Proving Marketing ROI Is Starving the Work That Creates It

Proving marketing ROI has become the test every dollar must pass, and that test quietly picks winners. It rewards work you can trace to a conversion and starves the slower work that fills the funnel first, so the reports improve while the business gets harder to grow.

I told you to find the middle ground. I was half wrong.

Two years ago I wrote a piece for MarTech called “The great debate: Activity vs. results. ” I landed where most reasonable people land. Activity without results is motion for its own sake. Results without activity is a fantasy. Find the middle ground, I said, and let the two feed each other (1. De Libero, 2024).

I still believe the middle-ground part. What I got wrong was an assumption buried underneath it. I assumed “results” meant the truth. I assumed that if you measured harder, you’d see more clearly.

You don’t. Proving marketing ROI works like a filter. It decides which work you can see and which work quietly disappears. And it’s throwing out some of your best.

The proof test only sees one half of the work

Here’s the mechanism, and it’s simple once you look at it.

When every dollar has to survive a monthly review, the work that survives is the work you can trace to a conversion. Paid search. Retargeting. The bottom-funnel channels that stamp their own receipts. That work is real and it matters. But it’s a slice of what marketing does.

The other slice builds the demand those channels harvest. Brand campaigns. Community education. Physician outreach in healthcare, developer relations in software, the content that earns trust a year before anyone fills out a form. None of it draws a clean line to a sale next Tuesday. So under a proof-first regime, it loses every budget argument it enters.

Nobody decides to defund the top of the funnel. It just keeps failing the test, quarter after quarter, until it’s gone.

What the numbers are actually showing

The pressure is real, and ROI now runs the budget conversation. 84% of chief marketers name it as their primary metric for allocating spend (2. NIQ, 2025). That sounds like discipline. Look at what moves alongside it.

In the same research, the share of marketers who believe their CEO and CFO back long-term brand investment fell 11 points in a single year (2. NIQ, 2025). Support for the slow work is sliding while the demand to prove the fast work climbs. The two move together, and the reason isn’t mysterious. Budget follows the primary metric, and the primary metric can’t see brand.

The measurement itself doesn’t rescue anyone. Teams now juggle a half-dozen tools to track performance, and only 18% say they get a clear view of what’s actually working (3. Bitly, 2026). The channel most marketers use most is the one they rank as their biggest blind spot. The work almost everyone is doing is the work almost no one can measure.

Neil Welsh, founder and CEO of the performance agency Silverback Strategies, named the trap plainly: “Attribution is a credit allocation problem dressed up as a measurement problem. Solving it with better rules still gets you the wrong answer” (4. Welsh, 2026). Better math won’t rescue you when the number you’re optimizing was the wrong number.

A marketing team that optimized itself into a corner

A healthcare organization’s marketing team lived this. Under pressure to tie every dollar to a measurable return, they gradually stopped funding anything they couldn’t connect to an appointment or an inquiry. Community education, physician outreach, brand campaigns, patient stories. All cut back. The money moved into paid search and retargeting, where attribution runs clean.

For a while it looked brilliant. Cost per lead dropped. Conversion rates climbed. The dashboards went green and the team could account for nearly every dollar. Leadership saw a leaner, sharper operation.

They weren’t sharper. They’d concentrated spending at the bottom of the funnel, harvesting demand the defunded work had created.

A year to eighteen months later, the bill arrived. Fewer patients searched for the organization by name. Physician referrals softened. Several service lines missed growth targets. Paid media got more expensive, because they now had to buy the attention they used to earn. The numbers looked better right up until the business got worse.

But isn’t proof just overdue discipline?

The strongest objection: marketing spent decades hiding behind “brand” to dodge accountability, and the proof demand is the correction. Fair. Plenty of unmeasured spend deserved to die.

But accountability to the wrong number just aims precision at the wrong target. Calling that discipline is generous. A regime that can only see attributable work will fund only attributable work, then act surprised when the pipeline it stopped feeding runs dry.

The other reasonable objection: better measurement fixes this. Incrementality testing and media mix models capture lift that last-click misses. True, and worth doing. But switching to them means admitting the old numbers were built on a flawed model, and for a CMO already under scrutiny, that admission can trigger budget cuts before the better measurement proves itself. The pressure and the cadence are the trap, not the model. Sharper tools don’t remove the incentive to report the number that survives the room.

What to defend when the pressure comes

You can’t opt out of proving marketing ROI. Your CFO is right to ask. The move is to change what counts as an answer.

Name the demand-building work before you cut it, and say out loud what it protects. Ask for a measurement window that matches how the work actually pays back, not the reporting calendar. Separate the demand you create from the demand you harvest, so a falling cost-per-lead can’t disguise a shrinking market. And when someone proposes killing the unmeasurable line, make them account for what it was holding up .

This is a bigger kind of accountability, one that can see the whole business, including the parts that never fit on this month’s dashboard.