The Distribution Artifact

Every engineering discipline is built around the thing it makes. Go-to-market has never named its own.

Try a small experiment. Ask five people who work in go-to-market what they build. Not what they do, what they build, the object their work produces. You will get five different answers, and the differences are the interesting part. One will say campaigns. One will say pipeline. One will say a brand. One will say a funnel, or a motion, or a growth engine, reaching for a word that sounds like a thing without quite being one. One will look at you as though the question is strange, because the honest answer, the one underneath the others, is that they are not sure there is a single object, that the work is a collection of activities rather than the construction of a thing.

I do not think this happens because the people are unclear thinkers. They are often very sharp about their parts, fluent and exact about the channel or the funnel or the brand they own. The vagueness appears only when you ask about the whole, and it appears reliably, across companies and across levels of seniority, which tells you it is a property of the field rather than a personal failing. The field has never handed its practitioners a whole to be clear about, and a person cannot be articulate about an object that their discipline has never named.

Now ask a structural engineer the same question and the strangeness disappears. They build structures. A bridge, a tower, a frame, but underneath the particular instances there is a clear and shared idea of the kind of object the discipline produces, a thing that carries loads across space without falling down. Ask a control engineer and you get control systems. Ask a chip designer and you get integrated circuits. In each case the answer is immediate and shared, because the discipline is organized around the object it makes, and everyone in it knows what that object is even when they are working on a small piece of one.

This difference is not a quirk of language. It is the difference between a discipline and a set of activities, and it is the thing I want to look at, because I have come to think that go-to-market’s failure to name its object is not a minor gap. It is close to the whole problem. You cannot have an engineering discipline without a clear account of the thing it engineers, and go-to-market has never settled what that thing is, which is why it has parts without a whole, tactics without a structure, and a great deal of motion that never quite adds up to an object anyone can point at, hand off, or be held responsible for.

What an artifact is, and why disciplines are built on them

Start with the bridge, because it makes the idea concrete. A bridge is not its girders. It is not its cables, its deck, its piers, its bolts. It is made of those things, and every one of them matters, and you could be an expert in any one of them, in the metallurgy of the cables or the geology of the foundations, without being the person responsible for the bridge. The bridge is the whole, the thing the parts add up to, and it has properties that none of its parts have on their own. It has a load it can carry, a property of how the components work together rather than the sum of their individual strengths. It has ways it can fail that no single part predicts, because the failure lives in the interaction. It has a behavior under wind and traffic and temperature that you can only understand by considering the structure as one object. The bridge is the artifact, and the discipline of structural engineering is organized around it.

The same is true of any engineered object once you look. A car is not its engine and its wheels and its brakes considered separately, it is the system they form, and its most important properties, how it behaves in an emergency stop on a wet road, belong to the whole and to the way the parts act on one another. Engineering failures of the worst kind live precisely in those interactions, in a place where every component meets its own specification and the system fails anyway, because the failure is a property of how the parts come together rather than of any part alone. This is the deep reason the artifact has to be the unit of attention. The most dangerous failures are the ones that no amount of care about the components would ever catch, because they do not live in the components, they live in the whole, and only someone responsible for the whole is positioned to see them coming.

This is what I mean by an artifact, the characteristic object a discipline exists to produce, considered as a whole rather than as a heap of parts. The artifact is the unit of design, the thing the engineer is actually designing when they design, even when the day’s work is a single joint. It is the unit of responsibility, the thing someone is answerable for, so that when a bridge falls there is a clear answer to the question of what failed and who was responsible for it as a whole. And it is the unit of knowledge, the thing around which the discipline’s understanding is organized, so that what structural engineers know is, in the end, knowledge about how structures behave.

Every mature engineering discipline has one of these, and the artifact is so basic to the discipline that the people inside it rarely think to mention it, the way a fish does not mention water. The aircraft. The engine. The circuit. The bridge. The artifact is the answer to what do you build, and the existence of a clear, shared answer is one of the surest signs that you are looking at a real discipline rather than a bundle of related skills. When the answer is immediate and everyone gives the same one, the discipline has an object. When the answer is five different words and a shrug, it does not.

The parts that go-to-market mistakes for the whole

Go-to-market is unusually rich in parts and unusually poor in wholes, which is part of why the object has stayed hidden. The field has names for all of its components, and it spends nearly all of its attention on them, which makes it easy to believe that the components are the thing.

Consider the candidates that get mistaken for the object. The campaign is a part, a single coordinated push, and a campaign is to the thing go-to-market builds roughly what a single load test is to a bridge, an event in the life of the object rather than the object itself. The channel is a part, a route through which attention or demand flows, and being expert in a channel is like being expert in cables, real and valuable and not the same as being responsible for the structure. The funnel is the most interesting false candidate, because it feels like the whole, but a funnel is a description, a way of measuring the object’s behavior by counting what passes through its stages, and a description of a thing is not the thing, in the same way that a record of how much traffic crossed a bridge is not the bridge. The brand is a part, an important one, a compression of meaning that the object uses, but a brand on its own creates nothing without the structure around it. The technology stack is a part, the set of tools the object is built with, and confusing the stack with the object is the error the whole field has been making, the one that treats assembling the tools as building the thing.

Two more candidates deserve mention, because they come closest and still fall short. The growth model, the spreadsheet that projects how customers will accumulate, feels like the object because it spans the whole, but a model in this sense is a forecast, a set of assumptions about the artifact’s behavior rather than the artifact itself, and a forecast of a thing is no more the thing than a weather forecast is the weather. The go-to-market motion, the phrase the field reaches for when it wants to sound structural, is a pattern of activity, a way of moving, and a motion is something the artifact does rather than something it is. Both come close because both gesture at the whole, and both stop short because neither names the object that is doing the growing or making the motion.

You can see the cost of the missing whole in a pattern that recurs in company after company. Each part is staffed by someone capable, the brand is sharp, the channels are well run, the funnel is instrumented, the tooling is modern, and the business still cannot reliably turn its market into customers, and no one can quite say why, because the failure is not in any part. It is in the way the parts do not compose, the brand promising one thing and the sales motion delivering another, the channels bringing in people the product was never built for, the funnel optimized for a conversion that does not last. Every component passes its own review. The artifact, which no one owns, quietly does not work, and the company concludes that go-to-market is simply hard, when what is actually true is that no one was ever responsible for the whole.

Each of these is real, and each of them is a component or a measurement or a material, and none of them is the whole. What is missing is the name for the thing they are all supposed to add up to, the system that takes a product and a market and produces durable demand, considered as a single designed object with its own behavior. Without that name, the field’s attention has nowhere to rest except on the parts, and so the parts get optimized endlessly while the whole goes unattended, because there is no whole in anyone’s mind to attend to. You cannot take responsibility for an object you have never named, and you cannot design one either.

Naming the object

So let me name it, plainly, and then defend the name. The object that go-to-market builds is the distribution artifact, the designed system that reliably creates and captures a market for a product. It is worth taking the definition apart word by word, because each word is load-bearing and each one rules out something the field currently confuses it with.

It is designed, or it ought to be, which means the artifact is something built on purpose to a specification rather than something you merely end up with. Most of what passes for go-to-market today is emergent in the other sense, the accidental sum of whatever campaigns and channels and tactics happened to get funded, and an emergent system of that kind is an artifact that no one designed, which is a weaker and more fragile thing than one built to a specification, because no one chose its structure and no one can reliably reproduce it. The word designed names the aspiration at the center of the whole idea, the difference between an artifact you intend and an artifact that merely happened to you. The object is present either way. Whether it was designed is the difference between engineering it and suffering it.

It is a system, which means it is a set of interacting parts whose behavior comes from the interaction, not a list of independent activities. This is the word that separates the artifact from the heap. A system has properties the parts do not, behaviors that appear only when the parts are connected, and to call go-to-market’s object a system is to claim that there is something to understand at the level of the whole, a structure with its own dynamics, which is exactly what the parts-focused view misses.

It is reliable, or it aims to be, which means the artifact is meant to produce its outcome repeatably and within known bounds, not once by luck. This is the word that makes it an engineering object rather than an event. A distribution artifact that worked once and cannot be made to work again is like a bridge that held until the day it did not, a failure dressed as a success. Reliability is the property the field most conspicuously lacks and most needs, and building it in is the whole point of designing the object rather than improvising it.

It creates and captures, both, which rules out the partial objects that handle only one side. A system that generates attention but never converts it is only half of the artifact, a generator of demand with no means of keeping what it creates. A system that converts demand it did not create is the other half, living off demand some other process produced. The distribution artifact is responsible for the whole arc, from a market that does not yet want the product to durable demand that does, and the word captures insists that creating the want is not enough, that the artifact is not finished until the want has become a relationship the business can keep.

And it produces a market, not a campaign result, which fixes the scale of the thing. The output of the distribution artifact is a market for the product, a population of customers in a sustained relationship with it, which is a far larger and more durable thing than the result of any single push. This is the word that keeps the artifact from collapsing back into the campaign, that insists the object is the thing that produces markets, not the thing that produces leads this quarter.

Why the name changes what you can do

Naming the object is not a semantic flourish, and I want to be concrete about what changes once the artifact has a name, because the changes are practical and large.

The first thing that changes is the unit of responsibility. In most companies today, no one is responsible for the distribution artifact, because it does not exist as a named thing to be responsible for. There is someone responsible for the campaigns, someone for the channels, someone for the brand, someone for the tooling, and the whole that these are supposed to compose is an orphan, attended to by no one because it has no name and therefore no owner. The predictable result is that every part can be doing well while the object does badly, each component optimized and the system as a whole incoherent, and no one is accountable for the incoherence because no one was ever made responsible for the whole. Naming the artifact creates the possibility of owning it, of saying that this person is responsible for the object itself, for whether the system as a whole creates and captures a market reliably, in place of responsibility for any single channel or campaign. That is a different job from any that currently exists, and it cannot exist until the object does.

It is worth picturing what owning the artifact would actually look like, because it is a role that barely exists today. The owner of the distribution artifact would not run the campaigns or the channels, in the way that the engineer responsible for a bridge does not personally pour the concrete. They would be responsible for the object as a whole, for its design and its coherence, for whether the parts compose into a system that creates and captures a market reliably, for knowing its capacity and its failure modes and the conditions it is rated for. They would be answerable when the whole failed even though every part had passed its own review, because catching exactly that kind of failure would be the substance of the job. No common title maps cleanly onto this. The closest existing roles own parts or own budgets, and the artifact, the thing itself, stays unowned in most organizations, which is the clearest possible sign that it stays unnamed.

The second thing that changes is the level at which you reason. Once you are thinking about the artifact as a whole, certain questions become askable that are invisible at the level of the parts. You can ask what the object’s capacity is, how much demand it is built to carry before it strains, which is a property of the whole system and not of any component. You can ask how it fails, what the failure modes of the structure are, which live in the interactions and not in the parts. You can ask what it is rated for, the conditions under which it performs as designed and the conditions outside which it does not. None of these questions can even be posed at the level of campaigns and channels, because they are properties of the artifact, and the artifact has to exist as an object of thought before its system-level properties can be reasoned about at all. Naming it is what makes the whole-system questions thinkable.

A concrete version of this makes the point. A company notices its growth slowing and, lacking an object to reason about, treats the slowdown as a collection of part-problems, a channel that is underperforming, a message that needs refreshing, a conversion rate that has dipped. It works each part and the slowing continues, because the real cause is a property of the whole. The artifact has reached the capacity of the market it was built to serve and is now straining against a limit that no individual part reveals. Only someone reasoning at the level of the artifact could even name that as the cause, because saturation is a system property, invisible in the parts and plain in the whole. The company without an object for its thoughts will keep adjusting components and wondering why the adjustments have stopped working.

The third thing that changes is what knowledge accumulates around. A discipline’s knowledge organizes around its artifact, and as long as go-to-market has no artifact, its knowledge cannot organize, which is why it stays a pile of disconnected tactics that expire with their channels. Knowledge about the distribution artifact, by contrast, would be knowledge about a persistent object, the kind that can accumulate across people and across years, because the object outlasts any particular campaign or tool. The artifact is the thing around which a real body of go-to-market knowledge could finally cohere, and the absence of the artifact is part of why no such body has ever cohered.

You can see the difference in how knowledge travels in the two kinds of field. In a discipline with an artifact, an engineer who has spent years learning how structures behave can move to a new firm, a new material, a new kind of structure, and carry almost all of that knowledge with them, because it is knowledge about the object and the object persists. In go-to-market, a practitioner who has mastered a channel watches much of that mastery expire when the channel changes, because the knowledge was attached to a part rather than to a persistent whole. Knowledge organized around the artifact would survive the churn of tools and channels in a way that the field’s present knowledge never does, and that durability is one of the quiet rewards of finally having an object to know.

The artifact is already there

The strangest part is that the distribution artifact already exists in every company, unnamed and therefore undesigned. Every business that sells anything has a system, of some kind, that creates and captures its market, even if that system was never designed and no one is responsible for it as a whole. The artifact is present, doing its work or failing to, exerting all the behaviors of a system, straining under loads no one measured and failing in modes no one anticipated, and the only thing missing is that no one has named it, looked at it as a whole, or taken it on as an object to be engineered.

This is why naming it matters even though the thing is not new. Naming it does not create the artifact, which was there all along. Naming it makes the artifact visible, turns it from an unexamined byproduct of scattered activity into an object that can be seen, designed, owned, and improved. The bridge was load-bearing before anyone could calculate a load, and the calculation did not create the bridge, it created the possibility of building bridges on purpose instead of by feel. Naming the distribution artifact is the same kind of move, the one that turns a thing that happens to you into a thing you make.

Once the object is named, the questions that follow are the ones every engineering discipline asks about its artifact, and they turn out to be surprisingly hard. The first of them is the one that sounds simplest and is not. If the distribution artifact is a system that can be designed and rated and pushed to capacity, then it must be a thing that can be measured, and measured means measured in some unit, and the moment you ask what the unit is, what the actual measurable quantity of the thing this artifact produces could be, you find that go-to-market has been counting everything and defining nothing, and that the question of what to measure is a foundation that was never poured rather than a detail to be settled later.

---

References

• On the artifact as the organizing object of an engineering discipline, and on system-level properties that do not reduce to component properties, see Walter G. Vincenti, What Engineers Know and How They Know It (Johns Hopkins University Press, 1990), and the broader literature on the philosophy of technical artifacts.