The Missing Role in the Mid-Market Revenue Stack

In The CRM Data-Debt Spiral I described how revenue stacks decay: one rational shortcut at a time, until reports disagree and every meeting starts with a debate about whose number is right. I ended by saying the spiral is broken by ownership, not by tools.

This essay is about why that ownership is missing in the first place — and why it's missing in almost exactly the same way at almost every mid-market company I've seen in 25+ years of working with revenue technology.

Everyone owns a piece. Nobody owns the picture.

Walk into a typical company between 50 and 500 employees and map who owns what in the revenue stack:

Marketing ops owns campaigns, automation, and attribution. Sales ops owns pipeline, process, and forecasting. Customer success owns their platform and playbooks. IT owns security, access, and infrastructure. Finance owns the tools budget, usually reluctantly, usually discovering new line items on the card statement.

Now ask a different question: who owns the data model? Who owns the integration map? Who decides what the source of truth is when two systems disagree? Who evaluates whether the stack can carry next year's AI plans?

In most of these companies, the honest answer is nobody. And it's worth being precise about why — because it's not what most people assume.

Not a competence problem

The tempting explanation is that nobody in the building can do this work. That's rarely true. Marketing ops people are often deeply technical. IT teams understand integration patterns. Plenty of sales ops leads could draw the data flows if you gave them a week.

The real reason is structural. The role falls between the chairs by design:

It's too technical for marketing. A marketing ops manager who spends her weeks deep in integration architecture is neglecting the campaigns she's measured on. Her incentives point at throughput, not foundations.

It's too commercial for IT. IT's mandate is security, uptime, and access. What the sales definition of "customer" should be is emphatically not their problem, and they will tell you so — correctly.

It's too strategic for an admin role. Companies that sense the gap often hire a CRM administrator. Good administrators are worth their weight in gold, but the role as usually scoped executes requests; it doesn't have the mandate to refuse them. And refusing requests — saying no to the shortcut in stage 1 — is most of the job.

So every function rationally optimises its own piece, every local decision makes sense, and the connecting layer decays exactly as the spiral predicts. Nobody is failing at their job. The job was never given to anyone.

Enterprises solved this. Mid-market inherited the problem without the solution.

Large enterprises hit this wall decades ago and built the machinery to deal with it: enterprise architects, platform teams, architecture review boards, data governance functions. The machinery is often slow and bureaucratic — but the ownership question, at least, has an answer.

That machinery assumes enterprise headcount. A dedicated enterprise architect plus a platform team is a seven-figure annual commitment. A 150-person company cannot justify it and never will.

Meanwhile, the mid-market company's stack is not meaningfully simpler than the enterprise's. Fifty to a hundred tools, a CRM at the centre, integrations in every direction, and now AI initiatives layering on top. The complexity arrived; the org design that manages it didn't. That's the squeeze: enterprise-grade architecture problems on a mid-market org chart.

For a long time this gap was survivable, because the cost of a messy foundation was diffuse — slower reporting, some duplicate work, the occasional broken sync. AI has changed the maths. A model reading your CRM inherits every undocumented workaround and every stale record, and produces confident conclusions from all of it. The messy foundation used to cost you efficiency. Now it costs you every AI initiative you attempt on top of it. The price of the missing role went up, sharply, and mostly invisibly.

What the role actually owns

Strip away the enterprise ceremony and the job is concrete. Whoever owns the revenue-stack architecture owns four things:

The map. A current picture of the data model and every integration, including the embarrassing workarounds. Not a governance document — a map people actually use. In my experience, producing this map for the first time changes the organisation's conversation more than any tool purchase, because for the first time everyone is arguing about the same picture.

The definitions. What is a customer, a lead, an active deal — written down, agreed across functions, and maintained. Boring beyond words. Also the difference between reports that get trusted and reports that get debated, and between AI that helps and AI that hallucinates plausibly.

The gate. A lightweight review for changes to the data model and new integrations. Not a committee — a person and a checklist. Most shortcuts are fine; the job is having someone with the whole picture in their head who knows which ones. This is the mechanism that stops stage 1 of the spiral from recurring.

The horizon. The connection between the stack and where the business is going. Can the current architecture carry the new product line, the new sales motion, the AI roadmap? Someone has to hold that question continuously, not just when the annual planning deck asks for it.

Notice the scale of this. It is not a platform team. At mid-market size it is one senior person's part-time attention — a day or two a week of genuinely architectural work, with a mandate attached. The mandate is the part companies get wrong: the map and the definitions without the authority to say no is just documentation of the decay.

Internal or external?

Should this be a hire? Sometimes. If you have a technically deep, commercially fluent senior person who wants the job and can be given real authority across marketing, sales, and IT territory — give it to them formally. Title it, mandate it, measure it. The worst version is the informal one, where the most knowledgeable person carries the map in their head as unpaid overhead until they resign and take the architecture with them.

But the honest maths is awkward for a full-time hire. A person senior enough to hold the mandate across functions is expensive, and the role at mid-market scale doesn't fill a week — which is why it usually gets bundled into another job and then neglected, restarting the cycle. This is the same structural reason fractional executive roles have grown across finance and marketing: some jobs need senior judgment at partial capacity, and the org chart has no natural slot for that.

I don't think it matters much whether the ownership is internal or external, permanent or fractional. What matters is that it exists, that it has a mandate, and that it is somebody's actual job rather than everybody's third priority.

The uncomfortable summary

The spiral from the previous essay is not caused by bad tools or bad people. It is caused by an org-design gap that mid-market companies inherited from a simpler era: a connecting layer that grew enterprise-grade complexity while remaining nobody's job.

The companies that close this gap in the next few years — by whatever staffing model — will compound a quiet advantage: trusted numbers, faster decisions, and AI initiatives built on foundations that can actually carry them. The companies that don't will keep buying tools, keep rebuilding platforms, and keep wondering why the new one feels like the old one within three years.

The platform was never the problem. The empty chair was.