Killing Features to Improve your Product

If we accept that 50% of features deployed fail to have the intended impact, then just like technical debt we need to have a regular process and commitment to actively prune these failed experiments.

Not doing so leads to increasingly obfuscated critical user journeys and of course adds to technical complexity and maintenance costs whilst delivering little or no value.

So how do you solve this problem?

1. Regularly audit your features and plot the results on a matrix showing how often a feature is used on one axis against the number of people using them on the other. You’ll end up with a grid showing features that are used all the time by all users in the top right, and those hardly ever used by anyone in the bottom left*.

2. For low use features on the kill list, make sure you understand why they were implemented in the first place. Speak to customers who are using them and understand what value they are deriving from them. How would they feel if it was removed?

3. Communicate internally showing your data and reasons. Remember, unfortunately discussions on features can be very political. The feature you want to kill might have been someones baby (CEO, CTO, VP Sales). In a healthy product culture this shouldn’t be the case, but many of us don’t operate in a perfect world!

4. If necessary, experiment by turning off the feature for new customers or a section of your user base.

Without this regular spring cleaning your product will bloat causing serious UX issues and making it harder for you to pinpoint where the problems are.

* Obviously some features are intended to only to be used once in a while (user profile changes for example), don’t kill these!