how to waste time building a startup
There is a recipe for wasting years of one's life on a company that will never work. It's reliable, well-documented, and many folks follow it.
To do this, you need to skip curiosity, avoid running experiments, replace discovery with performance, and treat your startup like a story you're telling about yourself rather than a question you're trying to answer. If you do these things, you will build something that looks, from the outside, like a company. It will have a pitch deck, it will have a logo, it may even have funding. But it will not have the one thing that matters, which is contact with a truth about the world that you didn't know before you started.
I want to talk about that contact, and about the specific machinery that produces it, and about the thousand ways founders avoid it while convincing themselves they're pursuing it.
Every great company sits on top of a graveyard of experiments. These are small, cheap, ugly tests that each produced a scrap of information. Most of them produced the information that something doesn't work. But a few — sometimes just one — produced the information that something does.
I sat in on a guest lecture by Andy Rachleff, who coined the term product-market fit, and he put it starkly: finding PMF is a discovery problem, not an optimization problem. You cannot execute your way to product-market fit. You cannot hire your way there, brand your way there, or network your way there. You can only learn your way there, and learning requires running experiments that might fail, that will probably fail, and being genuinely okay with that.
He puts forward an equation I keep coming back to: the probability of building something big is a function of the quality of your insight multiplied by the rate at which you learn. Which means the most productive thing a founder can do, at the earliest stages, is to run experiments and learn from them.
Here's what wasting time looks like in a startup.
One: you build before you've found desperation. You have an idea, the idea seems good, people nod when you describe it, so you start building. You hire engineers. You design screens. You choose a tech stack. You spend four months constructing something that, when you finally put it in front of actual users, produces a smile and the phrase "that's interesting."
Interesting is the kiss of death. It means they don't need it. Rachleff insists on a stronger word: desperate. If they're not desperate, you haven't found anything yet.
Two: you add features instead of changing audiences. If your first users aren't desperate, you think the problem is the product and you add a dashboard, integrations, a feature someone mentioned in a feedback call. But that is lunacy — adding features doesn't make someone more desperate. It makes your product harder to evaluate, harder to sell, and harder to learn from. The right move is to keep the product fixed and change who you're showing it to. Fix the what, iterate on the who.
Three: you hedge. You test three markets simultaneously because you're not sure which one will work and you don't want to miss. This is the most natural thing in the world and it is catastrophic. You can't learn from three experiments at once. You can't read the signal. If something fails, you don't know if it's the product, the audience, or the positioning. You've introduced so many variables that no outcome is interpretable. Hedging is the enemy of learning, and learning is the only thing that matters.
Four: you worry about competition. Barriers to entry, brand, margins — before PMF, these are luxurious distractions. They are answers to questions you haven't earned yet.
So what's the alternative? What does not wasting time look like?
To me, it looks like exercising curiosity. Curiosity as the central organizing principle of how you spend your time and money. Formalizing it into hypotheses, converting those hypotheses into experiments, running those experiments fast — with the explicit intention of being wrong — because being wrong fast is the fastest path to being right.
This is the scientific method applied to entrepreneurship. Running cheap, ugly experiments that will probably fail is not emotionally satisfying. It is, in fact, ego-destroying. It requires you to sit with the possibility that the thing you've been telling people about at dinner parties — the thing you've built your identity around, the thing you left your job for — might be wrong. That is hard.
But if it works, it works very well.
Howard Marks's two-by-two matrix maps all possible outcomes along two dimensions: right versus wrong, and consensus versus non-consensus. If you're wrong, nothing works. But if you're right and consensus — if you see the same opportunity everyone else sees — the returns get competed away. The only quadrant where outsized returns live is right and non-consensus. You see something others don't, and you're right about it.
Getting to that quadrant requires being genuinely, stubbornly, irrationally curious about whether something is true. Not performing confidence in a thesis. Not managing the narrative of your startup. Actually wanting to know.
If you're building a startup to prove something about yourself, you'll probably waste years building the architecture of vanity — unless you're unusually self-aware.
If you're building a startup because you are genuinely curious about whether something is true — whether this tech can do this thing, whether this person has this problem, whether this model can work — then the experiments will come naturally. The graveyard will fill itself. And somewhere in the dirt, you might find something worth building.
The difference between a wasted startup and a real one is whether the founder confused performing with discovering. It's whether they had the nerve to be curious rather than impressive.
Every other mistake is recoverable.
Onward,
Abhi