Yes to the Idea, No to the Product
I spent an evening reading a competitor's feature list and came out of it with a sentence I have been using ever since: **yes to the idea, no to copying the…
I spent an evening reading a competitor's feature list and came out of it with a sentence I have been using ever since: yes to the idea, no to copying the product.
The competitor was doing something I believe in. Broad agent platform, many surfaces, a lot of ambition on the page. And my first instinct — the honest one, before I talked myself out of it — was to go wide too. Match the surface area. Get the breadth onto my own roadmap so that nobody comparing the two could say I was the smaller thing.
That instinct is almost always wrong, and the reason it is wrong took me three separate encounters in one week to see properly.
The same decision, three times, in three different businesses
One: my own product, a week from a distribution milestone. I wrote the boundary down in the morning and it was the most useful thing I did all day: no new features before launch. Make the two desktop platforms and the admin page work properly. Mobile, scheduled jobs, machine control, dictation, shared workspaces and agent hand-offs all wait. And then the part that actually bites — the single launch-critical improvement is not a feature at all, it is the quality of the one answer the product already gives. Better verdicts. Real evidence retrieved. Nothing invented in a draft. Visible progress while it thinks.
Two: a consulting client's go-to-market for its AI platform. The platform is genuinely capable across several domains and the marketing kept drifting toward AI for engineering, which is a sentence that means nothing to a buyer. The direction we landed on was the opposite of a capability list: pick one offering, and build an internal certification programme for the delivery engineers first, so that the market-facing claim has a demonstrable capability standing behind it. Narrow the promise, then earn it, then say it.
Three: my own brand positioning. I worked through several broad narratives — the kind that gesture at work and freedom and leverage — and chose the narrowest of them: a right-hand, a chief of staff. One role, one job, plainly stated.
Three different businesses, three different weeks' worth of pressure, and the same answer each time. When the question is "we look small next to them, what do we add," the answer is "nothing, and make the thing we already promised actually true."
Why breadth is the tempting answer and the wrong one
Because breadth is legible. A feature list is a thing you can show a prospect, an investor, or yourself. Depth is invisible until someone uses the product, and by then you have either earned the sale or lost it on something they cannot articulate.
And because breadth feels like competing. Matching a competitor's surface area is a response to them, which feels like action. Deepening your own single promise is a response to your own users, who are quieter and fewer and have not yet asked you for anything.
But the arithmetic does not work. Every surface you add is a surface you now have to make good, and the failure mode of a wide product is not that any single thing is missing — it is that nothing in it is good enough to be the reason someone pays. A buyer does not purchase coverage. They purchase one job done better than their current arrangement does it. If I add six half-working surfaces, I have six opportunities to be mediocre and none to be chosen.
The sharper version, which is really a question about where the moat is: a competitor's feature list tells you what they have built, not what their customers return for. Copying the list copies the artefact and not the reason. And a feature list is the cheapest thing in any company to copy — which should tell you how little of the value is sitting in it.
What "yes to the idea" actually means
This is not an argument for ignoring competitors, which would be its own kind of stupidity. The research was worth doing. What I took from it was directional, not architectural:
- Yes to the idea — the category is real, the problem is worth solving, someone else's money is confirming the thesis. That is genuinely useful information and it should change your confidence.
- No to the product — their specific decomposition of the problem reflects their customers, their sequencing, their constraints and their mistakes. Importing it imports all four.
The practical test I now use when a competitor's capability looks attractive: does this deepen the promise I have already made, or does it add a second promise? Deepening is almost always worth it. A second promise, before the first one is reliably kept, is how you end up with two things that nearly work.
For my own product that test was decisive. Better answers deepen the promise — the whole proposition is that it knows what I committed to, so an answer that misses the evidence is a failure of the core thing, not a missing feature. Machine control, by contrast, is a second promise. Attractive, plausibly valuable, and entirely the wrong thing to build in the month before launch.
Where this goes wrong
Narrow is also the coward's choice, and it wears the same clothes as discipline. Freezing scope is what you do when you are confident about one thing; it is also what you do when you are avoiding something hard. I cannot always tell which I am doing from the inside, and the honest tell is whether the narrowing came with a harder standard. If I cut the roadmap and the remaining item got a specific, observable bar it has to clear, that is discipline. If I cut the roadmap and the remaining item stayed vague, I just did less work and called it focus.
Some markets genuinely are won on breadth, particularly where a buyer is consolidating vendors and the purchase is explicitly "replace these four tools." If the actual buying criterion is coverage, then coverage is the product and depth-first is malpractice. The question is not which philosophy is prettier; it is what the buyer in front of you is comparing. I believe mine is comparing reliability on one job. I should keep checking that, because it is the assumption the whole position rests on.
And there is a counterpoint I cannot fully dismiss: a single-promise product has a lower ceiling, and a competitor who goes wide early may be buying optionality I am declining. That is a real trade. My answer is that optionality is worth very little before you have one thing people pay repeatedly for, because until then you are not choosing between futures — you are choosing which of several unvalidated bets to spread yourself across. But it is a bet, not a certainty, and if the narrow product works and then stalls, this is the paragraph I will be re-reading.
The sentence I want to keep: a competitor's roadmap is evidence about the market and instructions for nobody.