Every pre-seed founder eventually asks some version of the same question: what do I need to raise? How many users, how much ARR, what growth rate, what retention curve. It’s a reasonable question, and the internet is full of people willing to answer it with a number.
The question isn’t wrong, exactly. But it’s pointed at the wrong thing.
Numbers are the readout. The engine underneath is an insight about your customer — one sharp enough to shape what you build and how you get it to people. Founders who optimize for the readout tend not to produce it. Founders who obsess over the insight usually do.
Numbers are evidence, not the goal
Early traction matters. I’m not going to pretend otherwise. When someone shows up, pays, and comes back, that’s a real signal — it means your understanding of the customer is grounded enough in reality to change someone’s behavior. That’s worth a lot, especially at pre-seed when almost nothing else is observable.
But the numbers are downstream. They’re the visible output of an underlying bet about who the customer is, what they actually want, and how to reach them. When the bet is right, the numbers follow with surprising ease. When the bet is wrong or vague, you can push hard on growth tactics and barely move — and you won’t know why, because the thing that would explain it is the thing you skipped.
The trap of optimizing the readout
Founders feel pressure to show progress, so they focus on what’s countable: signups, MRR, waitlist size, week-over-week growth. This is understandable. Progress needs to be legible — to investors, to cofounders, to yourself at the end of a hard week. A number is legible. A vague sense that you understand your customer better than you did last month is not.
The problem is that once the number becomes the goal, it pulls you toward tactics that goose it. Paid acquisition. Discounts. Chasing any customer who’ll convert. None of these are bad on their own, but they don’t sharpen your understanding of the customer — and often they blur it, because the people who respond to a 40% discount aren’t the people who’ll build your business.
You end up with numbers that look okay and a business you don’t understand. The next set of decisions — what to build, who to hire, what to charge, which segment to double down on — has nothing to anchor to. Worst case: you raise on the numbers, then can’t repeat them, because you never knew why they happened in the first place.
What the insight actually does
A real insight isn’t a slogan or a positioning statement. It’s a load-bearing piece of understanding that does specific work for you.
It tells you what to build. The product roadmap stops being a debate and starts feeling obvious — you’re not arguing about features, you’re working down a list the customer practically wrote for you.
It tells you how to reach them. If you understand the customer deeply, you know where they already are, what they already read, who they trust. Distribution stops being a separate problem bolted on after the product and becomes an extension of the insight itself.
It tells you what to ignore. Most founder overwhelm comes from not knowing what not to do. Every feature request, every potential partnership, every adjacent market looks plausible at 10pm on a Tuesday. A sharp insight is a filter — it lets you say no quickly and without regret.
And it compounds. Every customer conversation, every product decision, every hire sharpens it further. Vague positioning does the opposite: each decision makes the next one harder, because you’ve added complexity without clarity.
What a roadmap-anchoring insight looks like
The insight has to be specific enough to be wrong. “SMBs want better software” isn’t an insight; it’s a category. “Bookkeepers at 10-to-50-person services firms do month-end close in a shared spreadsheet because their revenue recognition breaks QuickBooks, and every firm has independently built roughly the same spreadsheet” — that’s an insight. You can picture the person. You can guess what they’d pay for. You can imagine the demo.
It has to be generative. It should produce at least a quarter of your product decisions without you having to strain for them. If every feature discussion starts from scratch, your insight isn’t doing the work it needs to.
It has to be load-bearing for distribution too. A good insight tells you not just what to build but who to build it for and how to find them — which Slack groups they’re in, which newsletters they read, which peers they trust. If your product insight doesn’t come with a distribution insight attached, you only have half of one.
And it has to be earned. It came from real exposure to the problem — watching customers, doing the job yourself, sitting with the weirdness long enough to notice what everyone else missed. Insights borrowed from a market map or a competitor’s blog post don’t survive contact with customers, because you can’t defend the parts that aren’t written down.
The sharpness test: can you defend your divergences?
Here’s the most useful test I know for whether an insight has actually landed.
A sharp insight shows up as product decisions that differ from competitors or from category convention — and you can explain exactly why. Not “we’re simpler” or “we’re AI-native,” which are aesthetic claims rather than insights. Something more like: “we don’t have a dashboard because our users check this on their phone between meetings, and a dashboard would make them feel behind instead of in control — so we send a single daily summary instead, and week-four retention is three times what it was when we had a dashboard.”
Notice what’s happening in that sentence. It has three parts, and all three have to be there:
The decision is genuinely different from what others do. Everyone in the category has a dashboard; you don’t.
You can trace the decision to a specific belief about the customer. You believe they check this on their phone between meetings, and that a sense of control matters more to them than a sense of completeness.
You can point to the customer behavior it drives. Retention moved.
If you can only do the first part, you’re being contrarian for its own sake, which is a common and expensive mistake. If you can do the first two but not the third, you have a hypothesis — a reasonable one, maybe — but not an insight yet. You haven’t earned it. When all three are there, the insight has landed. It’s producing decisions that produce behavior you can see.
Run this test on your last five product decisions. How many pass? That’s a better readiness signal than any ARR number, because it tells you whether the business you’re building is anchored in understanding or in guessing. The ones that pass are load-bearing. The ones that don’t are either momentum or noise.
How to tell if you have one yet
A few questions worth sitting with honestly.
Can you state, in one paragraph, what you believe about your customer that most people building in your space don’t believe? Not what you’d write on a landing page — what you’d say to another founder who knows the space. If the paragraph comes out generic, that’s information.
Does that belief tell you what to build next quarter and what to cut? Does it tell you the first three distribution channels to try and which to skip? An insight that doesn’t generate decisions isn’t pulling its weight.
When you describe it to a prospective customer, do they interrupt you because they recognize themselves? When you describe it to an investor, do they stop asking clarifying questions and start telling you things about the market? Those are the tells that your insight is legible to other people, not just to you.
If the answer to most of these is no, that’s your actual work — not more growth experiments, not more interviews for their own sake, not another round of funnel optimization. The work is going back to the customer until something sharpens.
The reframe
The numbers aren’t the point, but they’re also not a distraction. They’re the signal that your insight is real — that the bet you’ve made about the customer is connected to reality in a way that moves behavior.
The mistake is working on the numbers directly. The move is working on the insight and letting the numbers be the consequence. Founders who get this right tend to hit fundraising thresholds without having aimed at them. The traction is a byproduct of clarity, not the other way around.
When a founder asks me what numbers they need to raise, the honest answer is: fewer than you think, if your insight is sharp — and no amount will be enough if it isn’t. Investors at pre-seed are pattern-matching on whether you’ve seen something they haven’t. Numbers can help you prove it, but they can’t substitute for it.
Spend your energy on the thing that drives everything else. The readout will take care of itself.