This series started with a question from Lenworth Henry at Google Cloud, who works with early-stage founders on the Founders Advocates / Startup Ecosystem team. He asked how founders can figure out how much to build, how to price it, and how to make sure the numbers actually work. We talked it out and ended up with enough for three posts. This is post 1 of 3.
Most first-time founders operate on a version of the same advice: build something people want. It’s true. But it’s also almost impossible to act on because the assumption underneath it is that people can actually tell you what they want.
Nobody knows what they really want.
Not your customers. Not your prospects. Not even you. If you sit anyone down and ask them what they need, they’ll give you a version of what they can already imagine. What they can imagine is bounded by what they’ve seen before. What they’ve seen before is not the thing you’re trying to build. So the answer they give you tells you almost nothing about what would actually be valuable to them.
The gap between “people don’t know what they want” and “build something people want” is exactly where founders add value and make money. Closing that gap is the actual job. Almost everything else is downstream of doing this well.
People can’t tell you what they want. They can react to what you show them.
Humans are extraordinary at reacting. We do it all day, every day. Show a person something and they’ll immediately tell you whether it lands, what’s off about it, what’s missing. That instinct is fast, honest, and mostly unconscious.
That reaction is your actual data source. So the move, from day one, is to get things in front of the right people as fast as possible. Not because those first things will be right. Because their reactions will tell you what right looks like. You aren’t trying to ship a product yet. You’re trying to build a mental model of how your customer thinks by watching them react to real things.
For most of the last decade, this was expensive. You had to sketch prototypes on paper, mock things up in PowerPoint, cobble together no-code versions. Anything to have a thing to show that wasn’t just a description. The wall between “an idea in your head” and “a thing you can put in front of a customer” was tall enough that most founders skipped it and just started building the product.
That wall is gone. With AI, you can now generate high-fidelity prototypes in an afternoon. What used to take a week takes a night. Which means the main reason for sitting in a corner for months and building without customer contact, is habit. And that habit is expensive.
We should also note that there is another subtle reason many founders quietly build in a corner for months or years: Fear. Secretly, often unconsciously, they are afraid to bring their idea to customers. They are afraid to be wrong, afraid to feel rejected, afraid to have their dream burst. So, they procrastinate. They build more. They tweak. They refine. All with the excuse that it has to be just right before they show it to others.
Build something. Show it to someone. Watch their face. Learn. Iterate. Repeat. Do it a hundred times before you build the product you actually intend to sell. The founders who compress this loop are the ones who find product-market fit. The ones who don’t are still explaining why their idea is going to work six months from now.
Talk to buyers, not users
There’s a specific mistake almost every founder makes when they finally start doing this discovery work. They talk to the wrong people.
The instinct is to talk to the people who will actually use your software. Makes sense on the surface. They’re the ones with the day-to-day pain, so who better to describe it?
The problem is that the person who uses your software is almost never the person who decides to buy it. In enterprise and mid-sized organizations, the user is a frontline person, but the buyer is their boss, or their boss’s boss, or someone in a completely different function who owns the budget. In startups and small businesses, the founder might be both the user and the buyer, but the hats are different. What you’re paying for feels different from what you’re using.
If you only talk to users, you learn what would make their day slightly better. That’s real information. It’s not the information that closes deals. What closes deals is the felt pain of the buyer. What’s keeping them up at night. What they’re on the hook for. What their boss is going to ask them about at the next quarterly review. That’s the thing you have to solve for.
This is the reason enterprise software has always been so ugly. The user is not the buyer, and the buyer is the one who pays the bill. The buyer needs the strategic outcome. The user’s experience is a distant second consideration. Whether you like that reality or not, if you’re selling into a company with more than one decision-maker, you have to design your product for the person signing the check.
Go up the org chart. Talk to the CEO if you can. Talk to the person who owns the entire function you’re trying to serve. The buyer’s felt pain is what you’re building for. The user’s experience is a constraint you optimize against, not the target.
How deep is deep enough
Once you’re talking to the right people, the next question is how you know when you’ve actually understood their pain. This is where most discovery work stops too early.
The failure mode looks like this. You’re in a discovery call. You describe the problem your product solves. The buyer says “yeah, that would save us a lot of time” or “that would help us hit our numbers” or “that’s exactly what we’ve been looking for.” You think you’ve found the answer. You go back to the office and start building.
You haven’t found anything. You’ve found the polite, rational, top-of-mind response. That’s not the level where buying decisions actually get made.
The right move is to keep digging. What difference would that make? What does that problem have to do with the goal you just told me you have this quarter? What have you tried already? What happened? What are you actually worried about here?
The signal you’re looking for is emotional, not rational. You keep going until the person on the other side visibly shifts. They relax. They exhale. They say some version of “yes, that’s it!” That exclamation point, that deep sigh of anticipatory relief, that is the tell. That’s the moment they feel understood at a level they weren’t expecting.
A useful mental frame for what you’re digging toward — try to find the dreams they dare not to dream. Every buyer has a set of outcomes they consider realistic and a set they consider impossible. The realistic ones are what they’ll ask you for. The seemingly impossible ones are where the real value lives.
An import-export company hires a customs broker to fill out forms faster. That’s the realistic outcome. That’s what the RFP will describe. It’s also not why they actually pay premium rates to a broker they trust. What they’re actually paying for, if they’ve been through a difficult year in a difficult global environment, is a person who knows the guy who can call the guy and get the shipment out of the jam it’s stuck in. That’s the thing that keeps the importer up at night. That’s what they’d never write into a spec because they’ve assumed no software could ever do it.
Find that or the equivalent in your space. That’s the thing worth building around.
You don’t get to that layer by asking: “What do you need?” You get to it by putting things in front of people, watching their reactions, and asking one more question, and then one more, and then one more.
How this changes what you build
Once you’ve got a real mental model of your buyer and a real read on the dreams they’d not dared to dream, the “how much to build” question mostly answers itself.
You build the smallest possible version of the big idea. Not a slice of the pie. A small pie. This distinction matters. A slice of the pie is a feature that doesn’t stand on its own. A small pie is a complete version of the vision, sized down to something you can ship this month. When you show it to a design partner, they either say “when can I have this” or they don’t. If they don’t, you haven’t found it yet. Keep iterating.
The founders who do this well end up with a specific rhythm. They meet with a customer. They come home with an idea. They spin up a prototype overnight with AI. They show it the next day. They get a reaction. They iterate again. Compressed feedback loops, over and over, until the thing they’re building actually lands.
The founders who don’t do it well spend months in a corner building the wrong thing for the wrong person, and then wonder why nobody buys.
The gap between the two isn’t intelligence, and it isn’t work ethic. It’s whether you internalized the fact that nobody knows what they want, and then built your entire process around that reality instead of pretending otherwise.
Nobody knows what they want. You don’t either. But you can get in front of them, show them things, watch what happens, and build a version of the answer they’ll recognize when they see it.
That’s the job at the beginning. Everything else is downstream.