Essay · Market Structure
Start with a number that does not work. A competent Chief Information Security Officer costs north of five hundred thousand dollars a year once you count salary, equity, and the fact that good ones get recruited away the week they start. Now hold that number against a manufacturer doing eighty million in revenue, or a regional healthcare group, or a logistics firm with four hundred trucks. That company does not need a fifth of a CISO. It needs roughly a tenth of one, spread unevenly across the year, with the entire tenth arriving in a single week when something goes wrong.
There is no way to buy that. You cannot hire a tenth of a person, and the person you can hire at a tenth of the price is not the person you need. So the company does without, and the market records that as a customer who could not afford the service. It is not. It is a customer nobody built a product for.
Priced for someone ten times larger
This shape repeats across industries that have nothing to do with each other. Regulatory affairs. Clinical affairs. Financial planning and analysis. Data engineering. Anywhere the expertise is genuinely scarce, the incumbent providers organize around the largest buyers, and the pricing that results is not greed. It is arithmetic. A consulting firm's engagement model, bench economics, and partner leverage ratios all assume a Fortune 1000 buyer with a Fortune 1000 budget cycle. Given that cost structure, serving downmarket is a losing move, and the firms that decline to do it are correct.
But a correct decision, repeated by every provider in a category, produces a market failure that nobody chose. An entire tier of the economy goes unserved, and it is not a small tier. The middle market is most of the companies in the country and most of the jobs.
A need priced for a customer ten times larger is not a dead end. It is an opening that looks like a dead end from where the incumbents are standing.
The reason the opening stays open is that it is invisible from inside. From the incumbent's chair, that customer appears as someone who cannot pay. The information required to see it differently is not secret. It is just uninteresting to anyone whose economics already work.
The channel is already there
The second thing worth knowing about the unserved middle is that it is not unserved in general. It is unserved in the specific.
Mid-market companies rarely run technology the way large companies do. They outsource it, to a managed service provider or a regional integrator or a specialist firm that already has the relationship, the contract, and the CEO's mobile number. Whoever that is, they are already inside the account.
The instinct of a new entrant is to go around them. Call the CEO, explain that the incumbent provider is not qualified in this particular domain, take the account. It is the obvious move and it is almost always the wrong one, because it converts the one party with distribution into the one party motivated to stop you.
The alternative is older than most people building today. Whoever is already inside the account is either your distribution or your competitor, and you choose which. Going direct means winning accounts one at a time against an existing relationship. Going through means a single conversation buys access to an entire book of business, and it gives the incumbent something they genuinely need, which is the ability to answer a question in your domain without admitting they cannot.
This is not a clever idea. It is a thirty-year-old idea from enterprise software, where channel economics were worked out at scale and then largely stayed there. Most of what looks like insight when you enter a new industry is not insight. It is inventory.
What changes the arithmetic
All of the above has been true for decades. What is changing now is the cost of delivering the tenth of a person.
The reason expert services could not scale down was that expertise arrived bundled with a human being, and human beings do not divide. The unit of supply was a full-time salary and the unit of demand was a fraction of one. Every attempt to bridge that gap, fractional executives, retainers, shared services, was a scheduling workaround rather than an economic solution.
Applied intelligence changes the bundle. Not by replacing the expert, which is the version of this claim that keeps failing in practice, but by changing what the expert has to personally touch. When the routine ninety percent is handled by a system and the expert arrives only for the judgment, the unit of supply finally divides. A tenth of a person becomes something you can actually sell, at a price the middle market can actually pay, without the provider losing money on every account.
The constraint was never demand. It was that expertise arrived bundled with a human being, and human beings do not divide.
That is the opportunity worth watching, and it is a structural one rather than a technological one. The interesting question in any expert-services category right now is not whether a model can do the work. It is what the smallest sellable unit becomes when the routine part is absorbed, and whether the existing providers can reprice to meet it before somebody arrives who was built for that price from the start.
The value you cannot see
One more obstacle sits in the way, and it is the one most often underestimated.
Some categories deliver value that is invisible by design. Security is the clearest case: the entire return is an event that does not happen. Compliance, quality systems, and most of risk management share the shape. Senior managers file these under insurance, an unwanted expense defensible only in retrospect, competing every quarter against equipment that visibly makes money. That is not foolishness. It is a rational read of an expense whose payoff is a non-event.
You cannot sell a non-event. The industry's reflex is to reach for fear, which sells one contract and poisons the relationship, because a customer who bought out of fear resents you the whole time nothing goes wrong, which is to say the whole time you are succeeding.
The better move is to stop selling the disaster and sell the thing sitting next to it. Not: you will not be breached. Instead: you will answer the security questionnaire your largest customer just sent. You will pass the insurer's audit and stop paying the uninsured-risk premium. You will not lose the bid because a procurement officer asked something you could not answer. Those are real, countable, and happening on an ordinary Tuesday. If your value only becomes visible on the worst day of your customer's year, find the Tuesday where it also shows up, or expect to spend your life explaining why the thing that did not happen was worth the money.
The test
Three questions, and they apply to any expert-services category, not just this one.
Is there a real need here that is priced for a customer ten times larger than the one who has it? Who is already inside that account, and have you decided whether they are your distribution or your competitor? And if your value only shows up on the worst day, what is the ordinary Tuesday you can point to instead?
If the first answer is yes and you have good answers to the second and third, you are not looking at a market too small to serve. You are looking at one nobody has built for yet, which is a very different thing, and a considerably better one.
Juan Vegarra is the author of An Outsider's Playbook (forthcoming). The views here are his own. More essays · Write me