March 2015-August 2015
Today I lay to rest my first software app (web app) and what nearly was my first startup. Four months after launch, SurgeAlert has been displaced by the superior Surge Notifier out of California. It was a fun project and my consolation prize is that I may have been the first person in the world to push surge notifications to Uber drivers’ iPhones!
I hacked together SurgeAlert to exploit a big weakness in the Uber driver system. Uber doesn’t tell drivers who are offline that prices are going up, so they don’t know to get online! I think Uber wants to reward to drivers who stay out all the time and keep the market going. Contrary to popular belief, Uber apparently doesn’t mind that some rides don’t get filled. Regardless, the difference is pay is big — like $15/hr in normal times and $40/hr in surge times. So when I learned the API opened up in Fall 2014, I figured I could exploit the system. I only wanted to drive during peak hours, but I didn’t want to open the Uber app and check prices every 10 minutes — I was studying for law school. So I built the app and got surge alerts up to every 15 minutes — think random nightly concerts letting out, or sports games, or Terrapin’s beer parties, etc. It worked well and allowed me to boost pay up to $50/hr on some nights.
So I tried to commercialize it with a few friends. The app was cobbled together and it wasn’t in the app store, so it only worked for me. I envisioned a clean apps store product sold for $5 per month. But doing it right required mining data from Uber for locations around the country, storing it in a database, sending out alerts, and then building a front facing iphone app. That’s way beyond our ability, but we figured with could put together an SMS text service — drivers could just give us their phone number and preferred location, and we would send messages when the prices went up. We were almost finished when Surge Notifier dropped in the iPhone and Android app stores. It had more features and was easier to use, all for a $10 one time charge. And the kicker — the developer is some dude from SF that made the whole thing in a weekend just for fun. So that was the end of SurgeAlert.
- Anyone can learn to code, its just another language, but getting good at it takes a lot of time. It would have taken me months to do what the SF guy did in 2 days.
- While I right about the product fit, I think I picked a bad industry. Uber is too cool, so every developer wants to play around with the API. Next time I’ll try and build an app for some old industry that everyone ignores. That said, I’ll only build something that I actually need!
- Next up, QtrCall, a prettier version of this.
Here’s the final product:
And here’s how it worked:
First, I had to steal some code from a dude who posted a Los Angeles version on GitHub. Then I had to add the GPS coordinates for my city. Coding is super finicky — if you miss one period or space, the app just won’t work and you’ll have to find the bug. Finally, I had to register for an Uber API token. That’s how Uber keeps control — they could turn me off anytime.
Next, I had to run the code in the cloud on Heroku. It was free, either because I only had 1 app or it didn’t run very often. It produced this webpage with the surge multiplier — anyone could have gone to it to check the rate. https://surge-athens.herokuapp.com/
Finally, I had to figure out how to check the website constantly for changes and then send an alert to my phone. I rigged that together with this British guy’s software met to check for Ebay price changes ($1.20/mo), and this push notification app for iPhone ($5 one time). The software looked at my page ever 15 minutes and let me know if it changed from say “uberX 1.0” to “uberx 1.5”. Then it sent my phone an alert within about a minute.
Total project cost = few weeks of spare time + $10.
Value = extra $20/hr * 50 hrs driving = $1000.