How to Know What Features to Kill

When you build, and eventually maintain, a product, you’ll eventually have to make a hard decisions: what features should you kill? Those ideas you had early on, that made it into the product, aren’t always good ideas. Sometimes, users don’t use those features as much as you want them to. Unless you prune out these features, you’ll end up with bloated, unusable software that leaves your users looking for other solutions.

So, how do you know what to kill?

There are two things to consider: how frequent users use a feature, and how critical that feature is to the application. Grading features on those two axes will give you a way to have a productive conversation with your team about whether or not these features should stay or die.

Here’s how it works:


There are essentially 4 categories. Let’s break them down.

Frequently Used, Critical

These features are the core of your product, the ones that make up the very base of what the product is built around. For CrowdSync this is the ability to create a workflow, add actions to it, and add people to that workflow. These are the features that users are in the application using constantly, and without them, we wouldn’t have a product.

These features should get a lot of attention, should comprise your MVP, and should be at the forefront of refactoring and stability efforts.

Frequently Used, Non-Critical

These features are used often, but if they disappear, users can still get their core work completed in the product. Things like email notifications, sorting, searching and other enhancements fall into this category.

Watch yourself with these. They’re fine to add, but ensure you’re measuring their usage (more on that in a minute), to make sure they are indeed frequently used.

These features should be marketed, as they’re highly popular, and if you have an upsell path in your product, should be part of that upsell opportunity.

Infrequently Used, Critical

These are the unsexy features of your product: forgot password, account update, user profiles and the like. Critical for ensuring users can adequately use the product, and necessary to be there very occasionally, but not something people are using much.

These features should be made rock-solid stable, but this isn’t the place to innovate. Save your creativity for features in the previous two categories.

Infrequently Used, Non-Critical

Here we are - the features to kill. These features border on useless. They’re not important, and don’t get used often. Mark these features for deletion early and often.

What’s Frequent, and What’s Critical?

Now, onto the hard part - how do we measure frequency and criticality? The answer is, it depends. Each product is different, and the nature of use is way too wide to have any useful heuristics.

For determining criticality, ask yourself: “Can our users fully accomplish the core task in our product without this feature?” If the answer is yes, it’s probably not critical. It might in interesting, and might be popular, but it’s not critical.

Frequency requires measurement. For each feature in your product, you should be measuring the interactions with that feature, by user. Rolling a simple script to update a count in a Redis store or database whenever a user interacts with a feature (“Oh, they filtered that list, +1) is a great way to start getting an idea for this. There are other software packages out there that can also help with measuring this, but the fact is, you have to be measuring this, else you’ll end up in a “No, a lot of people use it, I KNOW it!” conversation, which you don’t want to have.

How frequent is frequent enough? Well, that all depends on the app, and it might take you a bit to start figuring that out. Look at usage by user, by session, and in the aggregate. Start computing some averages, some trends, and some deviations from other features you have. Put features on a frequency leaderboard, and let them fight it out. However you sort it out, just make sure you’ve got a process in place for measuring and reporting back on how often things are used.

Don’t Be Afraid to Commit (Feature) Murder

Building software products is one of the fastest ways to get lured by the sunk cost fallacy (perhaps right after investing). When you spend time building a feature, you become attached to it. That investment you made in bringing it into the world clouds your vision, and prevents you from objectively understanding if that feature should really be there or not.

By grading things on these two axes - frequency and criticality - you’re able to bring some objectivity to the process. Sure, there’s some subjectivity in determining where they land, but the fact that you’re being proactive about saying “What features should we kill this month?” means you’re well ahead of the pack.

Keep shipping, and keep trimming.

Next issue.

Related Posts

Default Yes vs. Default No

Are you and your team members default yes, or default no? One is good for startups, the other not so much.

Check Your Echo Chamber

The people you surround yourself with create your reality. Choose carefully.

Don't Forget the Goal

There's only one thing that matters when you're building software.

How To Get a Job Offer

Want to get a job in the field you love easily? This is how.

Required Reading for All Couples

If you're married, about to get married, or just committed to someone for a long period of time, these three books are absolutely required reading.

I Launched a New Podcast, and I Want You to Call In

I just launched a new podcast called Design By Committee, dedicated to answering your questions about UX, product design, content, strategy and anything else tech.

Shitty Sales Have Made Product Development Harder

Shitty, one-sided sales processes have made product development much more difficult for early stage startups.

Why I'm Cold Emailing You

You might have gotten a cold email from me. Tasteless? Some people think so. Here's why I'm doing it.

How I Found Your Email

I've been cold emailing a lot of people, and many folks are surprised that I found their email. Here's where I dug it up.

Sales is User Research, Undercover