startups need an asymmetry • Solving the decision problem skip to content

Close

startups need an asymmetry<br>12 August 2026 A quick note on startups

I tweeted some version of this and figured I’d expand on it a little. I don’t really feel qualified to turn it into a grand theory of startups, so take it or leave it.

I think a startup, or at least the kind of startup I’m interested in, starts with some bet that changes the physics of how something is already being done. that could be a new piece of technology, some IP, a business model, a regulatory change, or just noticing something about the world that other people haven’t noticed yet. I’m using “physics” pretty loosely here, but I mean that some important constraint has changed and you think you understand the consequences before most other people do.

This will probably make the thing look strange for a while, which is sort of the point. if the idea already fits neatly into how everyone understands the market then you’re probably competing much more directly with everyone already in it. instead you have this slightly awkward period where you can see why the thing should exist but other people don’t yet, and a surprising amount of the work is figuring out how to explain it in a way that makes them want it.

sometimes the product itself does that. sometimes it’s marketing, or positioning, or sales, or just finding the right group of people who immediately get it and become annoying about telling everyone else. however you do it, you need traction; being correct about the future in private isn’t especially useful.

the other problem is that the incumbents have all the normal stuff already. they’ve got the features people expect, integrations, sales, support, existing customers, years of boring work that you haven’t done yet. so whatever your bet is, it has to matter enough that people will put up with your product being worse in fairly obvious ways.

I actually think this is a useful test. people should be recommending your thing while complaining about it. there should be some feature where a customer says “how the hell do you not have this”, and then continues using the product anyway because the reason they came to you is still more important.

eventually, if you’re right, everyone else notices.

the feature gets copied, the idea stops sounding weird, competitors use the same words you use, and six months later somebody much bigger announces roughly your product as a checkbox in theirs. you can’t really be surprised by this; in some sense it’s evidence that the original bet worked.

so somewhere along the way you have to turn the original insight into an organisation that can keep moving. ideally getting the first thing right has taught you something, given you technology or distribution or customers or just a particular way of building, and that makes the next thing easier for you to see or do than it is for somebody copying the surface of what you’ve already shipped.

this is the bit I think people mean when they talk about speed, although “move fast” by itself has always felt a bit empty to me. it’s not particularly useful to type faster than everyone else. you want the company you’ve constructed to make certain things unusually cheap or obvious for you, so that by the time the thing you were originally right about is conventional wisdom you’ve already followed it somewhere else.

without something like that, you’re mostly trying to build a good business. which is fine! obviously. but now you’re competing over things like execution, price, sales, distribution, completeness, and capital, against companies that have spent years getting good at exactly those things.

maybe this is the distinction i’m trying to get at: a startup needs some asymmetry that lets it be unreasonably ambitious without having to win every normal contest at the same time.

at first that’s usually the bet. after that, hopefully, it’s the company you’ve built around it.

best of luck.