Every quarter I sit in a version of the same meeting. Marketing arrives with a presentation full of campaigns, content, events and a respectable pipeline figure, finance arrives with a different figure, and sales has a third one that nobody has reconciled against either. And the board, who only wanted to know what marketing generated this quarter, watches three teams argue about a source field in the CRM.

Nobody in that room is being dishonest. They're interpreting three systems that were never built to agree with one another, and the disagreement that follows gets filed under attribution when it's really a fight about credit.

Sound familiar?

The pressure arrived long before the answer did

According to the 36th edition of The CMO Survey, board pressure to demonstrate marketing's value jumped from 33% to 50% in a single year, with CFO pressure climbing from 52% to 63% over the same period. Research from HubSpot's 2026 State of Marketing puts ROI measurement at the top of the challenge list, and roughly three-quarters of marketers now report more budget scrutiny than they were subjected to previously. Almost 38% say their leadership treats marketing as less critical to the organisation than it used to be.

Not a comfortable year to be guessing.

The difficulty isn't that measurement became harder in some abstract, philosophical sense. Two things happened at once. The CFO became a gatekeeper on marketing investment, and the buying journey became a good deal less observable. Gartner research finds that B2B buyers spend only 17% of their purchasing cycle meeting with potential suppliers. More precision is being demanded, about a process that got harder to see.

The beer and the cashier

On one episode of the SaaS Stories podcast I had the pleasure of speaking with Jeff Greenfield, founder of Provalytics, who put the last-click problem rather better than I've managed to.

"It's like giving credit for a beer purchase to the cashier who scans it. Sure, they were the last step, but what about the ad, the recommendation, or the display that led to the purchase? Attribution isn't about the last step; it's about the journey."

Listen to the full episode of SaaS Stories with Jeff Greenfield

He was blunt about what this looks like inside a B2B organisation. Everybody lives in the CRM, every account carries a source field, and the field itself becomes contested territory. Marketing wants credit for the UTM code, sales wants credit for the outreach, and the events team wants credit for last year's conference. Jeff calls it a tug-of-war over attribution that doesn't help anyone understand what's driving account engagement.

That tug-of-war is the meeting I described at the top. And you can't win it with a better dashboard, since every team in the argument already has one.

The three agreements that shift that conversation are set out at the foot of this piece. None of them need new technology before they need a decision.

The timeline almost nobody plans for

The most useful thing Jeff said had little to do with attribution modelling. Every campaign carries its own timeline from first touchpoint to meaningful action, and understanding that timeline is what lets you plan properly when budgets or priorities shift underneath you. At his former company they analysed customer journeys that were sometimes hundreds of steps long, and found that early-stage awareness campaigns often needed 20 to 40 days before showing meaningful impact.

Now think about a campaign quietly killed at day fourteen.

Marketers under budget scrutiny cut early. It feels responsible. It looks decisive in a spend review, and it guarantees that the awareness investment will never produce the evidence it was going to produce. The team then reports a disappointing result on a campaign that was never allowed to finish, which makes the following budget conversation harder again, and the one after that harder still.

Whose fault does that look like, six months later?

Where this leaves you

Demonstrating marketing's contribution to revenue isn't a modelling problem that a cleverer tool will solve on your behalf. It's an agreement problem: which dataset everybody reads, which unit of measurement counts, and how long a campaign is given before somebody calls it. The organisations getting this right have usually reached those three agreements in a room with finance in it, well before the quarterly review.

The uncomfortable part? The agreement has to be negotiated while nothing is on fire. Trying to establish a measurement methodology halfway through a budget reduction is a conversation you will lose.

Start with the data question, since it's the cheapest of the three fixes to answer. Open the CRM, pick ten closed-won accounts from the past twelve months, and try to reconstruct what touched them and in what order. If you can't reconstruct that for ten accounts, no attribution model is going to rescue you. And if you can, you already have the beginnings of the timeline Jeff was describing.

In the next piece I'll take up the problem sitting directly underneath this one: buyers who have already assembled their vendor shortlist before marketing gets the chance to influence anything at all.