THE MACHINERY OF DISTRIBUTION THAT SHOWS ITSELF

Why the best product does not win, and what is actually being competed for

Two products. The same problem, solved about equally well.

The first is better. Genuinely better, in the way a person who has used both can demonstrate in ten minutes. It is faster, it breaks less, and the people who use it will not shut up about it.

The second one wins. It takes the market, it takes the money, and in four years the first one is a thing that a few people mention wistfully.

The story everyone tells afterwards is that the second company out-executed, or out-raised, or got lucky, or had better timing. Those stories are comfortable and they are downstream of the real thing.

The real thing is that the second product could be found and the first one could not.

Not that it was marketed harder. That it was structurally reachable. It sat in the path of people who were already moving, and the first one sat somewhere off the path, requiring a person to already know it existed in order to go looking for it, which is a requirement that eliminates almost everybody.

Quality decides whether a product is kept. Distribution decides whether it is ever encountered. And a product that is never encountered has its quality tested by nobody.

PART ONE: THE PRODUCT IS NOT COMPETING WITH THE OTHER PRODUCT

Here is the correction that reorganizes the whole picture.

A product is not competing with the rival product. It is competing for a slot in a channel, and the channel has its own physics, and those physics have nothing to do with which product is better.

A slot in a search result. A slot in a feed. A slot on a shelf at eye level. A slot in the sentence a friend says to another friend. A slot in the list a person already looks at every morning.

Each of those slots is finite, is governed by rules that the channel wrote for its own benefit and not for yours, and is allocated by a mechanism that has no access whatsoever to the quality of the thing being allocated.

Which means a business can be excellent, and lose, without ever having been evaluated. Not beaten. Never seen.

Distribution Is Not Promotion

This is the most expensive confusion in the field, and it costs companies years.

Promotion is what you do to a channel that already exists. You buy the slot, you make the creative, you pay per unit of attention.

Distribution is a structural property of the thing itself: whether it can travel, whether it fits where the people already are, whether the act of using it produces the conditions for someone else to encounter it.

Promotion is rented and it stops the moment the payment stops. Distribution is owned, and it keeps running.

A business with promotion and no distribution has to buy every customer, forever, at a price set by an auction that other people are also bidding in, and that price only rises. A business with distribution built into the product has a channel that does not send an invoice.

Hotmail in 1996 had no marketing budget worth the name. It had one line of text, appended automatically to the bottom of every message a user sent: PS: I love you. Get your free email at Hotmail. Every single email sent by every user became an advertisement, delivered by a trusted person, to a person who had already demonstrated that they read email, arriving inside the exact context where the product would be used. The recipient did not have to be found, did not have to be persuaded to pay attention, and did not have to be paid for. Twelve million users in eighteen months, at a fraction of the cost of any conventional campaign. The product was not better than its competitors in any way that a reasonable person could defend. What it had was a distribution mechanism welded to the act of use itself. Using it produced exposure. That is not promotion. Promotion would have been buying banner ads for an email product. This was the product distributing itself as a byproduct of doing what it did.

PART TWO: THE FOUR WAYS ANYTHING TRAVELS

There are only four, and every distribution strategy that has ever worked is one of them or a combination.

Search

Someone is already looking. The demand exists, formed, before the business does anything. All the business has to do is be present at the moment of looking.

This is the cheapest form of distribution in the world and it is not free, because it requires that the thing being sought is a thing people know how to name. If nobody knows the category exists, nobody is searching for it, and search does not work at all.

Word of Mouth

One person tells another. This is the only channel that gets cheaper as it scales, and the only one that arrives pre-trusted, and it is not a marketing activity. It is a consequence of the product.

It requires two things and it fails without either. The product must be good enough that a person’s reputation survives recommending it. And it must produce a moment where recommending it is natural, which is a design property, not a communications one.

The Borrowed Audience

Someone has already gathered the people. A platform, a publication, a person with a following, a store with foot traffic, a partner with a customer list.

The audience is rented, and the terms are set by whoever owns it, and those terms will get worse over time, because the owner of an audience always eventually charges what the audience is worth. Every business built entirely on a borrowed audience is a tenant, and every tenant is one policy change from having no business.

The Product Itself

The distribution is inside the thing. Using it exposes other people to it: a file that has to be shared to be useful, an invitation that must be sent, a link at the bottom of an email, a document that arrives with the sender’s tool visible on it.

This is the strongest form there is, because it does not decay and it cannot be shut off by a platform. It also cannot be added afterwards. It has to be designed into the mechanism of use, which means the decision that determines a company’s distribution is usually made by a product person, years before anyone in marketing is consulted.

PART THREE: WHY THE BEST PRODUCT LOSES

Now the mechanism, stated exactly.

Quality operates on people who have already arrived. It determines whether they stay, whether they pay, whether they return, and whether they tell anyone. Every one of those is conditional on arrival.

Distribution operates on arrival itself. It determines the size of the pool that quality is even allowed to act on.

They multiply. They do not add.

A product that is twice as good, reaching a tenth as many people, loses by a factor of five, and it loses without a single head-to-head comparison ever being made. The market never held a contest. The market simply never met one of the contestants.

And here is the trap inside the trap, and it is the reason good teams stay stuck for years.

The feedback a strong product gets from its small audience is excellent. The people who found it love it. They say so, loudly, and their enthusiasm is completely genuine, and it is the most misleading signal in business, because it is generated by exactly the people who are least representative of the market: the ones who found the thing despite it having no path.

That love is real. It is also a report from a survivorship-selected sample, and it is telling the team that everything is working, at the precise moment when the only thing that is not working is the one thing they are not looking at.

Betamax was, on most technical measures that mattered to engineers, the better format. Better picture, better build, and it came first. It lost to VHS, completely, and the reason had almost nothing to do with tape. Sony kept the format close, licensing it narrowly and controlling it. JVC gave VHS away to anyone who would build a machine. So within a few years there were many more VHS machines being manufactured by many more companies at many more price points, which meant more VHS machines in more homes, which meant the video rental store had to stock VHS tapes, which meant a person choosing a machine had to choose the one that could play what the store carried. Each step of that made the next step more certain. The tape did not get better. The path got wider, and the path was the whole game, and the format with the better picture was never actually evaluated by most of the people who decided against it. They did not reject Betamax. They rejected the empty shelf.

PART FOUR: THE CHANNEL DECIDES WHAT YOU BUILD

There is a deeper consequence, and it is uncomfortable, and it is true.

The channel does not merely deliver the product. It shapes it.

A product distributed by search must be a thing people already have a name for. A product distributed by word of mouth must contain a moment worth telling someone about. A product distributed on a platform must satisfy that platform’s algorithm, which is optimizing for the platform’s revenue and not for your customer’s outcome. A product that distributes itself must have sharing sitting on the critical path of its own use.

These are not marketing constraints applied to a finished product. They are design constraints on what the product can be at all.

Which is why the sequence that everyone runs is backwards. Build the thing, then figure out how to get it to people. That sequence produces, reliably, a beautiful object with no path to anyone, and by the time this is discovered the object is too finished to be reshaped around a path.

The order that works is the uncomfortable one. Name the path first. Then build the thing that can travel down it.

PART FIVE: WHAT CHANGES IN THE PERSON WHO SEES IT

Every product a person encounters starts to arrive with its path visible.

Not the ad. The path. How did this get in front of me, who paid for that, and what does the thing that put it here get out of it. That question, running automatically, is a different way of seeing a market, and it does not switch off.

And then something inverts, and it is the whole point.

The question stops being how do we tell people about this, which is a question about promotion, and becomes what would have to be true about this thing for it to travel on its own, which is a question about design.

The first question can be asked forever, and it costs money every time it is answered.

The second one can only be asked while the thing is still soft enough to change, and it is nearly always asked too late, if it is asked at all.

SYNTHESIS

WHAT IS ACTUALLY BEING COMPETED FOR

  not:  our product  vs  their product
  but:  a slot in a channel, allocated by a mechanism that
        cannot see quality at all


  QUALITY   ──►  acts only on the people who ARRIVED
  DISTRIBUTION ──► decides how many arrive

  they MULTIPLY.

  2x better × 0.1x reach = lose by 5, with no contest ever held


THE FOUR WAYS ANYTHING TRAVELS

   SEARCH            they are already looking     ── rented, cheap, needs a name
   BORROWED AUDIENCE someone gathered them        ── rented, and the rent goes up
   WORD OF MOUTH     one person tells another     ── owned, gets cheaper, pre-trusted
   THE PRODUCT       using it exposes others      ── owned, cannot be switched off
                                                     and cannot be added later


   THE TRAP:  the few who found you love you, loudly, honestly.
              they are the survivors of having no path.
              their love is the signal that hides the problem.

A product does not compete with the rival product. It competes for a slot in a channel, and the channel allocates that slot by rules that have no access to quality at all. So a business can be excellent, and lose, having never been evaluated by anyone.

Quality decides whether the people who arrive stay. Distribution decides how many arrive. The two multiply, which is why twice as good, reaching a tenth as many, loses by five without a single comparison being made.

Promotion is rented and stops when the payments stop. Distribution is structural: whether the thing can travel, whether it sits where people already are, whether using it exposes someone else to it. Only two of the four channels can be owned, and the strongest of them has to be designed into the mechanism of use, which means it is decided by a product choice made years before anyone asks a marketer for a plan.

And the channel does not just carry the product. It shapes what the product is allowed to be. Which makes the universal sequence, build it and then figure out how to reach people, an almost perfect way to produce something excellent that no one will ever meet.

The path comes first. Then the thing that can travel down it.

CITATIONS

Cusumano, M.A., Mylonadis, Y., & Rosenbloom, R.S. (1992). Strategic maneuvering and mass-market dynamics: the triumph of VHS over Beta. Business History Review, 66(1), 51-94. The format war was decided by licensing breadth, manufacturing volume and rental availability, not by picture quality.

Arthur, W.B. (1989). Competing technologies, increasing returns, and lock-in by historical small events. Economic Journal, 99(394), 116-131. Where adoption feeds adoption, the outcome is decided by early path advantages rather than by which technology is superior.

Ellsworth, R., & Hotmail growth data as reported in Draper, T. (1997). Viral marketing. Netscape M-Files. The original account of appending the offer to every outbound message, and the resulting adoption curve, at negligible marketing cost.

Berger, J., & Milkman, K.L. (2012). What makes online content viral? Journal of Marketing Research, 49(2), 192-205. Transmission is a property of the content’s structure, not of its quality: what gets passed on is what produces a specific state in the person passing it.

Christensen, C.M. (1997). The Innovator’s Dilemma. Harvard Business School Press. Products with inferior performance on the incumbent’s own measures routinely win, because they reach people the incumbent’s channel cannot serve.

THE MACHINERY OF THE EFFICIENT CHANNEL. Once the paths exist, the question becomes which one returns more than it costs, and why the blended average was hiding the answer.
THE MACHINERY OF NEW VERSUS RETURNING. Word of mouth is the channel that the returning half of the business produces for free, which is why the two subjects are the same subject.
THE MACHINERY OF THE COST OF A CUSTOMER. What it costs to buy a person when the product cannot carry itself.
THE MACHINERY OF UPSTREAM LEVERAGE. The general law underneath the whole of it. The path is decided at the origin, and it is nearly free there and nearly impossible later.