Using software to analyse and improve my workout performance — Part 2

Mathias
Atlon
Published in
3 min readDec 7, 2020

Our brand new functional fitness workout app, Wodscribe, will launch on the App Store next week. We’re super excited, and thought it would be a good time to take a step back and look at how we got here. This is the second (and last) part of a two-part series. Read Part 1 here.

Evolution of the workout screen in Wodscribe from a very early beta version to the first public version.
Evolution of the workout screen in Wodscribe from a very early beta version (July ’20) to the first public version available on the App Store (December ‘20).

In Part 1 I explained how the idea of automatic workout analysis started during CrossFit Open 2018 and how I built the first prototype to prove that exercise types and reps can be inferred from movement data. Going from scrappy prototype to an actual product that people want to use is no small step though. The user experience needed a serious upgrade (well, to be honest, there was basically no user experience), and the machine learning model required a lot of workout data. Much more than I could possibly generate myself.

I started tackling the data requirement first since that was the largest of the two problems. In order to get more data, I had to make an app that could be used by other people with minimal explanation. Since Apple Watch has more than 50% market share, I decided to go with an iOS app. I have never developed for iOS before (my prototype was for Android), but how hard could that be? Well, iOS is quite unlike anything else, and especially the Watch has some quite interesting limitations that took me longer than I care to admit to understand and get under control. However, after a month of not sleeping very much, I ended up with a fantastically unimpressive (but functional) first version (left screenshot below). I high-fived myself and went to find some users!

Evolution of the Wodscribe workout history view.
Evolution of the Wodscribe workout history view.

Lucky for me, Aarhus CrossFit is full of wonderful people who were happy to try it out and help me gather some data. However, a new problem emerged. Very few of them had an Apple Watch. So I bought 15 Apple Watches and handed them out, and finally, data started to flow in. This was around half a year ago. Today we have more than 200 hours of workout data and almost 100,000 individually marked repetitions (with wall balls being the most popular exercise at over 6000 reps)!

Since the first beta version, Peter (my co-founder) joined the company. He is a software engineer like me and we previously worked together at Uber. Being two is definitely better than being alone. To bring the user experience to the next level we were aided by our awesome graphic designer who is responsible for the design you will experience in the app. Looking at the screenshots above you can probably guess when he joined.

We are super excited about launching Wodscribe on the App Store very soon and can’t wait for more people to try it. Follow our Facebook page to stay updated, and to get notified when the app is available for download.

--

--