My job is too boring. It's raining too much. I have too many ideas.
It seems like every winter something in me snaps. I begin yet another "personal project".
Ryde started in the winter of 2013. I had a good job. I made video games. I got paid way too much. But I was bored. I needed a challenge to keep me occupied and game design didn't have one for me.
I was hanging out with my friend Grant (also a transportation biker) and we were talking about how annoying it was that the cycling tracker Strava had to be manually started and stopped. We loved the idea of timing our commute, but could never remember to pull out our phones every time we left work. We also found battery consumption to be a major barrier to everyday use.
The idea came when I realized that I could use geofences to automatically start the GPS. I could then record the user's ride until they stopped moving and upload it to a server. Just like Strava, except without the pesky start and stop requirement.
Going in to the project I never felt that this would be a blockbuster product, but I did feel there was potential for sustainable revenue. I could sell access to the data my software would collect. Unfortunately I decided to build Ryde first and then research the value of that data later...
Step 1: Automatic Start and Stop
I had never worked with the geofencing API before so I read a bunch of documentation. Eventually I started to implement the basic functionality of the app.
Here's how the first version worked: let's say the rider gets home from her commute. Her phone records her location and drops a geofence around her. If she leaves the radius of the geofence Ryde is launched in the background and flips on the GPS. If the rider stays stationary for long enough Ryde turns off the GPS and sets another geofence around her.
This plan worked pretty well. The app started in the background, recorded stuff, and even flipped off the GPS when the ride had ended. It also took 4 months and hundreds of test rides to fully sort it out. Debugging a suspended app in iOS is ridiculously difficult.
At this point I started using the app myself. I liked it. w00t
Step 2: Social
I wanted the rider's data in order to build a business. But I needed to give users a reason to want to share it with me. (I'm just being honest)
So I decided to gamify the app. Users would give their rides a start and end point like "home", "work", or "grocery store". This would allow a rider to compare their similar rides to each other. Riders could compete against their own times, share rides with friends, or just clock how many miles they had biked.
I used Parse for my backend (because it's a magical unicorn from internet land) and I was up and running in a day or so.
The trickiest part was packaging all of that location data into a file and saving it up on the server. Hack, hack, hack and a few days later it's all working.
At this point I added some friends and my wife on Testflight. Yay! I had a few users. For about a week. Then they all uninstalled the app because of battery consumption.
Step 3: Build a Team
I had a prototype and I had a few users. It was the perfect time to try and fill out my team.
Since I was building the app off of Parse I didn't need another engineer working on the backend. But I did need a designer that could add the level of polish necessary to stand out against Strava. Peter Nudo and I worked together at two different companies for about 3 years in total. We have great design chemistry and he was into the idea.
Peter put together some great mockups for the app and a sweet looking logo. The app and product was really starting to look polished. I loved showing it off to friends and people at networking events.
But it still had battery issues to tackle.
Step 4: Low Power Tracking
One day in July I was hanging out with Grant and mulling over this battery life issue. Suddenly, I had a crazy idea. Instead of running the GPS the whole time the rider was riding I could record several points along their path. I could later run those points through the Google Directions API and get a pretty good approximation of the rider's route.
This dramatically reduced power consumption and worked pretty well, but had some problems. The main issue was sometimes the routes I got back from Google Directions had arbitrary turns and loops. I figured out that a GPS location at an intersection could be easily misinterpreted by Google Directions. Google might think I was on the east/west road when in fact I was on the north/south one.
While I had some solutions in mind, I never solved this problem. It was just too early to put that much time into robustifying the tracking system.
I convinced friends to reinstall the app. This time the reaction was much more positive. For a week or so. Until it broke.
Step 5: Learn A Neat Lesson
Around this time I found Eric Lee at Bennett Midland. Eric and his company had done a transportation study in New York City on the economic impacts of the parking buffered bike lanes. He was nice enough to chat with me about what I was building.
The phone call was great and super informative. Also, I found out my data wasn't worth shit. It turns out cities all over the country already pay people to stand on street corners collect the same data for $10 an hour once or twice a year.
So I killed Ryde. And a bottle of Bulleit Rye. Because I was sad as fuck.
Step 6: Seek More Pain
Ryde may have been a crappy idea for a business, but I learned a valuable (and probably ridiculously obvious) lesson: prove your model first. Finding a dataset and spending a day on analysis can save a lot of work.
However, there was a silver lining. I identified a core schism in the biking audience: those who bike for transportation and those who bike for recreation. Strava, like most bike products, was targeted towards recreational bikers. But transportation biking was wildly underserved in the mainstream bike industry. But why?
Next time, before I built anything, I would look for data. I needed to figure out if the big bike companies were missing out on what I felt was a great opportunity or if they had found the only sustainable business model in the US bike industry.