The “popular” feature no one used

Hello!

Ever built a feature you thought your users wanted, only to find that no one uses it?

If you’ve been running a software company for long enough, this will have happened to you at least once.

The strange thing is it can even happen when you received lots of requests for the feature.

The “popular” feature no one used

The integrations in my product, Feature Upvote, are popular. Slack, Jira, MS Teams, Discord - almost every customer adds at least one of these integrations to their account. We receive a steady stream of requests to improve these integrations.

So we decided to add another oft-requested integration: Zendesk.

That was four years ago. Today, we have only THREE customers using the Zendesk integration. Hardly anyone has ever looked it at, let alone used it.

If I were a braver person, I would remove this integration altogether from our product to make room for other things: Add a button; remove a button as we wrote in Kill the HiPPO. But a couple of people did start using it, and even though I should remove little-used features, I’m not quite brave enough to do this to paying customers.

Why almost nobody used this highly requested feature is a puzzle. It shows how hard product management is. Despite all the sophisticated methodologies, customer research, and so on we still often get it wrong.

Start with the simplest thing that could possibly work

Luckily we didn’t lose too much time on this. We started with a “simplest thing that could possibly work” approach. We copy-and-pasted the initial version from one of our existing integrations, tweaked a few things, then announced it and waited for feedback. We also informed all the people that had requested it in the past, via Feature Upvote’s email notification system.

We receive zero feedback. Even worse, we could see in our analytics that few people even visited the “add Zendesk integration” screen.

Despite the requests, hardly anyone wanted this feature. So we knew there was no point in continuing to improve it.

Release early and often

I love using feature flags. They’re are a simple and flexible way to avoid spending too much time building the wrong thing.

Most new things we add to Feature Upvote get hidden initially behind a feature flag. We introduce the new feature to selected customers. Then we go into the deploy/listen/iterate cycle.

Recently, Jason Cohen posted on Bluesky:

I like the last line: “Or remove”.

Have you got a story about a feature hardly anyone used and you regret building? Let me know about it!

Kill the HiPPO news

It’s still all about podcast guest appearances at the moment.

It turns out that there is a big gap between recording a podcast guest appearance and it being published, and it is quite frustrating.

I’m awaiting with anticipation on the publishing of some podcast episodes where I chatted about Kill the HiPPO.

I’m learning to be discerning with my choice of podcast. Last week I turned down a couple of potential guest appearances because - to put it discreetly - the podcasts were not a good match. Finding the right podcast is quite the challenge!

Until next time,

Steve McLeod
​​​killthehippo.com​​