How to Tell If Your Marketing Spend Is Working

Your marketing is working if it moved a number you committed to before the period started, at a cost the business can carry, and you can name what caused the move. Three parts, all required. Miss one and what you have is an opinion wearing a chart.

Most owners who ask this question already suspect the answer. What they lack is proof — for themselves, for a finance lead, or for the agency sitting across the table. This is the diagnostic.

What "Working" Actually Means

Working does not mean busy.

It does not mean impressions climbed, the email list grew, or the monthly report ran to nineteen slides. Those are inputs. Inputs matter, because you cannot get output without them, but an input is not evidence. Traffic that never becomes pipeline is a cost centre with good posture.

A marketing program works when it changes something the business already cared about before marketing showed up. Revenue. Qualified pipeline. The cost of acquiring a customer. Your position on the shortlist when a buyer starts looking.

Write that list down and most dashboards get shorter in a hurry.

The Four Questions That Settle It

1. Did the number you committed to move?

Someone should have written down, before the quarter began, what marketing would deliver. A range is fine. "Between 40 and 55 qualified opportunities" is a commitment. "Grow brand awareness" is a mood. If nothing was committed in advance, you are not evaluating performance — you are reading a story written after the fact, by the people being graded.

2. Can you name the cause?

Pick the biggest move in your numbers last quarter, up or down, and explain it in one sentence without using the word "overall." If your team can do that, they understand the machine. If the explanation needs four caveats and a screenshot, they are guessing with better software.

3. Did the cost hold?

Volume without a cost ceiling is not a win. Leads doubled and cost per qualified lead tripled is a worse quarter than the one before it, and plenty of reports are built to hide exactly that trade. Look at cost per qualified opportunity, not cost per lead. The gap between those two numbers is where most budgets quietly leak.

4. Would it have happened anyway?

This is the hard one, and almost nobody asks it. Seasonality, a big referral, a competitor stumbling, a price change — any of these can carry a quarter while marketing takes the credit. The reverse happens too. Good work gets buried under a bad month and gets cut for it.

We built out the full system behind these questions in The Marketing Accountability Framework. The four above are the fast version, the one you can run on a Friday afternoon with the reports you already have.

How Long Before You Can Judge It?

Sooner than your agency says. Later than your board wants.

The 35th edition of The CMO Survey, fielded in January 2026 among 308 US marketing leaders, found that the median duration of marketing's impact on customers has stretched to six months, with a real shift toward a year or longer. In the same survey, more than 70 percent of leaders said they are prioritizing immediate results over long-run gains.

Read those two findings together. The effect is getting slower while patience is getting shorter, which is how good programs get killed at month four.

Set the clock by channel instead of by quarter:

• Paid search and paid social: two to four weeks for a directional read, once volume is real.
• Email and lifecycle: one full cycle, so roughly a month.
• SEO, content, and AI search visibility: three to six months before the trend line means anything.
• Brand, PR, and category positioning: two to four quarters, judged on shortlist presence rather than clicks.

Judging SEO at six weeks tells you nothing. Judging paid social at six months means you overspent for five of them.

Why the Dashboard Cannot Settle This Alone

Platform reporting is not a neutral referee. Every ad platform grades its own work, using a method you cannot audit, and each one counts conversions its competitors also claim.

The reflex is to assume the dashboard inflates everything. It is messier than that. Haus published an analysis of 640 Meta incrementality experiments run since January 2024, with the average advertiser in the set spending $14 million a year on the platform. Meta drove roughly 19 percent average lift to the brand's primary KPI — and seven-day-click reporting actually understated incrementality by about 15 percent for direct-to-consumer brands.

So the dashboard was wrong in both directions depending on where you looked. That is the real problem. Not a number that lies high, but a number whose error you cannot predict, which means you cannot correct for it by instinct or by discount.

The fix is not a better dashboard. It is a test. If the difference between models still feels abstract, start with Marketing Attribution, Explained Simply, then come back.

The Cheapest Test Almost Nobody Runs

Turn something off.

Pick one channel, one region, or one audience segment. Hold it dark for four weeks while everything else runs as normal. Then compare the held-out group against the rest. That is a geo holdout, and it costs you nothing but nerve.

Businesses resist this for an understandable reason. Nobody wants to switch off the tap. But if four weeks without a channel produces no measurable difference in pipeline, you just learned something worth far more than the spend you paused — and you learned it before you renewed the contract for another year.

Run one holdout a quarter. Rotate which channel gets tested. Within a year you will know which parts of your spend carry the business and which parts have been riding along.

Five Signs It Is Not Working

• The report changes shape every month. Metrics that flatter get promoted, metrics that embarrass quietly disappear. A stable report is the price of an honest one.
• Nobody owns a number. Everyone owns activities. Ask who is accountable for qualified pipeline and watch the room look at each other.
• Your cost per qualified opportunity is unknown. Not high — unknown. Five years into a program, that is a decision, not an oversight.
• Wins go to marketing, losses go to the market. One-directional causation is the oldest tell there is.
• The answer to "what would you cut first?" is silence. A team that understands its own spend can rank it. Ask the question cold at your next review.

If three or more of these describe your setup, the honest answer to "is my marketing working" is that you cannot tell — which is a different problem from marketing failing, and it gets fixed first.

What to Do When the Answer Is No

Do not fire anyone on Monday.

Start by separating the two failures, because they need opposite responses. A measurement failure means the work might be fine and you cannot see it. A performance failure means you can see it clearly and it is bad. Teams conflate these constantly and end up replacing a competent agency because of a broken tracking setup.

Fix measurement first. Agree on one source of truth for pipeline. Define a qualified lead in writing, in one sentence, and get sales to sign off on it. Set a target range for next quarter before it starts. Then run one holdout.

Give it a quarter under those conditions. If the numbers still miss, you now have something you did not have before — evidence, gathered under rules everyone agreed to in advance. That is a conversation you can actually have with a partner, and it is the one 10 Tough Questions to Ask Your Marketing Agency was written for.

If you would rather not referee your own spend, a marketing audit does it from outside, with no stake in the answer.

The Bottom Line

You can tell whether marketing is working. It takes a committed number, a named cause, a cost ceiling, and one test that proves the result would not have happened without you.

Most companies have none of the four and a dashboard instead. The dashboard is the expensive part. The four questions are free.