Let me start with the sentence that gets me in trouble.

Your P&L is the only positioning statement your company has ever written.

I don’t mean the number at the bottom. I mean the pattern underneath it. Where the money actually went. Who you actually hired. What you actually refused to build. Read that pattern, and I can tell you what your company stands for. Read your deck, and I can only tell you what a consultant sold you.

Most positioning work confuses the two. That is the whole problem, and it is worth taking apart slowly, because nearly everyone has it backwards, buyers and sellers alike.

I should say where I’m standing before I take anything apart. I’ve spent years reading companies from outside the building. Customer language, financial filings, hiring patterns, what firms ship and what they kill, what they fund and what they walk away from. No inside access, no NDAs, no workshops in the room. The pattern I’m about to describe is one I keep finding from the outside, which turns out to be part of the argument.

The core problem, stated plainly

Here is the problem in its smallest form. A position is a place you occupy in a customer’s mind. You don’t get to put it there by talking. You earn it by doing, over years, at real cost, in ways that are hard to fake and hard to copy. The talking comes after. It either describes a position that already exists or one that doesn’t.

The talking is the only part of this that is easy to sell. So an entire industry sells the talking and calls it the position. The company buys a new deck, a new tagline, a new “narrative,” and walks out believing it changed where it sits in the market. It changed how it describes itself. Those are different things, and the gap between them is where most positioning money dies.

That is the irreducible version. Now, let me show you why it is true rather than just asserting it.

Why talking can’t do the job

Strip away the frameworks and ask the basic question. Why would a customer believe anything you say about yourself?

They wouldn’t. And they don’t. People are built to discount self-promotion. Tell someone you’re the premium choice, and you trigger the part of their brain that asks “says who?” The claim and the skepticism arrive in the same breath. You have spent money to make someone evaluate you at the worst possible moment, the one where they were about to choose.

There is a cleaner way to see this, and economists have a name for it. Costly signalling. Michael Spence developed the idea in 1973, and a biologist named Amotz Zahavi arrived at a similar model while studying why a peacock grows a tail that should get it killed. The short version: talk is cheap, so talk carries no information. Anyone can say “premium.” Saying it costs nothing, so it proves nothing. What proves something is a decision that would hurt to fake.

Think about how you read other people. You don’t believe someone is generous because they tell you they are. You believe it because you watched them give something up when it cost them. The sacrifice is the signal. The self-description is noise, and a little suspicious on top, because generous people rarely announce it. Customers read companies the same way. They are not running a signalling model on purpose. They just don’t trust the label on the jar, and they do trust what they have seen the company actually do.

In-N-Out refuses to franchise and refuses to freeze a single patty. That refusal costs it billions in forgone expansion. You cannot fake it. A competitor can copy the menu by Friday. It cannot copy twenty years of saying no to growth. Patek Philippe discontinued the Nautilus 5711 in 2021, its single hottest product, while people were paying multiples of retail to get one. Walking away from your bestseller is not a move a company chasing the quarter can imitate.

The decision is the proof.
The decision is the position.

So the first principle is not “implicit good, explicit bad.” That is too crude, and I have said it too crudely myself in the past. The real principle is cost and verifiability. A claim works when the claim itself is expensive or checkable: a guarantee you would bleed to honour, a product you killed, a price you cannot walk back. A claim fails when it is only words, because words are free and everyone knows it.

This is also why I keep brain research at arm’s length, even though it points in the same direction. There is a well-known study in which patients with damage to a specific brain region stopped showing the usual brand-preference effect; the label no longer changed which cola they preferred. Real finding, good colour. But you don’t need a brain scan to know that earned associations stick and claimed ones bounce off. The behaviour already tells you. Keep the neuroscience as seasoning. Don’t build the meal on it.

Where positioning actually lives

If talking can’t build a position, what does? Costly decisions, made early, repeated for years. Hiring. Investment. What you build and, louder, what you refuse to build. The shape of your operating model. These are the things that put you somewhere specific in a customer’s mind, because these are the things that stay true about you whether or not anyone is watching.

Here is the part people miss. None of it lives in the communication layer. A press release ships by lunch. A position ships over quarters. Change your messaging, and you have changed something that takes an afternoon to make and an afternoon for a competitor to answer. Change your operating model, and you have changed something that took years to build and that a competitor cannot copy without becoming a different company.

That asymmetry is what the whole argument rests on. A tagline is copied in a day. An operating model cannot be copied at all. It can only be abandoned and rebuilt, which means becoming someone else.

Jim Farley, Ford’s CEO, said the quiet part out loud. Explaining why legacy carmakers can’t update their software the way Tesla does, he described Ford’s own setup: about 150 modules, each with its own chip, with software written by 150 different companies that, in his words, “don’t talk to each other.” He called it a “loose confederation of software providers” and admitted that, to change the code in a seat, Ford has to ask the supplier for permission because the software is the supplier’s property. Then the tell. Ford decided to insource the electrical architecture for its next-generation and, in Farley’s phrase, write the software running the car itself “for the first time ever.”

Sit with that.

Tesla’s position, the car that gets better after you buy it, is not a slogan. It is a decision made years earlier to own the whole stack from first principles. Ford can announce anything it likes on a stage. Until it rebuilds the underlying operating model, the position is not available to it. Farley knows this. It is why he is spending years and fortunes insourcing instead of rewriting the brand book.

And here is the proof that the moat was never the technology. A newer company with the same conviction can match it. BYD started around the same era as Tesla and built vertically from the inside, and it now competes on exactly the integration Tesla pioneered. That does not weaken the argument. It shows what the moat was actually made of. Not the parts, which can be bought or reverse-engineered, but the coherence of an operating model built that way from the start. A company that copies the surface, the words, and the product shots never gets there. The only thing that matches a coherent operating model is another coherent operating model, which means another company willing to become a different kind of company. That is why the moat holds against incumbents and falls only to a builder willing to pay the same price from scratch.

Why it’s invisible, even to you

Now the strange part. This position is hard to see. Not only for competitors. For the company itself.

Legibility is a function of time. In the present tense, a position is invisible to everyone. The market can’t see it yet. Competitors can’t see it yet. And the part founders resist most, the company can’t see it either. From the inside, it doesn’t feel like a position. It feels like conviction. Like how we do things. Like vision, or stubbornness, or plain common sense. Nobody at In-N-Out walks around saying “our position is restraint.” They just think freezing a patty is wrong.

The position becomes legible only in the rear-view mirror, once enough costly decisions have stacked up that the pattern is undeniable. By then it is too late to copy.

Which gives an uncomfortable law. The more readable a position is, the less defensible it is. Anything you can read off a website in five seconds, a competitor can copy in five seconds. Anything that is a real moat stays invisible from outside until it is too late to do anything about it. Readability and defensibility pull in opposite directions. If your position is easy to articulate and easy to recognize, that is not proof it is strong. It can be a sign it is hollow.

There is a second invisibility, and it sits on the other side of the transaction. Ask a loyal customer why they keep choosing you, and you’ll get a fluent answer that was assembled after the fact. They reach for you automatically, then they invent the reason. The choice runs below the explanation. Now look back at the founder. The founder enacts the position the same way, below articulation. The company proves it without being able to name it. The customer chooses it without being able to justify it. Positioning is something both sides do without being able to say.

The industry treats it as the opposite of this on both ends. As a thing the company consciously states, and the customer consciously weighs. A claim, made and assessed. That single inversion explains the whole error. It explains the deliverable, words, and it explains why the deliverable doesn’t work, because words only reach the part of a customer that fades the moment a louder claim shows up. The position that holds is the one running underneath all of it, on both sides, never surfacing as a sentence.

Where does the conviction come from in the first place? Usually one place. The founder. Someone builds the thing they personally wish existed, to solve a problem they personally hate, and their identity transfers into the company. That identity doesn’t show up as a positioning statement. It shows up as a decision rule. A standard for what we will sacrifice, what we will refuse, what we will pour money into. The founder runs every call through that rule without naming it. The rule produces the costly decisions. The decisions accumulate into a position.

So the founder has a standard, not a statement. The standard does the positioning. Which is exactly why it can’t be tidily articulated, and exactly why trying to articulate it first, before living it, gets you nowhere.

The mislabel

Now we can name what the positioning industry actually sells. It sells the map and calls it the territory. Articulation is real work with real value. Naming what is true can sharpen it, align a team, make a real position legible sooner. I am not against articulation. I am against the label. Calling the articulation “your positioning” is the error, because it tells the buyer they changed their position when they only changed their description of it.

Why does the industry do this? Follow the incentives. Articulation is packageable. It is a deck, a workshop, a deliverable you can sell off the shelf and finish in six weeks. Operating-model change is none of those. It is slow, it is specific to one company, it is mostly someone else’s decisions, and you can’t invoice for it cleanly. The market sells what it can package, not what actually moves a position. No conspiracy. Water running downhill.

There is a clean test a buyer can run on themselves after the engagement. Did your position in the market change, or did your deck change? Did customers start describing you differently, or did you just start describing yourself differently? If only the second thing happened, you bought a map. Maybe a good one. But you are standing in the same place.

And here is the signature of the mislabel, the thing to watch for. A heavy noun sitting on light proof. The deck claims a big word, “the leader,” “the premium choice,” “the platform,” and nothing costly underneath earns it. Every word you claim raises the proof standard. Big word, big proof required. When you see a giant claim resting on no costly decisions, you are looking at a description pretending to be a position.

The two honest ways to find your real position

So if you can’t just write it down, what do you do? There are exactly two honest paths. Both share one rule. You never describe yourself.

You can’t, and here is the mechanism. People are genuinely bad at knowing why they do what they do. Ask a founder why the company is the way it is, and you get a confident, fluent, after-the-fact story, and there is solid evidence that the story is partly invented even when the person fully believes it. Self-report isn’t a slightly unreliable witness here. It is invalid by construction for this particular question. The honest answer and the true answer are not the same thing. So self-description is off the table, and two methods remain.

Path one. Read the record.

Bring in a third party with full context and have them read your position the way an outsider reads a label, not by asking you what you stand for but by looking at what you have actually done. The decisions. The refusals. The allocation pattern in the P&L. The words your customers use, unprompted, about what you changed for them. An outsider’s whole edge is pattern extraction, seeing the single thread running through choices you experienced as unrelated, because from the inside they just felt like Tuesdays.

This is where it can turn into a horoscope, so it needs discipline. The reading has to be checkable by you, not authoritative because the consultant said so. Three checks keep it honest. Remove all the words, strip the self-description entirely, and see what the behaviour alone says. Anchor to customer voice, because what buyers independently say you did for them outranks anything you say about yourself; your own copy, unrepeated by customers, is intent, not proof. And match the noun to the proof, so for every big word in the reading you can point at the costly decision that earns it. Survive all three, and it is a finding. Fail one, and it is a flattering story.

Path two. Write the negative filter.

This is for the rare founder with real conviction, working forward. You don’t write what you are. You write what you are not, what you won’t do, and the rule that governs the hard calls. A positive claim like “we’re customer-obsessed” constrains nothing; you can do anything at all and still say it. A negative filter like “we will not ship anything that needs a manual” is costly, checkable, and self-enforcing. It generates the very decisions that become your position. A position is defined by what it excludes. Positioning is subtraction. And you can’t tell people a subtraction. You can only show it, one decision at a time. This is Bezos refusing to optimize for the short-term number, Musk refusing to outsource the hard parts, Chouinard refusing to let Patagonia become a company that needs the planet to lose.

The two paths look like they argue with each other. Only an outsider can see it; against a founder can author it. They don’t argue. They sit at two ends of a line measured in self-awareness. Most companies never wrote a filter; their position emerged by accident, so it has to be discovered after the fact, which is path one. A rare few wrote the filter on purpose and have lived it, which is path two. Same underlying truth, two starting points. And the strongest case is both, in sequence. The founder runs the filter for years without ever naming the word. An outsider reads the pattern those years deposited and hands back the noun the decisions were proving. Only then does the company hold it on purpose, as an internal rule that governs hiring and capital and refusals, never as a line it says out loud.

Where this argument breaks

I would be doing the thing I criticize if I sold you this clean and hid the cracks. Three honest limits.

First. Costly does not mean correct.

A decision being expensive and irreversible makes it hard to copy. It does not make it right. Quibi spent close to two billion dollars building a beautifully specific position, premium short-form mobile-only video, and shut down six months after launch because almost nobody wanted it. A negative filter is only a moat if it points at a conviction the market will eventually agree with. Point it at a fantasy, and you have built an expensive, hard-to-copy way to be wrong.

Second. Signals can spike and fade.

Earlier this year, Anthropic refused to let the Pentagon use its models for mass domestic surveillance or fully autonomous weapons, even after the government moved to bar federal agencies from using its products. Claude went from outside the top 40 in early February to the number-one free app on the US App Store by the end of the month, past ChatGPT, and the company’s chief product officer said more than a million people a day were signing up. A costly refusal moved the market, hard and fast. Then the spike settled. By late March, the daily download lead had returned to ChatGPT, and Claude’s surge had plateaued, though it remained well above its starting level. A costly signal can move the needle in a weekend. Whether it builds a durable position is a separate question, settled over years, not weekends.

Third, and deepest.

This whole argument can become unfalsifiable if I let it. The same logic that says “discount the founder’s explanation, it’s confabulated” can be turned on any inconvenient fact, relabel it post-rationalization, wave it away. And reading a position backward from a winner is its own trap, because everyone who won looks like they meant to. I can’t fully escape this. What I can do is bind it with the three checks, remove the words, anchor to customer voice, match the noun to the proof, so a reading has to survive evidence I didn’t choose. That is the line between analysis and astrology. Not certainty. Discipline.

What this changes on Monday

Enough theory. If I were across the table from you, here is what I would say to do. Stop buying articulation and calling it positioning. The next time someone pitches a positioning project that ends in a deck, ask one question. Will my position in the market change, or just my description of it? If they can’t point at decisions, refusals, and dollars, you are buying a map.

Then pick your path.

If you are like most companies and your position emerged by accident, get it read properly, off your actual record and your customers’ actual words, with the three checks applied so it comes back a finding and not a flattering story. If you are the rare founder with real conviction, stop writing what you are and write what you are not. The things you will refuse. The rule you will run the hard calls through. Then go make the costly decisions that filter demands, and let the position accumulate where it always does.

In the P&L.
In the allocation.
In the years of saying no.

That is the only positioning statement that was ever real.
Everything else is the label on the jar.