Marketing Operations

Back when a team ran on two or three tools, nobody needed this job. Email went out from the email tool, ads ran in the ad account, and the contact list lived in a spreadsheet. Everyone watched their own screen and that was enough.

Then the number of channels grew, and the number of tools grew with it. Four ad accounts, an email platform, an analytics tool recording where visitors came from, and a customer relationship management (CRM) system holding the contacts. A CRM is the system that keeps every interaction with a customer in one place.

The trouble was that none of these tools talked to each other. To connect the person who clicked an ad to the moment they joined the mailing list, and then to the moment they paid, somebody had to line up spreadsheets by hand. That somebody was whoever had time that week, which meant the method changed every time it was done.

So the Monday marketing meeting always opened with an argument about numbers. The ads person reported 420 signups this month. The email person reported 380. Finance, counting completed payments, reported 310. All three numbers came straight out of a real tool; the three tools simply counted different things. Forty minutes of a one-hour meeting went to deciding which number to believe.

Marketing operations exists to delete those forty minutes. The job is not to run one more campaign — it's to lay the floor that campaigns run on.

The day starts by checking that yesterday's data actually moved between tools. Did every new contact land in the CRM, and does each one carry the source it came from? If 40 of 500 ad-driven contacts arrived with an empty source field, those 40 will never be credited to any channel later.

The work falls into four groups.

1. Connecting and maintaining tools. When a new tool joins the stack, decide what information flows in which direction; then watch the connections that already exist so a silent break doesn't go unnoticed for a month.
2. Setting naming rules and enforcing them. If each person writes UTM parameters — the tags added to a link so the traffic it brings can be told apart — in their own style, one campaign splits into three rows on the dashboard. spring_sale, Spring-Sale, and springsale all register as separate campaigns. The rules go in a document, and links that break them get caught before they ship.
3. Building the flows that run on their own. This means putting conditions and sequences into a marketing automation tool. A lead — a prospect who has handed over contact details — downloads a guide, gets a follow-up email two days later, and moves to sales after opening that email twice.
4. Producing one set of numbers. Pull values from each tool into a single view and align what each one counts, so 420, 380, and 310 never show up in the same room again.

Revisiting the lead scoring model each quarter belongs here too. Lead scoring assigns each contact a number representing how likely they are to buy: 20 points for viewing the pricing page, 15 for a company email address, 0 for a personal one. Cross 60 and the contact becomes a marketing qualified lead (MQL) that sales will call.

Almost every seat inside and outside marketing. Nothing ships under this role's name; what ships under everyone else's name depends on conditions this role sets.

With the performance marketer, the shared ground is ad accounts and conversion records. When the ad platform reports more conversions than the CRM actually received, marketing operations is the side that finds the leak. With the CRM marketer, the work is building send lists: turning a phrase like "free users who logged in within the last 30 days" into a query the tool can execute.

With sales, the conversation is about the handoff bar and about what a contact record has to contain before it crosses over — company size, role, and how the person arrived, at minimum. When marketing passes leads and sales lets them pile up untouched, the scoring rules usually need another look, and both sides revise them together. With data analysts, the topic is attribution — the rule deciding which of the several touchpoints a customer passed through gets credit for a conversion. With legal, it's checking list conditions so no email goes to an address that never consented, since a marketing message to a non-consenting address is a fine rather than a wasted send.

"That campaign's UTMs weren't tagged to spec, so nothing is showing up on the dashboard."

The campaign may have worked fine; the tracking just isn't attached, so the results are invisible.

"Send that over as a marketing ops ticket."

It's a tool configuration or integration problem, and someone specific owns that.

"Can we revisit the lead scoring model? Sales is ignoring too many MQLs."

The handoff bar no longer matches real buying intent, so the point rules need fixing.

At Pagedeck, Ravi opens a new ad campaign and asks for tracked links. Marketing operations hands back links tagged to the naming convention and, before the campaign starts, decides how people arriving through them will be recorded in the CRM. Two weeks later, when Ravi asks how the campaign did, the answer is already on the dashboard. Without that setup, the answer would begin with opening a spreadsheet and matching names by hand.

Because the job involves tools, people often file it under IT rather than marketing. The surface of the work — settings screens and integrations — invites that reading. But what this role decides is not tooling, it's definitions. What counts as a lead, which touchpoint earns credit for a conversion, what threshold sends a contact to sales: those are marketing judgments, and they become the definitions behind every number the team reports. Treat the role as IT support and it ends up configuring decisions made elsewhere, with no authority over the standards — which brings the original argument about numbers right back.

The other common claim is that small teams don't need the role. With five people, a dedicated hire is hard to justify, so the thought is reasonable enough. The work doesn't disappear, though; it just gets spread thin across several people, each naming things their own way and counting by their own rule. After a year of that, the first dedicated hire spends their first six months cleaning up the historical record before they can report anything.