The most efficient way to kill a startup is to become a dropdown.

Not a direct competitor. Not a lawsuit. A dropdown. A menu item in a product someone already uses every day. “Now with AI writing.” “Now with summarization.” “Now with X, the thing your whole startup was about.”

The wrapper era learned this lesson expensively. The companies built to sit on top of foundation models — adding one layer of value, one specific prompt pattern, one particular UI — found that the models themselves kept growing. The layer they occupied kept getting absorbed into the foundation.

What Gets Commoditized First

Features with clear value propositions and no switching costs die first. If someone can describe what your product does in one sentence, and that sentence is “it does X for you,” and X is something a large platform might want to add — you’re vulnerable.

The tell is: does your value live in the feature, or in everything surrounding the feature? Data integrations, user context, workflow position, trained behavior — these are harder to replicate than the capability itself.

A tool that summarizes text is a feature. A tool that summarizes the specific kind of technical documents your team produces, in the format your review process requires, integrated into your existing workflow — that’s a product. The summarization part will get commoditized. The rest is harder.

The Three Exits From the Dropdown

Go deeper into the niche. If the general-purpose version isn’t good enough for your specific users — if their needs are particular enough that the dropdown doesn’t actually solve their problem — you’re safe. The pediatric dental agency software that does exactly what pediatric dental offices need is not threatened by generic CRM features. You win on fit, not on capability.

Become infrastructure. Products that other things depend on have stickiness that features don’t. If removing you breaks workflows, you’re not a feature anymore. The more embedded your tool is in how something works, the harder you are to replace with a dropdown.

Get small enough to be irrelevant to big players. There are markets that are valuable to the people in them and not worth addressing for anyone building for general audiences. A $2M ARR niche software business is uninteresting to every major platform. That’s protection, not consolation.

The Timing Question

The real danger isn’t that big platforms will commoditize your feature — it’s that they’ll do it before you’ve built enough defensibility to survive it.

The race isn’t to have a unique feature permanently. It’s to build something that can outlast the inevitable commoditization of that feature’s functionality. Use the feature as the hook, then build everything that makes you a product rather than a wrapper around it.

The dropdown is coming eventually. The question is whether you’ve built something worth keeping when it arrives.