Wistia brings together continuous deployment and release management to help their teams ship 20-30 times per day.
Releases used to feel like flying blindfolded. Now they feel a lot safer, and their product team is freed up for segmented feature rollouts with more flexibility.
Wistia has multiple streams of work, with a majority of teams working inside of one big app. No one wants to hold anyone else up from getting stuff done, so they need a solid process for managing releases.
Before they started using feature flags, there were a lot of incidents and post-mortems where someone was saying, “We should have put a feature flag around this.”
First, they started with a homegrown flag system for their main app on Ruby on Rails. They used an open-source flag solution with a bunch of caveats before they hit a breaking point and needed a better system because they were running into problems during releases.
Their product team wanted to segment releases and target accounts with rollouts, but the homegrown solution didn't have enough flexibility to define who they were rolling out to. Targeting specific segments or experimenting with the way features were released created a lot more work and was often met with a “no.”
Targeting specific segments could add weeks to the project estimate because the release management system couldn’t support their proposal as it stood. It wasn’t flexible enough for where they were as an organisation.
Wistia wanted to invert that control and have account properties that they were able to target.
They could have written their own advanced feature toggle solution, but building and maintaining it—and dealing with it when it inevitably broke—would have been difficult.
Feature flagging is mission-critical. If it doesn't work, the app doesn't work.
Stability and reliability are crucial, too. Wistia didn’t want to invest more time in building out a system, so they needed a solution that could handle their workflows—without the nerve-wracking feeling of being locked in forever.
By moving to Flagsmith, Wistia was able to improve the developer experience. They were able to unlock the product team and allow them more flexibility in the ways they roll out features to users.
They no longer have to manually tinker with flags to support the product team’s ideas.
Now, they can easily segment features by plan type (say, Pro, Advanced, or Free plan). They can pull all of the Pro customers and just release features to them. They can segment further and just pull the Pro customers that have less than 10 users. The list goes on.
This supports the way that their product team thinks and segments. And it allows them to grow their targeting.
Wistia deploys 20 to 30 times a day and the train doesn't stop.
They have about 80 users right now that use flags. It’s a mix of product managers, engineers, and designers. They encourage their teams to use feature flags across the board. An engineer can create a new feature and wrap it in a flag, and the right people can go in and turn it on for the right accounts.
They also have an audit log, so if something happens in the code or flags are in a weird state, they know who did what.
As a result, Wistia has an overarching trend of getting more done overall. Flagsmith has been part of enabling that. Their changelog looks healthier across the organisation.
They were always able to deploy continuously, but before, it felt like flying blindfolded. Now, it feels a lot safer.
Their multiple releases per day are part of the overarching trajectory of healthier releases. And they can rely on the open source aspect for peace of mind and the ability to just look into the code.
Find the full interview with Nick Quaranto from Wistia here.
Contact us to find out how Flagsmith could work with your releases and teams.