Help Wizard

Step 1

NEXT STEP

FAQs

Please see below the most popular frequently asked questions.

Loading article...

Loading faqs...

VIEW ALL

Ongoing Issues

Please see below the current ongoing issues which are under investigation.

Loading issue...

Loading ongoing issues...

VIEW ALL

How Spotify's mobile app release process works

meahtenoha
2,167 Views

Our product owners have been reading your feedback here in the Community and in our Beta boards. In doing so they've realized there's quite a few questions from you all about how our release process works. One of our POs @henka has drafted this post for you all to help outline our process at Spotify and hopefully answer some questions you all might have. 

 

Hello,

 

We’d like to quickly introduce you to how we deliver our Spotify client software and experience to you on mobile. There are key three points we’d like to share:

 

1. Releases are time-based, rather than content-based

 

Instead of delivering a fixed set of improvements and cutting the release when all the content criteria have been met, we generally deliver on a fixed schedule. This schedule is currently once a week.

 

What this means is that with every release, in addition to all the issues that get fixed in that one week duration, you will also get whatever feature set improvements we can deliver to the deadline.



2. New features are rolled out incrementally

 

So, while we do deliver new features and improvements with most client releases, some users will experience new features in Spotify quicker than others. This is because most of the feature set can be controlled from within the Spotify service itself. We call this incremental or gradual rollout and this is a means for us to make sure that the new features are of top-notch quality when they are shipped to all users.



3. We do A/B tests.

 

At Spotify, we commonly employ a technique called A/B testing to try out tweaks to our product. The point of A/B testing is to ship a particular change to a small enough test group and then evaluate performance against a control group.

 

If the A/B test is successful, with respect to the metrics important to that specific test, we may opt to roll out the change to the full user base.

 

If the A/B test is unsuccessful, we will most likely not pursue that change any further. This is why sometimes you’ll have a friend with a feature that hasn’t reached you yet. Users can also reach out to us if they would like to be removed from a specific A/B test at any time. 

 

Regards,

Henrik Österdahl Djurfelter