The Art of Killing Features: Why Sunsetting Is a Growth Strategy

The invisible mortgage

No one gets a bonus for shutting down a server or deleting a repository. But they should.

We naturally create emotional bonds with what we build. It's human. We call features our "babies," we defend them in meetings, and we ignore low usage metrics with excuses like "marketing didn't push it enough" or "the market just doesn't get it yet."

This isn't persistence. It is the Sunk Cost Fallacy draining your team's energy.

What many Junior PMs fail to realize is that every new line of code carries a lifetime mortgage. It’s not just storage space; it’s the cost of cloud computing, support tickets, documentation updates, and, most importantly, cognitive complexity.

Case Study: The 0.8% Anchor

I remember a specific "trench tale" from a previous role. Our app's operation was being held hostage by Jurassic versions of Android.

When I dug into the data, I found that only 0.8% of our user base was on these legacy systems. In the revenue spreadsheet, this impact looked irrelevant. But in the invisible cost spreadsheet, it was brutal.

Every release cycle took an eternity because it required manual testing on obsolete devices. Worse, we couldn't implement modern architecture or new features because the old code couldn't support them. We were leveling our innovation down to accommodate less than 1% of users.

The Decision Matrix

Taking this to stakeholders is never easy. The fear of complaints is paralyzed.

(This is a classic scenario where I now advise using Decisor.ai. You need to strip away the emotion and present a weighted matrix: The cost of delayed releases vs. the revenue risk of 0.8% churn. When you put numbers on the "fear," it becomes a manageable risk).

I presented the case. I got the buy-in (loaded with skepticism), and we executed the sunsetting of support.

The Aftermath

The result? The volume of complaints was statistically irrelevant.

On the flip side, the engineering team's velocity skyrocketed. We were finally able to modernize the stack and ship value faster to the 99.2% of users who actually drove our revenue.

Conclusion: You Are an Allocator of Capital

A Senior Product Leader isn't paid just to decide what goes into the backlog. You are paid to have the stomach to decide what comes out.

Keeping zombie features alive to avoid a few angry emails isn't "customer obsession." It is poor capital allocation.

Just like a gardener must prune dead branches to allow the tree to grow stronger, you must prune your product. If you don't, the dead weight will eventually drag the whole thing down.

(Thinking of auditing your own life or diet? The same logic applies. Whether it’s calories in MealAudit or features in a backlog, the principle is the same: cut the excess to focus on the nutritious).