The Quiet Funeral: What a Company's Sunsets Say About Its Soul
Photo by Photo by leannk. on Unsplash on Unsplash
Launches get keynotes. Sunsets get a support-forum post published on a Friday afternoon. That asymmetry tells you most of what you need to know about how the technology industry regards the people who depend on it.
Every product team ships features that fail to find an audience. This is healthy. Discipline about what to kill is a hallmark of good product management, and no reasonable user expects a company to maintain every experiment forever. But there is a difference between pruning a garden and boarding up a house while the lights are still on inside. The industry has become remarkably skilled at the second and almost entirely unpracticed at the first.
Anatomy of a Ghost Feature
Call them ghost features: capabilities announced with fanfare, adopted by a committed minority, and then left to decay through neglect. They do not get formally killed — that would require a decision and a communication plan. Instead, they stop receiving updates. The documentation grows stale. Bugs pile up unanswered. Eventually, a changelog entry mentions, almost in passing, that the feature will be removed in the next major release.
The pattern is especially acute in fast-moving corners of the industry. AI platforms, racing to iterate, routinely ship experimental capabilities, watch a community form around them, and then withdraw them when strategic winds shift — leaving workflows, integrations, and in some cases businesses built on top. The users who adopted early, the ones every company claims to prize, are precisely the ones most exposed.
Deprecation Is a Trust Event
Here is the claim this article wants to stake out: how a company kills a feature reveals more about its character than how it launches one. A launch demonstrates ambition. A sunset demonstrates whether the company's promises meant anything.
When a platform deprecates a feature, it is making an implicit statement to every current and prospective user: this is what happens to commitments we no longer find convenient. Developers, who are professionally allergic to this kind of risk, remember. The most damning phrase in software is not "this product is bad." It is "I no longer build on that platform," and ghost features are how that sentence gets written.
Some companies understand this. The best sunsets in the industry share a anatomy: long notice measured in quarters, not weeks; a documented migration path, ideally with tooling; continued security maintenance during the transition; and an honest explanation of why. Users can forgive a feature dying. What they do not forgive is discovering its death from a 404 error.
A User Bill of Rights for Feature Sunsets
If deprecation is a trust event, it deserves codifying. Siiir proposes the following baseline, which any platform could adopt tomorrow:
- The right to notice. A minimum of ninety days between announcement and removal for any feature with meaningful usage — longer for developer-facing APIs.
- The right to a path. No feature dies without a documented alternative, even if that alternative is a third-party tool the company does not control.
- The right to the data. Everything a user created through a dying feature must be exportable in a standard format, permanently, not merely during a grace window.
- The right to an explanation. "Business priorities have shifted" is honest. Silence is not.
- The right to be counted. Platforms should publish usage numbers before killing a feature, so the community can assess — and, if warranted, contest — the decision.
None of these rights would prevent a company from sunsetting anything. They would simply make the cost of abandonment visible, to the company and to the market.
The Competitive Case for a Good Funeral
There is a commercial argument here that extends beyond ethics. In a market where switching costs are high but memories are long, a reputation for honorable sunsets is a genuine asset. It tells the next wave of adopters that betting on your platform is a calculated risk rather than a gamble.
The companies that will earn durable loyalty over the next decade are not necessarily those with the most spectacular launches. They are the ones that treat every withdrawal of a promise — a feature, an API, a pricing tier — as the trust event it is. Launches are how products are born. Sunsets are how companies are judged.