Building my first 100,000 users app: Lessons learned
First, some lightning fast background information:
Rio de Janeiro state (like the city, but bigger) has a draconian( and well branded) law that goes after anyone who might be DUI. Every night from 10pm to 5am, police build checkpoints around critical traffic bottlenecks and take turns breathalyzing drivers. The traffic jams can get massive on Fridays.
At first applauded, the law received its 1.02 patch, alcohol tolerance went down to 0. Ate one of those licor filled chocolates? You might be sleeping in a cell.
Besides removing drunk drivers, the state had a groundbreaking epiphany: It could use the LeiSeca checkpoints for informal tax collection. Recent statistics published by the LeiSeca crew showed 5% of all apprehended drivers being alcohol related. The remaining 95% were drivers with overdue fines or behind on yearly driver fees.
It's 3am and your car is towed and fined.
So almost a year ago I got together with the guys behind the well known @LeiSecaRJ twitter account, a feed dedicated to ratting out LeiSeca checkpoints in real time.
Between beers we thought up what the MVP should do: Show users where the checkpoints are (physical reference, address, lane…) complete with a map.
Here are the lessons learned:
Building the simplest yet functional of your app is key. The hypothesis had been validated by the twitter account, but we still had to figure out what was necessary for users to use the app instead. We were, after all, giving out the same information.
Scaling (stability) is mission critical when it's a utility app. What if someone lost their car because we couldn't deliver the information reliably? In the end we were really just competing with our Twitter account, could we outbeat Twitter's stability?
Platform is important, use it to your advantage. Remembering the importance of having the app's value proposition validated, we made v1.0 a paid iOS app. We paired the smaller iOS market penetration with a $1.00 paywall. As our userbase grew steadily we knew: 1) people wanted it. 2) people were willing to pay for it. 3) how much we were growing per month and could figure out what the next steps could be.
What we hadn't expected: Because the checkpoints only take place after 10pm, our servers would simply get hammered every night. Until we got our act together we had nights where the servers would go down up to seven times a night.
Lots of small percent improvements are better than anything else. In one week I:
- A/B tested the landing page, Twitter prompts, App store page.
- Installed Twitter Cards so that our now 1 million followers can see a [Download] button for their relevant platform.
- Had our practically dormant Instagram account promote the iOS app.
- Reached out from a separate Twitter account and asked users if they were willing to recommend the app to a friend (Do things that don't scale).
We hit #3 late September 2014. Not only did we have a permanent increase in user growth but being on the top3 gave us momentum.
Figure out how little you can make. The free android app meant we had to learn how to make money with advertising. A newbie, I went with the Walmart of Ad-networks, AdMob. We made way less per month than we expected but were pleased that it was enough to cover all expenses. This gave us a lot of room to experiment.
Figured out that it's not all banners/It's not one size fits all. Our Admob revenue came from banners located at the bottom of the list. The staggeringly low revenue told us this: People open your app, look at the list for a few seconds and close the app. In my opinion, banners are only remotely worth it if the session time is long enough to warrant it. ex: 30 min game vs 5 second glance.
You can only be successful with ads if you know your app inside out.
Resumo LeiSeca's users are >90% cariocas (Rio native), maybe hyper localized ads will pay more? What about interstitials? I played around, questioned everything, and learned loads. Know how little you can grow/make with your app and run weekly experiments to grow it.
Congratulations to the @LeiSecaRJ team for 1Million Twitter followers and over 100,000 active users.
P.S: I just launched Bebedouro, a Reddit clone for startups to discuss best practices. Go check it out!
