Marketing automation for mid-market is dead past the first 40k

You signed the enterprise contract for one feature. You are paying for forty. That gap is the leak, and for a sub-30M business it runs 25k to 60k a year in license you never touch. Run the math before you renew, because the platform that made sense at your first ten seats stops making sense long before you hit your revenue ceiling.

This is not an anti-software rant. The CRM of record earns its keep. What does not earn its keep is the bolted-on automation suite you bought with it - the seats, the modules, the marketing-hub tier that sits at 9% utilization while the invoice clears every month.

The reason this happens is structural. Enterprise platforms sell the whole catalog as a bundle, scoped against where you might be in three years. You buy the future-state org chart, then run the present-day team on it. The contract renews on autopilot, the champion who signed it moves on, and nobody is left to question the line items.

The leak: enterprise pricing, hobbyist usage

Here is the pattern we see across audits. A 12M operator buys a top-tier platform. The CRM gets used daily. The automation suite ships 5 workflows: lead routing, a welcome sequence, a re-engagement drip, deal-stage alerts, and a renewal reminder. That is it. Five flows. The other modules - lead scoring models, ABM orchestration, custom journey builders, the AI add-on tier - sit dark.

Gartner reports that marketers use only a fraction of the martech capability they pay for, and utilization has fallen year over year. When your stack utilization sits below 60%, every renewal dollar past the CRM is funding shelfware. The gap is widest at companies that bought up-market early, which is exactly the sub-30M operator who took the enterprise tier on a growth bet that never required the extra modules.

The seat math compounds it. Enterprise marketing tiers price per contact and per seat. Cross 40k a year and you are typically buying capacity for a 50-person marketing org you do not have. A 6-person team does not need 40 builder seats. It needs 5 workflows that fire reliably. Count the logins that have not touched the platform in 90 days. On most stacks we audit, more than half the provisioned seats are dormant, and every dormant seat is a fixed cost with no return attached to it.

The contact-tier trap is quieter but worse. These platforms bill on total contacts in the database, not active ones. Your list grows, your bill grows, and most of those records have not opened an email in 18 months. You are paying a per-contact rate on a graveyard. Run a filter on last-engagement date and you will usually find that 40 to 70% of your billed contacts are inert. That portion of the contact bill buys you nothing but a bigger renewal number next cycle.

We ran the numbers across 50 mid-market stacks and published the breakdown in our audit of 50 broken stacks. The headline: most operators cannot name the 5 workflows their suite actually runs, which is the tell that they are paying for capacity, not outcomes.

The named answer: a closed-loop build

Keep the CRM of record. Stop paying for the automation suite. Rebuild the 5 workflows you run on Make.com or n8n, wired straight into the CRM you already trust. We call this the closed-loop build, and it ships in 14 days.

The logic is plain. Your CRM holds the records. The automation layer moves data between systems and fires the sequences. You do not need an enterprise marketing tier to do that. You need a workflow engine and the integrations. n8n self-hosts, and its published pricing starts near free for the engine itself, with paid tiers in the low hundreds per month. Compare that to the suite line on your platform invoice.

We see operators do the same on the front end with GoHighLevel for outbound and Shopify flows for ecommerce, all stitched through the same engine. The build is boring on purpose. Boring ships, and boring does not break.

Boring also means observable. When a Make.com scenario fails, you see the exact module that broke and the payload that broke it. When an enterprise journey builder silently drops a contact, you find out from the prospect who never got the email. The flows are yours, portable, and readable by anyone on your team in an afternoon.

There is a second-order win operators miss until they live with it: change velocity. On a suite, editing a journey means filing a ticket, waiting for a sandbox, and testing inside the vendor's release calendar. On a workflow engine, you open the scenario, change one module, run it once against a test record, and ship. A change that took a week now takes an afternoon.

What to keep, what to cut

Keep: the CRM of record, your billing system, your email-sending domain reputation. Cut: the unused automation tier, the seats you provisioned and never assigned, the AI add-on you trialed once. The test is utilization. If a module runs fewer than 2 live workflows, it is shelfware, and shelfware is the line item to kill at renewal.

One caution before you cut: export everything first. The contacts, the email templates, the form data, the historical engagement. The suite makes leaving costly by making export tedious. Pull it all before the renewal date, not after. Export the contact list as CSV, the templates as HTML, the workflow logic as screenshots or a written spec, and the engagement history as a dated dump. That archive is also your migration source: the 5 flows you rebuild come straight from those screenshots.

A worked example: the 12M operator's invoice

Here is a representative sub-30M case built from the audit pattern, not a single client. Treat the figures as a model you can run against your own invoice, not a promised outcome.

An operator at 12M in revenue runs a marketing-hub tier at roughly 3,200 a month, or 38,400 a year. On top of that sits a contact bill: 85,000 records at a per-contact rate that adds about 1,400 a month. They provisioned 18 builder and admin seats during onboarding; 7 are still logging in. The AI add-on, trialed once, runs another 600 a month. Total platform spend past the CRM lands near 64,000 a year.

Now count what runs. Five workflows, the same five every audit turns up: lead routing, welcome sequence, re-engagement drip, deal-stage alerts, renewal reminder. Of the 85,000 contacts, a last-engagement filter shows 31,000 active in the past year. The rest is the graveyard, billed in full.

Rebuild those five flows on a self-hosted n8n instance plus the integrations, and the recurring engine cost sits in the low hundreds a month once migration is done. Even with a maintenance tier, the steady-state number is a fraction of the 64,000. The one-time cost is the build itself, priced as the real work it is, but it is a one-time number against a recurring leak. Break-even on most stacks at this size arrives inside the first year, and every year after is recovered margin.

Platform suite vs closed-loop build

Price the difference yourself with the open-loop tax calculator. Most operators find the suite line is 4 to 7 times what the equivalent workflows cost to run.

Common mistakes when you cut the cord

Killing the suite is the right move at this size, but the rebuild fails in predictable ways when operators rush it.

Cutting before exporting. The most expensive mistake is canceling first and discovering the export tools are locked behind an active subscription. Pull every asset while the account is live. Assume the door closes the moment you give notice.

Rebuilding flows nobody uses. Some teams try to port all 40 modules onto the new engine out of completeness. That misses the point. You rebuild the 5 flows that run, not the 35 that never did. The whole gain comes from declining to recreate shelfware in a cheaper place.

Ignoring sending-domain reputation. Your email deliverability lives in your domain authentication and sending history, not in the platform. Warm the new sending path and keep the SPF, DKIM, and DMARC records clean. A botched migration here costs you inbox placement, which is worse than any license fee. The deliverability research from Litmus is clear that authentication and consistent volume matter more than the platform name on the invoice.

No owner on the new stack. A portable workflow with nobody watching it is a future outage. Name one person responsible before you migrate, give them the run logs, and build alerting so a failed scenario pages a human instead of going quiet.

A four-question decision framework

Before you renew or rebuild, answer four questions in order. Each one is a gate that points you to the right call without a sales rep deciding it for you.

Can your team name the live workflows?

Ask your marketing lead to list every automation the suite currently runs, with what triggers each one. If the list runs past 10 distinct, genuinely active flows, the suite may be earning its tier. If it stalls at 5, you have your answer. The named-flow count is the cleanest proxy for whether you are paying for capacity or outcomes.

What fraction of seats logged in last quarter?

Pull the login report. Divide active seats by provisioned seats. Below 50% and the seat charge is funding empty chairs. This number alone often justifies a tier downgrade even if you keep the platform, so run it regardless of which way the rest of the decision goes.

How much of your billed list is alive?

Filter contacts by last engagement. The active fraction, not the total, is what your marketing touches. If you are billed on 85,000 and you engage 30,000, you are paying a per-record rate on 55,000 dead entries. That is the contact-tier leak in one number.

Will someone own the new stack?

This is the honest gate that sends some operators back to the suite. A closed-loop build is cheaper and portable, but it needs an owner. If you have a capable ops or engineering hand who will keep the lights on, rebuild. If you have nobody and never will, the suite's hand-holding is a real feature and you should price it as one. Pretending you have capacity you do not have is how a cheaper stack becomes a more expensive outage.

When the big platform is right

Sometimes it is the correct call. If you run at enterprise scale, with 50-plus marketing seats, multi-region compliance, and an ABM motion that orchestrates named accounts across thousands of contacts, the suite earns its price. The integrated reporting and governance are worth real money at that size. McKinsey research on operating-model scale makes the same point: centralized tooling pays off when coordination cost across many teams would exceed the license, and that crossover sits well above the sub-30M band.

It is also right when you have no engineering or ops capacity at all and need a single throat to choke. The platform support contract is a feature for a team that cannot maintain anything itself. The math we ran in our piece on agency retainers applies here too: pay for the thing only when the alternative costs you more.

The suite is a fit at the extremes. Real enterprise on one end, a team with zero technical capacity on the other. The trap is the broad middle: the profitable sub-30M operator who bought enterprise tooling because the sales deck made it feel safe, and who now funds 40 modules to run 5 flows.

Who this is NOT for

This is not for true enterprise. If you clear 100M and your marketing org is its own department, ignore everything above and keep your suite. It is also not for the operator who wants to fire-and-forget. A closed-loop build is portable and cheap, but someone has to own it. If nobody on your side will touch a workflow, the suite hand-holding may be worth the premium. And it is not for the team running zero workflows today. If you have no sequences live, you have a strategy problem, not a tooling problem, and no platform fixes that.

The next action

Pull your last platform invoice. Circle every line that is not the CRM of record. Count the live workflows each line supports. If the number is below 2 per module, you found your leak. The fix is a 14-day rebuild on a workflow engine, priced at 3,500 to 10,000 a month while we migrate, then scoped down to a maintenance tier once the 5 flows run clean. Voice agents, if outbound is part of the gap, go live in 5 days on a separate track.

Do it in that order and nothing breaks. Export first, rebuild the five flows in parallel with the suite still live, run both side by side for a week to confirm parity, then give notice. That week of overlap costs one extra invoice cycle and buys certainty that no contact falls through during the switch. It is the cheapest insurance in the migration, and skipping it is the one shortcut not worth taking.

Start with the free Closed Loop Audit to see your own utilization gap in numbers. Browse the case studies for the migration pattern, read how the automation service scopes a build, and book the audit when you want the math run against your stack. Receipts, not slides.