new user process personally
Baker completed the process to become a driver. He walked out of his house with the app in driver mode, turned around and went back inside. Driving for the first time made him too nervous. He saw the same story in the numbers for lots of first time drivers. They realized that first time behind the wheel was critical to driver retention and started running experiments to make sure they had a positive experience that first time out.