The hardest product decisions I have made were rarely about obviously bad ideas. Bad ideas tend to lose quickly once the evidence is visible.
The dangerous ideas are the ones with something real behind them: revenue, funding, a credible customer, an adjacent market, a founder who can see exactly how it might work.
During a clinical research AI initiative, there was pressure to pull an adjacent hospital intervention idea into the product. It was not absurd. That was the problem. It was close enough to sound strategic and different enough to pull the product away from what we were actually trying to prove.
Individually rational decisions can create an incoherent product
Imagine saying yes to five adjacent opportunities over two years. Every decision has a business case. Every customer is real. Every feature has an owner.
The result can still be a product nobody can explain in one sentence.
I think of this as product entropy. Not as a framework to put on a slide, but as a useful warning. Every adjacency changes the internal definition of what the product is allowed to become.
The question I ask before accepting adjacency
Does this strengthen the reason the product exists?
What stops receiving attention if we pursue it?
Would we still want this if the immediate revenue or funding disappeared?
Does this create a second product hiding inside the first one?
Focus is an economic choice
Teams sometimes talk about focus as discipline or taste. I think it is more concrete than that. Every yes reallocates engineering time, product attention, commercial narrative and future maintenance.
When I pushed back on that adjacent initiative, the argument was not “this is a bad idea.” It was that mixing the two would weaken the case for the product we were already trying to build. The initiative stayed focused and later secured major European funding.
I do not take that outcome as proof that saying no is always correct. I take it as evidence that coherence has economic value.