Before Tempo came along, the idea of exercising from home was bleak, expensive, and altogether daunting for a lot of people

When working out from home, it’s difficult to stay motivated, critique your form, time your workouts properly, etc.—and adding a personal trainer attaches a costly price tag to the equation.

Enter Michael Koch, Tempo’s engineering manager, and local Statsig evangelist. “We noticed that the barrier entry to get healthy is very high,” he states.

One of the things that attracted Michael to Tempo was its product Move, which was just an idea when he joined. “Move takes all the high-performing features and engineering of smart home fitness products and crams them all into your phone,” says Michael. Taking the advanced functionality of the high-powered workout studios and applying them to a mobile app was no simple feat, either. “Our very early prototypes were computationally taxing,” he adds.

During his early days with Tempo, Michael found himself building Move, and an entire mobile team soon followed. As a leader, Michael values good culture, good engineering, and good data practices. Thus, it was only a matter of time before a core experimentation platform entered the mix.

Adopting experimentation early

Michael reflects on the state of Tempo’s experimentation early on—before Statsig came along. “I vividly remember those days,” recounts Michael. “There weren’t a lot of tools and processes back then. Experimentation was done through Gitlab and only supported Studio, preventing us from coordinating across platforms.”

In fact, the first experimentation tool Michael vetted was a miss:

“The biggest issue was that they weren’t open source. It was extremely difficult to figure out how their SDK operated.” He says. “Further in the discovery process, we also learned that they intercept a lot of code to do UI things, which was very inconvenient.”

“A lot of experimentation platforms used to break our experiments and product in unexpected ways.”

Michael Koch

Engineering Manager

Enter Statsig

As more and more pain points were adding up, Tempo eventually decided to vet Statsig instead for its experimentation needs. “We had to accept that we had gone too far down the road with the other platform and needed a solution quickly. Ultimately we went with Statsig,” Michael comments.

Getting started with Statsig was quick and painless, as Michael explains: “Setting up Statsig was very easy. They actually built out special SDK functionality for us. Other than initially developing our own SDK wrapper, there has been no impact to our own engineers’ time.”

Michael also commends Statsig’s responsiveness to its customers—during the setup process and persisting throughout Tempo’s customership of Statsig. “Our quick setup was largely in part to the Slack channel and the solutions engineering team. We joined the Slack channel and it was astounding how fast the team was responding to requests.”

He laughs, “[Statsig was] building features faster than we were.”

“[Statsig was] building features faster than we were. They generated immediate value.”

Michael Koch

Engineering Manager

Upping the tempo on experimentation

With Statsig up and running, Tempo was able to incorporate experimentation into everything across the platform. “We were focused on the Move, and were just about ready to release it. The need to experiment across all platforms was critical, and we used Statsig for all of our features,” says Michael.

“We finally had a tool we could rely on, and which enabled us to gather data intelligently. Statsig is similar to an IT department. It’s just always working—because if it wasn’t, you’d notice.”

Michael Koch

Engineering Manager

Michael recalls a particular customer delight: “One of our vendors plays music in our classes. Since we have volume on a dynamic config, we can react to customer complaints about volume and change the volume instantly.”

Tempo experienced another unique challenge. “With the launch of the Move, we were going to have two very different UIs that would exist simultaneously,” notes Michael. “We needed to fork the experience very early, and really trust the tool we used.” Because of this, Statsig was included very early in Move’s lifecycle. “We needed something timely that will give us information quickly,”

A true culture of experimentation

Michael describes Tempo’s data shareout meetings—a bi-weekly meeting wherein data collected with Statsig, and visualized with Mode, is shared with the team. These meetings “didn’t exist without Statsig,” according to Michael.

“Every department uses Statsig is some way, shape, or form, to make decisions. We can make choices about what and when instead of doing the entire process at once.”

Michael Koch

Engineering Manager

A fun example, Tempo was able to increase its app store rating by testing different placements within the app for the “leave a review” popup. “We also experimented with different copy,” advises Michael.

After successful experimentation, Tempo’s rating jumped from 3.7 stars to 4.5 stars. “Prompting users after their workout was the best time and place. The other variant was upon users opening the app.”

Michael explains that users who have just finished a workout are using the app as intended, and are more likely to leave a fair review. Prompting users who have just opened the app—the old variant—turned out to be an inconvenience for any workout-ready user.

“All of our major feature releases are made possible because of Statsig,” Michael admits. And since Statsig integrates with Tempo’s data analysis tools, it helps stakeholders ensure they’re meeting their KPIs as well.

“Developers should not be afraid to experiment. At Tempo, it takes zero extra time to do it,” Michael shares. “There are often cases where people are afraid to add new events. But with Statsig, it’s painless.”

To date, Michael’s team has been able to increase their release cadence by 50%, thanks to Statsig, the data shareout meetings, and Tempo’s culture of experimentation.

His outlook is optimistic for the future of experimentation at Tempo: At this point it’s simply part of their identify.