Art, Ops, and the Thing in Between

Two traditions have a claim on go-to-market, and the work that matters most falls in the space between them.

I have sat in a particular kind of meeting enough times to recognize it from the first few minutes. Two people are in the room, both responsible, in some sense, for whether the company grows. One of them came up through brand and creative and story. They talk about how the product makes people feel, about the line that will lodge in someone’s memory, about the difference between a campaign that is correct and a campaign that is alive. The other came up through operations and systems and data. They talk about pipeline coverage, about the conversion rate between the third and fourth stages, about which sequence variant won and by how much. Both of them are intelligent and both are good at their work, and they are not really able to talk to each other. Each suspects, quietly, that the other is doing something slightly unserious. The brand person thinks the ops person is counting things that do not matter. The ops person thinks the brand person is decorating something that cannot be measured. They are both partly right, and the company is paying for the gap between them without anyone naming it.

What makes the standoff worth noticing is that it is not a personality clash, though it always looks like one. It is two real bodies of knowledge that have grown up far enough apart that they no longer share a language, each able to see part of the problem clearly and neither able to see the whole. The company needs both of the things in that room, and it needs a third thing that lives in neither of them, and the third thing is what nobody hired for, because there is not yet a settled name for the person who would carry it.

That gap is what I want to look at, because it is the most interesting territory in go-to-market and almost nobody lives there on purpose. The two traditions in that room have each developed real knowledge over a long time. Neither of them, on its own, is engineering, and the reason neither is engineering tells you something about what an engineering discipline of go-to-market would actually have to be.

What the art tradition knows

Start with the brand and creative tradition, because it is the older of the two and the one most often dismissed by people who like numbers. It is worth taking seriously, because it knows things the other tradition cannot see.

The art tradition understands that markets are made of people, and that people do not move for reasons that fit cleanly in a spreadsheet. It understands attention, which is the scarcest resource in any market and the one least responsive to brute force. It understands that a brand is a compression of meaning, a way of carrying a complicated promise in a single recognizable form, and that this compression is worth more than any individual message. It understands taste, which is a kind of judgment rather than decoration, the ability to tell the difference between something that will resonate and something that will slide off. People who are good at this can look at two versions of a thing and know, before any data exists, which one will work, and they are right often enough that the skill is clearly real.

You can watch this happen and still not learn it. I have sat with someone who had this kind of judgment while they rejected a piece of work the rest of the room considered finished, named the one thing wrong with it in a single sentence, and then, asked how they knew, could not say. The knowing was real, and it was faster and more accurate than the committee around them, and it lived somewhere below the level where reasons are kept. That is the gift and the limit of the art tradition at the same time. The judgment is real and the judgment is trapped, available to the person who holds it and almost impossible to extract, so a company that depends on it is really depending on a small number of people it has to keep and cannot easily replace. The same is true of the deeper thing the tradition understands, which is that a brand is a compression of meaning, a way of carrying a complicated promise in a form a person can recognize in an instant. Building that compression well is among the most valuable work in any market, and the people who can do it are working from a sense of rightness they cannot fully hand to anyone else.

What the art tradition cannot do is guarantee anything. Its knowledge lives in individuals and travels poorly. You cannot take the instinct of a gifted brand builder and install it in someone else, or write it down in a form that survives the person leaving. You cannot say in advance how much of an outcome a given creative decision will produce, or how the same decision will behave in a different market, because the knowledge is intuitive rather than explicit, and intuition does not come with error bars. When a campaign built this way succeeds, the explanation offered afterward is a story, and the story is usually persuasive and usually unfalsifiable. When it fails, the explanation is also a story. The tradition produces brilliant outcomes and cannot tell you reliably which of its decisions caused them, which means it cannot be taught except by apprenticeship and cannot be trusted to repeat.

Apprenticeship, in turn, does not scale. It moves knowledge one person at a time, slowly, with heavy losses along the way, and it leaves the field permanently short of the talent it runs on. A discipline that could write down even part of what the gifted brand builder knows would change the economics of the entire activity, because it would let competence spread faster than rare people can be found and trained. The art tradition, for all its real depth, has never produced that writing-down, and it is not obvious it can from the inside, because the very knowledge that makes someone good at the work is the knowledge that most resists being made explicit.

This is the signature of an art. An art is a body of practice in which excellence is real, recognizable, and concentrated in talented individuals, and in which the knowledge cannot be fully separated from the practitioner. Cooking at the highest level is an art in this sense, and so is diagnosis at the edge of medicine, and so is a great deal of brand work. The art tradition in go-to-market is not unserious. It is simply pre-explicit, and a field built only on it will always depend on finding the right rare people and will never be able to say, in general, why the thing works.

What the ops tradition knows

The operations and systems tradition grew up partly in reaction to all of that. Its founding intuition is that the vagueness of the art tradition is a problem to be solved, and that the way to solve it is to measure, to instrument, and to systematize. This tradition has also developed real knowledge, and it is the knowledge the current moment rewards most heavily.

The ops tradition understands process. It knows how to take something that happened once by luck and make it happen repeatedly on purpose. It understands instrumentation, the discipline of putting a number on each step of a flow so that you can see where people fall out and where they continue. It understands the funnel as a structure, the idea that demand moves through stages and that each stage has a rate, and that the overall outcome is the product of those rates. It understands automation, how to take a manual sequence of actions and make it run without a person, which is the skill the whole go-to-market tooling economy is now built around. People who are good at this can take a chaotic, intuition-driven growth effort and turn it into something legible, with dashboards and stages and conversion rates, and the legibility is valuable in itself, because you cannot manage what you cannot see.

The power of this is real, and so is its characteristic failure. I have watched a team lift the conversion rate of the stage they were measuring by making the step just before it harder, which filtered out the people who were never going to buy and made the surviving rate look wonderful, while the total number of customers quietly fell, because the same filter also turned away people who would have bought if the path had been easier. Every number on the dashboard moved in the right direction. The business got worse. Nothing in the ops toolkit flagged it, because the toolkit measures the stages it has defined, and the harm happened in the space between the stages, in a part of the system no one had instrumented, because no one had a model that said that part mattered. This is the recurring shape of the ops tradition’s blind spot. It sees the system as the set of things it has chosen to measure, and it is confident in proportion to how much it has measured, and the confidence is most dangerous exactly where the measurement stops.

What the ops tradition cannot do is explain. It can tell you, with precision, what happened. It can tell you that the conversion rate between two stages is eleven percent, and that it was nine percent last quarter, and that a particular change preceded the rise. What it usually cannot tell you is why, in any sense deep enough to predict what will happen next. The funnel it measures is a description of the past, not a model of the system, and the difference matters enormously the moment conditions change. A funnel that has only ever been observed in one regime tells you almost nothing about how it behaves in another, and the ops tradition, because it is built on measurement rather than on a theory of the underlying object, tends to mistake a detailed record of what occurred for an understanding of why it occurred. It optimizes locally, improving the number in front of it, and it is often blind to the larger structure that the number sits inside, the structure that determines whether the local improvement helps or quietly does harm somewhere it is not looking.

This is the signature of operations, as distinct from engineering. Operations runs and improves a system that already exists; it is the discipline of keeping the machine going and making it incrementally better. That is essential work, and it is not the same as understanding the machine well enough to design it, or to predict how it will behave under a load it has never carried. A field built only on operations will be very good at measuring itself and very poor at explaining itself, which is a precarious place to be when the conditions stop holding still.

The curve every field travels

There is a pattern to how human practices turn into disciplines, and go-to-market is sitting at a recognizable point on it.

A practice usually begins as an art, dependent on gifted individuals whose knowledge cannot be transferred. Over time, as the practice is repeated and observed, it accumulates a body of craft, a set of rules of thumb and standard methods that can be taught to ordinary practitioners and that make the typical outcome reliable even without genius. Craft is a real advance over art, because it spreads competence beyond the talented few. Then, sometimes, a practice makes a further transition, when someone develops an explicit theory of the object the craft has been shaping by feel, a model precise enough to design with and to predict from, and at that point the craft becomes an engineering discipline. Building went this way, from master builders whose knowledge died with them, to the accumulated craft of the guilds, to structural engineering with its calculations. Medicine has been making the same passage for a century, from an art practiced by individuals, through the craft of accumulated clinical method, toward something that can increasingly say why an intervention works and predict who it will help.

The craft stage is worth pausing on, because it is where go-to-market mostly lives, and it explains the shape of the field’s writing. Craft is the realm of the playbook, the named motion, the best practice, and craft knowledge is a real advance on pure art, because an ordinary practitioner can pick up a playbook and get a decent result without the rare instinct. The trouble with go-to-market craft is that it is unusually shallow and unusually unstable, because the playbooks are tied to particular tools and particular channels that change every few years, so the craft has to be relearned constantly and never settles into anything deeper. A building code accumulates for a century and the buildings get safer. A go-to-market playbook is half obsolete by the time the channel it depended on saturates, which is why the field rewrites its tactics every few years and tends to mistake the churn for progress. Craft that cannot compound stays craft, and a field whose craft cannot compound never lays down the foundation a discipline would stand on.

Go-to-market has not made the second transition, and it has barely made the first. It still depends heavily on gifted individuals, which is the mark of an art. It has accumulated some craft, the playbooks and the standard motions, though the craft is shallow and changes with the tools. It has no explicit theory of its object worth the name, which is why it cannot predict and why its outcomes vary so wildly, and why the explanations it offers for success and failure are almost always stories told after the fact. The two traditions in that meeting room are an art and an operation sitting side by side, each compensating a little for the other’s blind spots, and the obvious question is why putting them together does not simply produce the discipline, which is the question worth answering next.

Why neither is engineering

It is tempting to think the answer is simply to combine the two, to put the brand person and the ops person on the same team and let taste and measurement balance each other. That helps, and most good go-to-market efforts already do some version of it. It is not the same as having a discipline, because a discipline is more than the sum of an art and an operation. It is a third thing, with its own kind of knowledge, and the third thing is what is missing.

The reason the synthesis is more than the sum is that taste and measurement, set side by side, still leave the central question unanswered. Taste tells you what might resonate. Measurement tells you what already happened. Neither of them tells you how the system works, in the sense of a mechanism you could describe to another person and use to predict an outcome you have not yet observed. You can put a brilliant brand mind and a rigorous operator in the same room for a year, and unless one of them quietly builds an explicit model of the thing they are both touching, what you get at the end is a well-decorated, well-measured system that no one actually understands. Here is the reason, stated as plainly as I can. Measurement can only describe states the system has already entered, and taste can only register states a person has already felt, so neither method, however refined, ever yields a claim about a state the system has not yet reached. A claim of that kind, a prediction about behavior not yet observed, is a different sort of object, and it cannot be produced by adding a record of the past to a feeling about the present. It has to be constructed from a model of how the system works, and a model is precisely what neither tradition’s method generates. Understanding is a separate achievement, and it does not fall out of stapling intuition to telemetry. It has to be built on purpose, and building it is the work the field has not done.

Engineering, in the sense worth holding the field to, is principled design that produces predictable outcomes under stated conditions. The key words are design and predictable and conditions. An engineer does not only have taste, though good engineers often do. An engineer does not only measure what happened, though measurement is essential to the work. An engineer has a model of the object being built, an account of how it behaves that is explicit enough to reason with before the thing exists, and that model is what lets them predict, within a stated range, what the object will do when it meets the world. The structural engineer can tell you how the beam will behave under a load it has never carried, because they have a model of the beam, not merely a record of beams that stood and beams that fell. That capacity, to say in advance and within bounds what a thing will do under conditions you have not yet seen, is the thing both the art tradition and the ops tradition lack, and it is the center of what engineering is.

It is worth dwelling on what a model buys you, because it is easy to mistake for a fancier kind of measurement. A structural engineer working from a model can take a beam that has never been built, in a material and a configuration no one has tested at that size, and state how much it will carry before it bends and how much before it breaks, with a margin they can defend to another engineer. They are not recalling a beam like it. They are reasoning from an explicit account of how beams of that kind behave, an account compressed into relationships they can calculate with. That is why engineering can build things that have never existed and have them stand the first time, which neither intuition nor observation can manage on its own, because intuition has not seen the new thing and observation has no record of it. A go-to-market effort that wanted the same power would need an account of its own object that explicit, and the plain position today is that we do not have one, which is why entering a truly new market still feels like a leap of faith rather than a calculation with a margin.

The art tradition cannot do this because its knowledge is intuitive and does not yield predictions with bounds. The ops tradition cannot do this because its knowledge is observational and does not extend past the conditions it has already seen. The space between them, where someone has an explicit model of the go-to-market system, rich enough to design with and to predict from, is the space an engineering discipline would occupy, and at the moment it is mostly empty. People wander into it occasionally, the rare operator who has both taste and rigor and has privately built something like a working theory of how their particular market behaves, and when you meet one you notice immediately, because they can do something the others cannot. They can tell you what will happen if they change something, and then change it, and be right, repeatedly, for reasons they can articulate. I have met perhaps a handful of people like this, and what stays with you afterward is not their intuition, which the artists also have, and not their dashboards, which the operators also have. It is that they treat the market as a thing with describable behavior, something that responds in patterns they have worked out and can state out loud. Ask them why a channel stopped working and they do not reach for a story, they reach for a mechanism. That is the texture of the missing discipline, visible in the few who have built a private version of it. There are very few of them, and almost nothing they know has been written down in a way that anyone else could learn.

The cost of the gap

The gap between the two traditions is not only an intellectual curiosity, and the cost of it shows up in ways anyone who has worked inside a growing company will recognize. Companies hire for one tradition or the other and then swing between them on a cycle, bringing in a brand leader to restore some soul after a season of joyless optimization, then an operator to restore some rigor after a season of unaccountable spending, never quite landing on the thing that would actually help, which is a way of thinking that contains both and collapses into neither. Much of the variance in outcomes that everyone treats as simply the nature of the business is this, the absence of a discipline that could make growth both alive and accountable, so that whether a given effort works depends far too much on whether the company happened to find one of the rare people who carry the whole thing in their head. When you depend on rare people, you are not running a discipline. You are running a search for talent and hoping it pays out, which is a fine description of an art and a poor description of a field that moves serious money.

The thing in between

So the interesting question is not which tradition is right, because both of them know real things and both of them are missing the same thing. The interesting question is what would have to exist in the space between them. What would a person need to know to stand in that gap on purpose, with the art tradition’s understanding that markets are made of people and the ops tradition’s insistence on measurement, and something further that neither has, an explicit model of the system precise enough to design with.

I do not think that person exists yet in any numbers, and I do not think the knowledge they would need has been assembled. The few who approach it have built it privately, for one market, and cannot fully say what they know.

It is a strange vacancy, when you consider how much money depends on the work. Almost every other activity that moves this much capital has, at some point, grown a class of people who specialize in understanding it rather than only practicing it, the people who can say why and not only how. Go-to-market has produced an enormous amount of practice and very little of that understanding, and the few who hold some of it tend to keep it tacit, partly because the incentives reward doing over explaining and partly because no shared language exists to explain it in. The space stays empty for a harder reason than size, because entering it means building the vocabulary and the models from something close to nothing. The space between art and operations is where the discipline would live, and the fact that it is mostly empty is not a sign that nothing belongs there. It is a sign that the work of building it has not been done. The first piece of that work is the one the two traditions in the room never quite get to, because the art person is talking about people and the ops person is talking about rates, and neither is talking about the thing itself. Before you can model a system, or measure it well, or even tell a true story about it, you have to be able to say what it is, with enough precision that the saying could be checked, argued with, and shown to be wrong. Naming the object that go-to-market actually builds turns out to be much harder than it sounds, and much rarer than you would expect, and it is where the real difficulty begins.

---

References

• On the progression of practices from art to craft to engineering discipline, the history of structural engineering and of clinical medicine offers the clearest parallels; see Walter G. Vincenti, What Engineers Know and How They Know It (Johns Hopkins University Press, 1990), on engineering knowledge as distinct from both intuition and observation.