I hold a few personal disciplines around building businesses. One of them is: don't build a business on supply-side logic.

Put differently: never let go of the premise that a business must start from the customer's real problem.

That said, I run into the exact opposite of this discipline more often than I'd like.

Selling "What the Builder Can Suddenly Do"

I feel this pattern shows up especially easily in AI product development these days.

With AI, builders suddenly gain abilities they didn't have before, things they used to be bad at or simply couldn't do. Coding is my own example of this. It's true that the technology raises the ceiling on what you're capable of.

Products that embed AI also become able to handle problems that used to be hard to solve with deterministic computation.

And then that newly acquired capability gets turned straight into something to sell, packaged as a business. I see this pattern constantly, and it feels like pure supply-side logic to me.

The starting point isn't the customer. It's the tool sitting in your own hands.

The temptation gets stronger exactly when the tools change, and because everyone moves the same way at the same time, it ends up looking like a trend from the outside.

As an aside, LayerX's phrase "selling the completion of work" strikes me as the exact opposite stance, a strong, customer-first way of framing things, and I think it's wonderful.

"The Customer Asked for It" Is a Drug

People who build businesses on supply-side logic almost always justify it as "this is customer logic."

Something like, "we interviewed the customer, and they said they wanted XXX."

That phrasing is a powerful weapon. It legitimizes the business and gets it past internal scrutiny. It's genuinely something the customer said, there's a record of it, and it's hard to argue with.

But I think the real question isn't "what did the customer say," it's "what does the customer actually need."

What people truly need often lives in the nonverbal, in the unconscious. What comes out as words is, I think, only a small fraction of that.

And whatever does get put into words tends to be framed as an extension of an existing solution, an expression of dissatisfaction with what already exists, or a request to improve it.

So if you take the customer's words at face value, they become indistinguishable from supply-side logic. It's easy for a builder's own wish to dress itself up in the form of "the customer said so."

Catch the 80% Who Stay Silent, Not the 10% Who Speak Loudly

What I'm always trying to find isn't "what 10% of customers say loudly," it's "what 80% of customers deeply want but can't put into words."

The first kind is loud, and anyone can spot it. It always turns up in interviews and in quantitative research alike. And it's easy to mistake volume for demand.

The second kind is hard to find. Nobody can picture what "it" would even look like, and customers themselves can't recognize what's missing. But the moment it's satisfied, there's no going back to a world without it.

The second kind is what makes for a strong business, and it's what I want to build.

I think this is exactly where the real difficulty of founding something lies. You have to surface a psychology that hasn't yet surfaced, and go after it, saying: this, right here, is what you actually wanted.

So what do you use as a clue? Not the customer's words, I think, but their behavior.

First, workarounds. The gaps people quietly patch with Excel, paper, or their own unwritten rules, without anyone asking them to. They never put it into words because they've come to think of it as just part of the job.

Second, traces of resignation. Tasks explained away with "well, that's just how it is." A habit nobody can explain the reason for is usually the fossil record of some past constraint.

Third, friction that never shows up in a metric. Judgment calls that never get logged as yield loss or labor hours, and exist only inside one particular person's head.

None of these come out if you just ask. You can only catch them by being there and observing. Compared to picking up something someone actually said out loud, finding the voice that never became a voice is far harder. There are fewer clues to go on, and you have no way of knowing whether you got it right until you actually ship it.

Which is exactly why you can't let yourself take the easy path.

Existing Operations Have Their Own Gravity Toward Supply-Side Logic

There's one more thing worth thinking about: the gravity that an already-running business exerts.

Any existing business that's still standing must have cleared this same difficulty at least once. There's a history there, it got off the ground because someone caught a voice that had never become a voice.

Once it clears that bar, though, the operation that follows is almost entirely about listening to the customers who are already using the product, rather than the latent voices of customers who aren't. Day-to-day decisions end up running entirely within the loop of existing metrics and existing requests.

On top of that, the people operating that business almost never get a chance to train the muscle of noticing. Seven years ago I wrote that noticing takes training, and that people who actually train for it are exceedingly rare. I still believe that.

A handful of people are actively trying to notice something in the current state of things, but most people aren't. Noticing takes training, training in asking "why" of the present, and in finding what's missing.

The biggest realization I've come to, no matter how capable the executive, the founder, the product manager, or the creator, is that people who actually train themselves to notice things are exceedingly rare. (abridged)

Source: Building 10x Products (2019)

As a result, when you try to spin up a new business out of an existing business's habits, you almost inevitably end up building something new on supply-side logic. And that kind of business tends not to have much long-term staying power.

An organization forgets, once its phase changes, the difficulty of founding it once had to overcome itself. Maybe that's inevitable, since the muscles a company needs are different at each phase.

At 10X, we've been fortunate that Stailer, the business we launched in 2020, has kept going for a long time and has carried the company forward. At the same time, creating new businesses remains one of the most important issues we face.

I want to keep chewing on this problem as my own, and keep moving forward with it.