I watched the brand growth debate over the past week. Six people walked up to the same empty room and stopped. The one who got closest got there thirty years ago.

For a whole week, I read the same argument made by different people, and the same thing kept bothering me each time. Not the disagreement. The disagreement is easy to follow, and most of it is fake. What bothered me was the part everyone walked right up to and then stepped around.

It came to a head at Cannes.

Mark Ritson and Byron Sharp shared a stage for the first time in a decade. The data man and the strategy man, the two camps everyone assumes are at war, sat down to mark where they agree. They named five things. Mental availability matters most. Distinctive brand assets are non-negotiable. Sophisticated mass marketing beats niche thinking for brand building. Brand purpose is largely nonsense. Consistency is criminally undervalued (Ehrenberg-Bass Institute, The Drum).

Read that fifth one again. Consistency is criminally undervalued. Consistency over time, of decision, of cue, of meaning. Five things settled, and the fifth one sat there with no argument around it, the way the obvious things do when everyone has agreed and moved on.

Then the disagreement, the one they’ve been having for a decade, came back in the same shape it always has. Sharp says distinctiveness, being recognizable. Ritson says distinctiveness plus relative differentiation, being recognizable while carrying a meaning that a buyer can attach to. Sharp says the data shows that the difference barely matters. Ritson says when you can get it, like Volvo with safety, it’s worth defending. Five things settled. One thing not settled. And the one thing not settled was positioning.

The whole famous war between reach and positioning had been reduced to a single unresolved question: positioning. Everything around it was settled. The thing nobody could agree on was the one thing the stage couldn’t resolve, and I wanted to know why it kept slipping out of everyone’s hands. So I went looking at who else was talking, and the talk was everywhere that month.

The first door

Erich Joachimsthaler put up a post that took the whole argument up a floor. His move was to point out that the entire penetration debate is happening at the wrong altitude. Mars had recently paid around thirty-six billion dollars for Kellanova. Erich’s point was that Mars did not spend that to buy more penetration for candy bars. Pringles, Cheez-It, Pop-Tarts, those are brands, and the brands were not the point. Mars bought access to a larger demand space. Snacking. Penetration is an operating metric inside a market. It is not a way of deciding which market deserves your capital. A brand can grow penetration and still destroy value by winning more buyers in a structurally weak space at thin margins. The better question sits above penetration entirely. Where can the company build demand, earn pricing power, and compound economic profit over time?

A few days later, he sharpened it. He went after Category Entry Points directly, the idea that Sharp made famous, the buying situations that trigger a category purchase. Erich’s argument was that the older idea of brand salience, going back to the early nineties, was broader. A great brand becomes mentally accessible throughout life, not only at the moment of purchase. Sharp narrowed that to the entry points, the specific triggers. And the narrowing, Erich said, dropped something. People did not suddenly develop a Category Entry Point for SUVs. They wanted safe family travel, adventure, and freedom, and the category formed around that pre-existing want. Nobody woke up wanting a smartwatch. They wanted health, fitness, and self-improvement, and the category followed. The demand space came first. The category came after. Entry points explain how a brand captures existing demand. They do not explain how a brand creates demand that doesn’t exist yet.

Erich tells you which room to walk into. He tells you to pick the demand space before anyone else can see it’s a demand space, while it’s still a tension, a habit and an aspiration with no product attached. That’s real, and it’s upstream of everything Sharp measures. But the demand space, once you’ve proven it, becomes a thing everyone can see. Once SUVs are obviously selling, every carmaker can read the sales chart. So choosing the space gets you in early. It does not tell you how you keep the position inside that space when the money arrives, and twelve competitors who can also now see the space come in behind you. He named the front door. He didn’t name what stops the next person through it from taking the room.

The two doors at Cannes

Sharp and Ritson are the second and third doors, and they’re standing closer together than the decade of argument suggests.

Sharp tells you to be available and recognizable. Be in the cooler, be on the shelf, own a colour, a sound, and a shape, so the brand surfaces fast when the buying moment arrives. He has a wall of data behind it, and most of the data is right. The brand that comes to mind first, in most situations, sells the most. He plays down the difference because his evidence shows it’s hard to find, easy to lose, and rarely the thing that moves volume.

Ritson stands a foot to the left. He takes everything Sharp says about availability and adds the half Sharp discounts. When you can get a buyer to attach a meaning to you, the way a buyer attaches safety to Volvo, that meaning is worth having and worth defending. He calls it relative differentiation. Same cooler, same shelf, same need for recognition, and, on top of that, a meaning that makes the buyer choose you over the cheaper option next to you.

Both of them are telling you how to win the moment of choice once you’re already in the category. Sharp on recall. Ritson on recall plus meaning. Neither one tells you where the meaning comes from or what keeps it yours after a funded competitor decides to claim it. They tell you the meaning is worth having. They stop before the question of what builds it.

The fourth door

Then Jelena Veselinovic wrote something that came at the same thing from the opposite side, from inside the product. Her post was a swing at the Silicon Valley line that advertising is the tax you pay for being unremarkable. She said the line makes her cringe, and then she said the opposite idea is equally wrong, the idea that you pour advertising on top of a finished product and wait for the business to grow. The product, in her words, is the strongest brand lever you will ever have. The strongest companies don’t add the brand later. They bake it in before a single ad is written. Coca-Cola knew this in 1915 when it briefed a bottle so distinctive you’d know it in the dark, or recognize it from a single shard if it shattered on the ground. Apple knows it daily, taking components anyone can buy and turning them into the most wanted objects on earth. And then a brand she’d stumbled on, Ritual, a supplement so well made you stop comparing it to the hundred cheaper options sitting next to it on the shelf.

I agree with most of that. The brand is built into the thing, or it isn’t built at all. The proof is the product. She’s standing on ground I’ve spent years defending. Looking more closely, I noticed that the word she used and the word she meant were two different words.

She said distinctiveness. Everything she described was difference. Those are not the same thing, and the gap between them is the whole argument she didn’t know she was having. Distinctiveness is being recognizable. The bottle you’d know in the dark. The colour, the shape, the asset that makes a brand easy to spot. Difference is being un-substitutable. The Ritual supplement that makes you stop comparing. Sharp’s camp says be distinctive, be recognizable, and openly plays down difference. Jelena says distinctiveness is everything and then proves difference with every example. She’s using Sharp’s words to make Ritson’s case, and the seam between them is the same question Cannes left open.

The Oxford team at Saïd Business School modelled 872 brands from 2006 to 2022 using Kantar’s BrandZ data and found that Difference accounts for 35% of brand impact on share-price outperformance, and Salience for 0.6% (Kantar Think Different). The same body of work shows Difference matters about 2.5 times more for Pricing Power than for Demand Power.

I’m going to name the conflict on it before I use it for anything. Kantar sells difference measurement, so the model rewards what the seller is paid to find. The Difference variable in their definition includes “leading the way,” which is partly an outcome of winning rather than a cause of it. The dependent variable is share price, not buyer behaviour, so the chain runs through investor sentiment too. None of that makes the number wrong. It does mean the number can’t carry an argument on its own. It points in a direction with a thumb on the scale, and the thumb has to be named every time the number comes up. Even taken at face value, notice what it doesn’t tell you. It doesn’t tell you what stops your difference from being copied next quarter.

That’s where Jelena stopped. She tells you the answer lives inside the product. Ritual made a beautiful supplement. But beautiful supplements are copyable by morning, by her own rule, the rule she opened the post with. So what stops the next funded brand from making an equally beautiful one and collapsing Ritual’s advantage back into a comparison? She named where difference has to live. She didn’t name what defends it once it’s living there.

The fifth door

I kept going, because there was one voice I couldn’t leave out, the one whose whole argument runs against mine.

Rory Sutherland has spent his career on what he calls psycho-logic, the idea that much of what looks irrational about how people choose is rational at a level a spreadsheet can’t see. His examples are the ones that stick. An expensive painkiller relieves more pain than a cheap one with the same molecule, because the price changed what the body expected. Eurostar wanted to spend a fortune making the train run faster, and his line was that you could spend a fraction of that making the journey feel better, free champagne and good company, and people might ask you to slow the train down. Some goods sell more when you raise the price, because the price is the product.

That should worry me, because my whole argument says perceived value sits downstream of costly decisions repeated over years, and Sutherland has a stack of cases where the feeling came first, and the substance came later, or never arrived.

Reading him slowly, I found there are two Sutherlands, and I was about to collapse them. The expensive painkiller and Eurostar Sutherland are mostly about cheap perceptual moves, the kind of thing any operator with imagination can borrow by morning. That Sutherland is the opposite of my argument, and I should say so. The other Sutherland is the one who writes about costly signalling, the argument that advertising works partly because it’s expensive, and the expense itself is the proof. That Sutherland is parallel to my argument, on one slice. Not identical. He’s making it about ad spend as the signal. I’m making it about operating decisions as the signal, taken every day and accumulated over the years. Cousins, not the same person.

The way I account for cheap perceptual moves is by specifying what they can and can’t do. They can move a single transaction. They cannot hold a position against a competitor doing the same trick next Tuesday. The expensive-aspirin effect works for every aspirin brand at once because it’s a fact about human wiring, not a moat. Anyone can charge more and call it premium. A cheap perceptual trick is, by its nature, cheap to copy. Sutherland’s main case concerns how perceived value is created. He says less about how it stays yours. He owns how the feeling starts. He leaves the rest of the question alone.

The sixth door, and the floor under all of them

By this point, the same gap had opened from every direction I looked, and I wanted to get underneath it. So I stopped reading the debate and asked the older question. Why does a buyer’s mind work in a way that lets any of this happen at all?

That question has answers, older than the marketing fight by decades, and they come from fields that weren’t talking to each other.

A psychologist named George Miller worked out in the 1950s that working memory holds about seven things at once, and later work pulled that down to three or five for the part that really counts. The number doesn’t matter as much as the ceiling. It’s low. In any category a person can name, they can hold a small handful of brands in mind without effort, and everyone else is invisible. You’re not competing for the whole of someone’s attention. You’re competing for one of a few slots.

Eleanor Rosch worked out in the 1970s that categories don’t have clean edges. They have centers. A robin is a more obvious bird than an ostrich. People don’t run a checklist when they sort the world; they judge by how close a thing sits to the center of its category. A buyer doesn’t pick by comparing features. They reach for whatever sits closest to the center of the category in their head, and whoever owns the center owns the category.

Amos Tversky and Daniel Kahneman worked out around the same time that when people don’t know, they go with what comes to mind easily. A thing that surfaces quickly feels more common, more likely, more right. Being remembered first does work the product can’t do. The brand that comes up easiest gets chosen, and it feels like a reasoned choice from the inside, even when it isn’t.

Two economists, Edward Chamberlin and Joan Robinson, working separately in 1933, both wrote down that real markets are not perfectly competitive. Sellers differentiate, buyers form preferences, and small differences create little monopolies inside the buyer’s mind. The part most people skip is the second half of what they found. Those little monopolies don’t last on their own. Once a difference is visible and worth having, rivals copy it, and the advantage gets competed away unless something keeps them out.

That last sentence is the whole problem, and it’s where the sixth door comes in, the oldest one, the one that was standing there before any of the five showed up.

Michael Porter walked up to that door in 1996 and came closer to the room than anyone else. His answer to what keeps the little monopoly from being competed away was a real answer. Strategy is choosing what not to do. A defended position isn’t one clever choice; it’s a system of choices that fit together and reinforce each other, and the fit is what makes it hard to copy, because a rival has to match the whole system at once and not just one piece of it. The sacrifice is the structural part. You don’t own a place by adding, you own it by refusing, and the refusal is the half that most people skip. Roger Martin turned that instinct into a discipline in Playing to Win: strategy as an explicit set of cascading choices, starting with where to play and how to win, rather than a company that tries to be everything to everyone, making no real choice at all. And Jay Barney, working the same decade from the resource side, wrote down why some advantages persist. They persist when they’re built through a path a rival can’t retrace cheaply, when the cause of the advantage is hard to see from outside, when you can’t buy the result without living through the history.

Read those three together, and you have most of my argument already on the page, decades early. A position is a system of costly choices, welded so the pieces hold each other up, defended because a rival would have to retrace a history he can’t shortcut. So why isn’t the room full? Why did Porter stop at the threshold, too?

Because all three of them described the position from the firm’s side of the glass. The system of activities. The choice cascade. The inimitable resource. Every one of those lives on the company’s balance sheet and in the company’s operating model. None of them crossed the glass into the one place the position has to exist to be worth anything.

The buyer’s head.

Porter told you how to build a defended position. He never told you that the defence only matters because it gets written into a stranger’s memory, one decision at a time, and that the writing is the asset. The strategy people built the half about what makes an advantage hard to copy. The cognitive people built the half about how a buyer’s mind selects and remembers. The two halves were sitting in different buildings for thirty years, and nobody walked them across the street to each other.

The thing that walks them across the street was set down in 1949, before either side had the language for it. Donald Hebb wrote that when one cell repeatedly takes part in firing another, the connection between them strengthens, and when the firing stops, the connection fades. Carla Shatz, decades later, condensed it into the line everyone now quotes: “cells that fire together, wire together.” The mechanism predated the catchphrase by forty years, which is the thesis of this essay in miniature. Apply it to a buyer. Every time a company makes a decision that proves a concept, the concept and the company wire together a little tighter in the people watching. Every time the company makes a decision that contradicts it, a competing pattern fires, and the original wiring gets diluted. The brain doesn’t average them. It gets confused by them.

But Hebb is the place I have to slow down, because Hebb is easy to over-read, and the over-reading is where most of the bad thinking in this field lives. Hebb tells you that co-activation builds the wire. He says nothing about what the co-activation does or what it costs to make it happen. And that second question is the one that matters, because two very different things build wire the same way.

Both kinds are built by repetition. The buyer who sees the same logo in the same places ten thousand times wires a connection through sheer exposure. The buyer who watches a company refuse the easy money, kill the product that doesn’t fit, and build the thing the category said couldn’t be built, again and again for a decade, also wires a connection through repetition. Same mechanism, co-activation over and over until the pattern holds. So repetition can’t be what separates them, because both kinds repeat.

What separates them is what’s being repeated, and what it costs. A logo on repeat is cheap to lay down, because anyone with a budget and enough years can lay it, and cheap to copy, because the next brand with a budget and enough years lays the same kind right beside it. That cheap kind is mostly what Sharp is measuring. A decade of costly decisions runs counter to both counts. It’s expensive to lay down, because the decisions cost real money and real revenue, and expensive to copy, because a rival has to live through the same decade of the same choices to wire the same thing. That’s the whole distinction. The cost and copyability of whatever is on repeat. A budget buys the first. The second takes a decade, no budget shortcuts.

Hold that, because I’m about to walk into the cleanest experiment in the pile, and the experiment proves that wiring beats the product. It does not, on its own, tell you which of the two kinds of wiring it’s looking at. And the whole field reads it as if it does.

The soda experiment

People reach for this one when they want to prove that brand power is a real force, not a story marketers tell themselves.

In 2004, a team at Baylor ran sixty-seven people through a soda test (McClure et al., Neuron, 2004). With the labels hidden, two unmarked cups, one Coke, one Pepsi, taste and choose. Preference split roughly evenly. Then they did something sharper. They gave people two identical cups, both Coke, and put a Coke label on one of them. People picked the labelled cup about three-quarters of the time, choosing one Coke over the same Coke because of a label. Then they ran the same setup with Pepsi. Two identical cups of Pepsi, one labelled. The label moved choice much less. The asymmetry was the result. The Coke label did real work. The Pepsi label barely did any.

When they scanned the brains, the regions that lit up when the Coke label worked were memory regions, the hippocampus and the dorsolateral prefrontal cortex. Four years later, Michael Koenigs and Daniel Tranel at Iowa took the same setup to patients with damage to one specific area of the brain, the ventromedial prefrontal cortex. Their finding, published in 2008, was that the brand-driven shift in preference disappeared in those patients (Koenigs and Tranel, SCAN, 2008). Damage that one piece of wiring and the label loses its grip.

Here’s what this shows, stated as narrowly as the evidence allows. A memory trace is doing real work behind the Coke label, a much weaker one behind the Pepsi label, and when you take out the brain’s machinery for turning brand cues into preference, the effect goes with it. Choice in the cola aisle runs on a wired memory, and barely on the sugar water. That much is solid, and it’s enough to kill the idea that fast, cheap categories are purely about availability with no meaning.

Here’s what it does not show, and what I almost let it show. It does not tell you what built Coke’s trace. I was ready to put Coke at the front and the back of this essay as the clean proof of my whole argument, the proof that costly coherent decisions wire a defensible position. Then I ran my own method on Coke and it came back ambiguous, and the ambiguity is the honest part. Coke has a century of being everywhere, which is the cheap repetition kind of wiring, the kind anyone with a hundred years could lay. Coke also has real costly structure underneath it. I cannot cleanly separate the two from the outside. Coke’s trace might be mostly the expensive kind, mostly the cheap kind, or some split I can’t measure. The soda study proves the trace is there and is doing the work. It is silent on which kind of wire it is.

I want to be plain about what that costs me, because it hands Sharp the most famous brand-power case on the board. If Coke’s wiring is mostly a century of ubiquity, that’s Sharp’s exact mechanism, cheap repetition at a planetary scale, and a Sharp reader is right to take it. I’m spotting him Coke on purpose, and the reason it’s cheap to spot is that Coke barely out-prices Pepsi at the shelf. The two are priced in lockstep, fighting on availability and promotion, not on a premium. So Coke was never a pricing-power case to begin with. It’s the wrong brand to test a margin claim on, which is exactly why I can give it away without losing anything I need. Coke proves that choice runs on memory. Coke does not prove my argument about what kind of memory defends a premium. For that, I need a case where the costly structure is legible and cheap repetition can’t account for the result; I’ll get to it.

The shape

By the time I’d read all of them, I could see the shape, and it was the same shape every time. Erich tells you which market to enter. Sharp tells you how to be available once you’re in it. Ritson tells you to add meaning beyond being available. Jelena tells you the meaning lives in the product. Sutherland tells you the feeling can be built directly. Porter tells you a defended position is a welded system of costly choices. Six doors. And underneath all six, the same floor. The buyer holds a few brands, reaches for the one at the center, picks the one that comes to mind easiest, and runs the whole thing on a wired memory.

Every door opens onto the moment of choice. How the buyer chooses, once the choosing is happening. Not one of them is about the years before the choice, when the wiring that decides the choice was being laid down. Porter came closest, and even Porter described the welded system from the firm’s side and never followed it across the glass into the memory it’s supposed to build. The room they’re all standing around is the same one. It’s the place where a company’s costly decisions get written into a stranger’s memory and held there, and where that writing turns into the thing a competitor can’t cheaply copy.

Six doors around one room, and every one of them stops at the threshold.

Why nobody walks in

The room stays empty for a reason, and the reason is structural, and it splits in two.

The five practitioners stop at the threshold because of what they sell. Walking into the room means telling a company that the work isn’t a deliverable. The work is the next ten years of decisions. The work is who you hire, what you refuse to build, what you walk away from when the easier money is right there, and the discipline to keep doing it after the founder has stopped looking. Nobody can sell that on a six-week engagement. Nobody can put it in a deck. Nobody can finish it on a Friday. So an entire industry stops at the threshold and sells the part it can ship. Sharp can sell mental availability because it shows up in a tracking study. Ritson can sell distinctiveness because it can be audited in a brand book. Jelena can sell product craft because it has a shipping date. Sutherland can sell perceptual moves because they fit on a slide. Erich can sell market selection because it shows up in a board memo. All real work. All shippable. None of them is the room.

Porter and the strategy people stop for a different reason. They weren’t selling a six-week engagement; they were describing how advantage works, and they described it accurately. They stopped because the question they were answering ended at the company’s boundary. What makes an advantage hard to copy? They answered it. The question they didn’t ask was where the advantage has to live to be worth defending, and the answer to that one is in a place economists and strategists don’t usually look, inside the head of a person who will never read their work. The cognitive scientists were looking inside that head the whole time and never asked what a company’s capital allocation had to do with it. Two correct halves, in two different buildings, and the bridge between them is unsellable, which is most of why nobody built it.

There’s a political fact under the commercial one, too. The work in the room is a CEO’s job. It’s capital allocation, hiring, the things you refuse, the markets you walk away from. Nobody who isn’t the Founder or CEO can do it, including me. The most anyone outside the company can do is help an operator read what the operator’s own decisions already say. The room stays empty because outsiders can’t live in it, even when they’re invited in. The work has to be done by the person whose name is on the capital. It can’t be packaged. It can only be done.

What keeps it yours

The room raises the obvious question. How costly is it, and why does the cost matter? Could a company skip it and achieve the same result more cheaply?

There’s a model for this, and it’s older than marketing. Michael Spence wrote it in 1973 while working in information economics and won the Nobel Prize in 2001 for the work. A signal is only believed when it’s expensive enough that a weaker player couldn’t afford to fake it. A degree signals capability because the cost in time and effort is one only a capable person would pay, with the productivity gain a smaller part of the story than the price of admission. Amotz Zahavi arrived at the same model, studying biology. The peacock’s tail is reliable as a signal precisely because it’s costly, costly enough that a low-quality bird could not afford to grow one. Two disciplines, no contact, the same answer. Costly signals are credible because they’re costly. Cheap talk is ignored because anyone can produce it.

Apply that to a buyer reading a company. The buyer is not running a signalling model on purpose. They don’t trust what the company says about itself, but they do trust what they have seen the company do. Self-description triggers the part of their head that asks, “Who says?” The claim and the skepticism arrive in the same breath. A decision the company would bleed to fake, a refusal that cost real revenue, a product killed instead of harvested, that lands without skepticism because the cost itself is the evidence.

There’s a step in the middle I have to be honest about, because the chain breaks without it. The buyer never sees your capital allocation. They don’t read your income statement. They can’t watch you kill the product or turn down the deal. The costly decision only wires the buyer when it surfaces in something the buyer can perceive. Costco’s markup cap is invisible as a number. What the buyer experiences is that the prices feel low, every visit, for years. Patagonia’s trust structure is a legal document almost no customer has read. What the buyer experiences is the repair counter, the jacket that lasts, and the ad that told them not to buy it. So the real chain has four links, not three. A costly decision, then a perceptible proof of that decision, then the proof repeated until it’s a pattern, then the pattern wired into memory. The buyer reads the consequences of the money when the consequences are visible. The operator is the only one who can read the money itself. Hold both of those, because the Monday test at the end depends on the gap between them.

The same cost that makes the signal credible also makes it durable. A rival can copy a claim by morning. A rival cannot copy a decade of refusals without living through the decade. This is Barney’s inimitability, finally routed to the place it pays off. The buyer doesn’t have to do the inimitability calculation in their head because their head was wired by the cost as it happened, and that wiring is what they reach for at the moment of choice. Credibility and durability come out of the same substrate. That’s the part the field keeps splitting into two arguments when it’s one.

So the room now has a mechanism. The room is where the company runs its operating model, producing costly signals every day for years. The signals surface as things a buyer can see. The wiring in the buyer’s mind accumulates from those visible signals one decision at a time. The cost makes the signal credible the moment it lands and inimitable across the years it takes to build. Pull the signals, and the wiring fades. Replace the signals with claims, and the wiring never builds in the first place.

The force

This is where the thing I’d been circling for the whole week finally sits still. People keep reaching for a force. Brand power. Brand magic. Something beyond what the spreadsheet can measure, the thing that makes the Coke label work and the Pepsi label fail. They reach for it because they can feel it doing something, and they can’t see what it’s made of, so they name the feeling and treat the name as an explanation.

The scan turned up no separate force anywhere. Only memory, firing the patterns it was built to fire. Damage the wiring, the effect goes with it. The label moves you in proportion to what was wired behind it over the years, which is why Coke’s label moves you, while Pepsi’s, with less consistent wiring behind it, does not.

The thing they were hunting was the absence of a thing. There was never a separate force floating above the business. There was the company’s pattern of decisions, surfacing where the buyer could see them, firing when the cue arrives, fading when the pattern fades, and deepening when the pattern holds. The force is the wiring, and the wiring is what the cue triggers, and what the cue triggers is what the company spent the last decade paying for.

Reading it backward

I tried to build the whole thing backward, starting from the only moment that’s real, the moment someone pays.

A person pays for the option that, right at the moment of deciding, feels like the least risky and best-fitting answer to whatever they were trying to do. That’s the end of the line. Everything upstream exists to produce that feeling at that moment.

For the option to feel that way, it had to carry a meaning that fit the need, and no rival could carry that meaning as strongly, or the choice gets harder, and price starts to matter. For that meaning to be in the room at all, the brand had to come to mind when the need showed up, because a brand a buyer can’t recall can’t be chosen. For coming to mind to be worth anything, the need and the category had to exist in the first place. And for the meaning to be wired strongly enough to survive a competitor copying the surface of it next quarter, it had to be built through a pattern of costly decisions the company made every day for years, decisions a rival would need the same years of the same discipline to match.

Read in that order, it’s a stack. The demand space at the bottom. The category formed on top of it. Recall inside the category. The meaning that gets recalled. And the ongoing pattern of costly decisions that wires the meaning into memory and keeps it wired. Each person in the debate stands on one stair. Erich on the demand space. Sharp and Ritson on recall. Jelena pointing at the meaning stair and saying it lives in the product. Porter describing the staircase as a welded structure without following it up into the head at the top. The top stair, the costly pattern that wires the meaning and keeps wiring it, is the one with nobody on it.

The economic payoff of the whole stack is not one word, and I used to think it was. I want to say pricing power and stop. The cleaner term is Erich’s. Compound economic profit over time. Sometimes that shows up as margin, the premium a buyer pays before substituting. Sometimes it shows up as retention and repeat, and the cost of acquisition falls year over year. Sometimes both. Volume mostly comes from reach, and being easy to find, and Sharp’s people have that half right. What the costly substrate buys is the durable half, the profit that holds when a competitor shows up with a functionally similar thing and a similar story. Strip the cost out from underneath, and the story is copyable, and the durable half goes with it.

I’m leaning toward the split on the mechanism, not on the data, and I want to be plain about why. The only industry number close to on point is the Kantar finding, that difference matters more for pricing power than for demand power, and I already named why that number can’t carry an argument. Seller of difference measurement, a variable that smuggles in the outcome, a dependent variable that runs through the stock market. It points the same way the mechanism points. It does not prove the mechanism. The split between the durable half and the volume half rests on the logic of costly signalling: a cheap signal moves a transaction, and an expensive one builds a wall; the compromised number is a thumb on the scale, agreeing with the wall, nothing more.

And there’s a condition that the stack makes obvious the moment you read it as a stack. The costly pattern at the top is necessary. It is not sufficient. If the demand space at the bottom is too thin, or the category sitting on it is too small to pay, then a decade of costly coherent welded decisions wires a beautiful position that nobody pays enough for. You can do everything in the room exactly right and still earn no premium, because the room was built on top of a market that was never big enough to matter. That failure is real, and I’ll come back to it, because it’s the one the whole argument has to survive.

The category everyone hands to availability

The reflex move at this point is to concede a piece. Draw a line at category involvement, hand the considered buys to the costly-signal stack, hand the fast, cheap shelf to Sharp, and call that modesty. I tried to write it that way. The cases won’t let me, and the worst of them is the one I already put at the front of this essay.

The McClure soda study is a Coke study. Coca-Cola is the cola aisle, and the cola aisle is the purest grab-and-go shelf there is. The whole result was that the wired memory, and not the sugar water, moved choice three quarters of the time. Then I was about to write that for the cold drink off the shelf, there’s a hand and a habit, as if the habit were the absence of meaning. The habit is the wiring firing. I had spent the best evidence proving that meaning wins even in the cola aisle, and I was about to hand the cola aisle to Sharp three sections later. The boundary was the bug.

Coke proves wiring beats the molecule there. It can’t, on its own, prove the wiring is the costly kind, for the reason I already gave. So I need the case on the impulse shelf where the costly structure is legible, and a budget alone can’t account for it, and Red Bull is the cleanest one anyone could ask for. It pours somewhere between a quarter and a third of its revenue into marketing, on the order of three billion euros a year, against an industry norm in the single digits, though the exact figures are estimates because the company is private and reports nothing. The money doesn’t buy ads in the ordinary sense. It buys Formula 1 teams, the man stepping off the edge of the stratosphere, a media house that produces its own films, and three decades of association with anything fast and dangerous. This is the distinction from the Hebb section made concrete. Monster can buy advertising. Monster can match a sponsorship line for a season, which is the cheap repetition kind of wiring. Monster cannot reproduce the media house, the sports portfolio, or thirty years of the same meaning held in the same place, which is the costly kind. Same cooler, same impulse, and a per-ounce price and a margin Monster doesn’t reach. If the costly signal stack only worked for considered buys, Red Bull would be impossible. Red Bull is the strongest exhibit in the whole argument. Strip the logo, read the spending, and the money says energy before anyone mentions the can.

Liquid Death is the lighter second case, and worth keeping for what it proves rather than what it spends. It sells canned water for two to three times the price of the bottle next to it and has reached a $ 1.4 billion valuation doing so (Bloomberg). Its signal is cheaper than Red Bull’s, mostly creative, with one real operating refusal underneath it, aluminum instead of plastic, the enemy it named on day one. It built a following on stunts for years before buying its first national Super Bowl spot in 2025, a roughly seven-million-dollar buy it had pointedly avoided until then (CNBC). Commodity water, premium price, a meaning carried by relentless creativity, and one refusal, a rival would have to retool a supply chain to match.

So the line that matters isn’t where the category sits. What builds a position is a costly, hard-to-copy signal, and the cost can sit in operations, like Costco capping its own markup, or in marketing, like Red Bull’s media operation, as long as it’s high enough that a rival can’t cheaply match it. The lane doesn’t matter. The cost does. What does not build a position is the claim on the homepage or the ordinary ad anyone with a budget can run, because that’s the cheap repetition wire, and the next budget lays the same one. The homepage was never the thing.

Now, the seventeen-year study, read carefully, because at a glance it buries me. Distribution and assortment dominate brand equity among 325 national brands across 35 categories (Rajavi, Kushwaha, and Steenkamp, Journal of Marketing, 2023). That’s an average across 325 mostly undifferentiated brands. Most brands in fast-moving categories never built an exclusive meaning, so availability is the only lever they have, and availability is what the data measures. The median is the median. Red Bull is the tail that the average hides. The study captures what most brands do and says nothing about the ceiling. This is the true half of Sharp’s case, and I’ll grant it without flinching. Most brands, in fast-moving categories and everywhere else, don’t escape the availability game. The claim he doesn’t get to make from this data is that they can’t.

The durable-half-versus-volume-half split does the reconciling, and the cleanest proof of it is Red Bull back on the impulse shelf. Its ubiquity is availability, the volume half, and Sharp is right that the cooler being everywhere is what moves the cans. The premium it holds over equally available rivals is the meaning, the durable half. Even in the purest impulse category the field can name, the part availability can’t explain is the premium, and the premium is the costly wiring. The model isn’t strained on the fast shelf. The fast shelf is its best demonstration.

The cases that look like slogans

The same shape holds where the costly signal sits more obviously in operations, and these are the cases that look like slogans until you read the books.

Costco holds the idea that it’s on your side by capping its own gross markup in the mid-teens, where most retailers charge two or three times that, and making its money on the annual membership instead. A rival can copy the slogan in an afternoon. It cannot copy the cap without tearing up its own income statement, because the cap isn’t a line in a brand book; it’s the architecture the whole business is built to survive on. And Costco is worth pausing on, because it looks like it breaks my own split and ends up sharpening it. The costly signal here is a refusal of margin. So the thing it buys can’t be margin. What the cap buys is trust, renewal, a member who walks past three cheaper-looking options because the price is already known to be fair, and a renewal rate that turns into the most reliable profit stream in retail. That’s compound economic profit arriving as retention instead of as price. The split runs between the durable profit half and the raw volume half, and durable profit wears more than one outfit. Costco wears the retention one. Patagonia wears closer to the margin one, holding the environmental idea by pledging one percent of sales to environmental causes since 1985, co-founding 1% for the Planet in 2002, running a repair operation instead of selling new gear, telling people in a Black Friday ad not to buy the jacket, and in 2022 transferring the entire company to a trust so that profit not reinvested goes to fighting climate change. A competitor can run the same ad next week. It cannot run the same four decades.

Notice what those two share, because it matters in a minute. In both, the position is welded into the operating model so deeply that the business is built around it. That weld is the reason a rival can’t lift the position out and wear it. Hold that thought.

What challengers reveal about decay

There’s a related point that earns its own paragraph, because it’s a different mechanism running in reverse. On and Alo are challengers, not incumbents, and they take premium share inside categories where the incumbent had the availability advantage Sharp prizes.

On did it with a product difference under the meaning, the CloudTec sole, so the costly signal is in the engineering as much as the marketing, and the proof is on the bottom of the shoe. Alo is the richer case because the signal is mostly aesthetic, and the case still works. It grew from 0.4 percent of the US athleisure market in late 2021 to 1.3 percent three years later (Earnest Analytics), while Lululemon held around 21 percent, with revenue near 9.6 billion dollars, compared to Alo’s roughly 1 billion. So Alo didn’t unseat Lululemon. At roughly a tenth of the revenue, it took the cultural lead and the premium momentum while Lululemon’s meaning aged, its stock fell by close to half through 2024, and the coverage started describing its best decade in the past tense.

The wiring you stop firing stops being wiring. A challenger with a more relevant meaning takes the premium first and the share later, if the incumbent doesn’t reset. That’s the decay half of the mechanism in motion. The same logic that builds a position pulls it apart when the costly decisions stop being made.

Trying to break it

I keep trying to break this, because an idea you can’t break isn’t worth holding. There are four ways to come at it, and the second and third are the ones that can land.

The first way tests whether the cost is necessary. Find a brand in any category that holds a price on an owned concept with nothing costly underneath it. Claims on a homepage, no matching decisions in the books, and still the premium, year after year. If that brand exists, the wiring argument is wrong, and brand power is a free-floating force you can summon with language. I’ve looked hard and haven’t found one. Every brand that looks like it pulled the position out of thin air turns out, when you read its decisions instead of its slogans, to have paid for the word somewhere, in operations or in marketing. The lane changes. The cost doesn’t. But notice what this test checks. It checks that the cost is necessary. It does not check that the cost is enough.

The second way tests whether the cost is enough, and this is the one I have to take on the chin, because the answer is no. Costly, coherent, welded, and still no premium. It happens, and the stack already told you where. When the demand space at the bottom is too thin, or the category is too small to pay, a company can run a decade of disciplined, costly decisions, wire a real and exclusive meaning, and earn nothing for it, because there were never enough buyers who valued the thing it came to own. That company doesn’t die from a competitor copying it, nor from a reframe. It dies on demand.

Here’s the honest trouble, and I’m going to make it a claim instead of an excuse. This failure mode is nearly invisible from the outside by its very nature. A company that wires a real position on top of a market too thin to pay for it doesn’t leave a clean record saying so. It looks, in the obituary, like an ordinary flop, and the cause of death gets written up as bad management or bad timing or a competitor, because those are legible and thin demand for an owned meaning is not. So the cases that would most cleanly prove cost isn’t enough are exactly the cases survivorship erases. The proof is structurally hard to produce, and the theory predicts that difficulty rather than hiding behind it.

The closest instance I can hand you is Vertu, and I’m handing it over with the contamination flagged, because the contamination is the lesson. Vertu made handmade luxury phones, sapphire screens, a concierge button wired to a live human, a meaning it owned, costly and coherent for nearly two decades. It went into liquidation in 2017, with around 200 jobs lost, after selling phones that started at around $14,000 and topped out in the six figures. Read one way, it died on thin demand because the set of people who would pay that much for a technically mediocre phone was never large enough to support the structure. Read another way, it died on a reframe because the iPhone shifted the value axis from materials and concierge to software and ecosystem, and the obituaries split almost evenly between the two stories. I can’t separate them cleanly, and neither could anyone writing at the time. That inseparability is the whole point. A costly owned position dying on thin demand almost never arrives labelled, because by the time it dies, an axis has usually shifted too, and the shift gets the credit. The clean example is rare for a reason the theory explains.

So the discipline I owe you, the reader, is to say plainly that the costly pattern is necessary and not sufficient, that the demand space underneath has to be real and big enough to pay, and that anyone selling the room without saying so is selling half a map. Cost builds the wall. It can’t conjure the territory the wall is supposed to enclose.

The third way is the sharpest, because it comes for the survivors. You can own a position, pay for it with a decade of costly decisions, win, and still die. BlackBerry owned the keyboard. Kodak owned film and invented the digital camera that killed it. Axe owned the mating game in India for more than a decade, and not as a slogan. The position was carried by everything at once: a patented spray nozzle, the casting of the nerd who sprays the can and gets swarmed, the events, and the comms held in one line for years. Then it fell out of the top five and effectively ceded the market. If a position can be that owned and that costly and still die, the argument looks finished.

It isn’t, and the reason is the same fact read from the other side. The weld that makes a position impossible to copy is the weld that makes it impossible to drop. Costco can’t un-cap its markup without tearing up its income statement, which is the exact reason no rival can copy the cap. Axe couldn’t walk away from aerosol and seduction when the Indian market moved, for the exact reason that no competitor could lift that position off it. The business was built around those refusals. Its parent is Hindustan Unilever, which owns the deepest distribution in the country, so the loss didn’t come from the shelf. FOGG never made a better mating-game spray. It changed the axis. No gas, more sprays per bottle, a deodorant positioned as an everyday thing with Indian sanskars baked in, and it grew the market past the narrow space Axe’s meaning could reach (ThePrint). Axe’s patented nozzle was defending an aerosol can; the challenger simply abandoned. The costly asset was guarding the wrong layer.

So the argument doesn’t predict that strong positions live forever. It predicts how they die. Strong incumbents fall to challengers who change the game and almost never to challengers who copy it. Nobody out-keyboarded BlackBerry. Someone made the keyboard beside the point. The most welded fall hardest, which is why Kodak invented the future and died of its past. And they fall only when the valued axis shifts, not before. That’s a falsifiable shape. If strong incumbents fell to direct copies as often as to reframes, the argument would be wrong, and they don’t. The refusal that builds the moat is the refusal you can’t take back. The moat and the trap are one wall, read from two sides.

There’s a partial exception worth naming because it bounds the claim rather than breaking it. A welded firm can sometimes survive the axis shift, by tearing up the weld itself before it kills them. IBM nearly died in the early nineties, welded to mainframe hardware, and Lou Gerstner pulled it off the thing it was built around and rebuilt it on services. That’s rare, it’s brutal, and it’s exactly as expensive as the theory predicts, because you’re paying to unbuild a decade of structure and build another. It’s the same move the winning challenger has to make eventually, when its own fresh meaning ages and its own weld becomes its own trap. The challenger who wins doesn’t escape any of this. It builds the next position, pours its own costly structure into new ground, and waits for the day that structure is the thing holding it in place.

The fourth way is the one I have to watch in myself, because it’s the way the argument rots from the inside. It can become unfalsifiable in the wrong hands, and the wrong hands might be mine. The same logic that says to discount the founder’s words because they’re confabulated can be used to wave off any inconvenient finding by relabeling it. Read a position backward from a winner, and everyone who won looks like they meant to. The discipline that keeps it honest is that the reading has to be able to come back empty, and earlier in this essay, it did. I wanted Coke to be clean proof, and it came back ambiguous, and I said so, and the ambiguity stayed in. That’s the test working. Sometimes a company’s decisions don’t cohere, no identity is transferred into the operating model, and there’s no pattern to find. The honest answer then is that the company owns nothing yet. A method that always finds a clean noun is doing horoscopes.

There’s one real floor left, the true commodity, a bulk input where the buyer can’t perceive or care about difference. Even that one is thinner than it sounds, because Sutherland spends his career showing perceived difference can be manufactured almost anywhere. But far enough down, the wiring stops mattering because the buyer stops looking. The cola aisle is not on that floor. The brain scan settled that at the beginning of this essay.

Finally

If you run a company, there’s a version of all this you can do on a Monday morning, and it doesn’t require any of the theory.

Pull your last twelve months of spending. Not the deck. Not the homepage. The actual capital. Where it went, what it bought, what you said no to. You can read the money directly, which is the one thing your buyer can’t do. They only get the consequences of it, the product, the price, and the refusal made visible. You get the source. So use the access. Take your logo off the spending and hand the whole thing to a competitor’s CFO, someone who has never seen your marketing. Ask them one question. From the money alone, what word would you guess this company is trying to own?

If they’d name the word that’s on your homepage, the pattern and the word agree, and the wiring is being built. If they’d name a different word, or shrug and say they can’t tell, then the word and the money are telling two different stories, and the buyer is reading the money, in the only form the buyer ever sees it, the things it bought. The buyer always ends up reading the money. The homepage was never the thing.

The room is yours when the last twelve months of decisions could only have come from a company that meant the word. And then the next twelve months have to do it again, because the wiring you stop firing stops being wiring. A campaign won’t fix that. It’s in the spending, and it shows.

One last thing, because the argument cuts at me too. This essay is a claim. By its own logic, a claim is cheap, and cheaper than the thing it describes, and I’d be a fraud not to say so. The essay can’t give me the room. What it is, at most, is the visible surface of something more expensive underneath, which is years of reading companies one set of decisions at a time, getting some of them wrong, and keeping the ones the evidence holds. If the reading underneath is real, this surfaces it. If it isn’t, no sentence in here saves it. Which is the whole point, turned back on the person making it. You’ll know what I’ve been building by what I keep doing after you’ve stopped reading.