How to Run a Marketing Audit

A marketing audit is a fixed-scope review that answers three questions: where the money went, what it produced, and which part of your reporting cannot be trusted. You can run one yourself in about two weeks, using reports you already own. No vendor required to start.

What follows is the framework we use, stripped down to the version a founder or a marketing lead can run without outside help. It assumes nothing about your stack and nothing about your budget.

What a Marketing Audit Actually Is

An audit is not a strategy session. It is not a brand refresh, a creative critique, or a list of ideas for next quarter.

An audit is an inventory with a verdict attached. You collect what gets spent, what gets measured, and what gets produced. Then you rank the gaps by what each one costs you. The output is a single page a CEO can read in four minutes, not a sixty-slide deck nobody opens twice.

That distinction matters more than it sounds. Most audits handed out by agencies are pitch documents wearing a lab coat. They are built to arrive at a scope of work. Yours is built to arrive at a decision.

When to Run One

Four triggers. Any one of them is enough:

• Spend went up and revenue did not follow.
• A retainer is coming up for signature or renewal.
• Nobody in the building can explain last quarter's biggest move, up or down.
• Leadership changed — new owner, new CEO, new head of marketing.

Annual is the right default cadence, with a lighter check at the half-year mark. Run one every quarter and you are auditing instead of operating.

Before You Open a Single Dashboard

Three decisions, written down first.

Who runs it

One person owns the audit end to end. An audit by committee turns into a negotiation, and negotiations produce consensus rather than findings. If your agency runs it, you are reading a self-assessment.

How long it gets

Two weeks for a company under $20M in revenue. Three if the data is scattered across tools nobody has logged into since onboarding. A timebox forces you to rank what matters, which is the whole point.

What it covers

The six areas below, and nothing else. Add brand perception, competitor positioning, and creative quality to the list and you will produce a document that never gets finished.

Tell the team it is happening and say why. An audit run in secret reads as a firing in progress, and frightened people answer slowly.

Step 1: Gather the Inputs

Send one request for all of it:

• Twelve months of spend by channel, by month.
• Platform exports from every ad account, same date range.
• A CRM export of leads, opportunities, and closed revenue by source.
• Analytics data covering sessions, conversions, and landing page performance.
• The software invoice list — every tool, its annual cost, its renewal date, its named owner.
• The last twelve monthly reports.
• Contracts and scopes of work for every vendor.
• Your written definitions: what counts as a lead, what counts as qualified, what counts as a conversion.

If any item takes more than a week to produce, stop and write that down. It is not a delay. It is your first finding.

Step 2: Audit the Measurement Before Anything Else

Every number you look at later depends on this step, so do it first.

Pull platform-reported revenue and booked revenue for the same period and put them side by side. They will not match. They never match. What you are testing is the size of the gap and whether anyone can explain it. A 20 percent spread between Meta's reported purchases and your bank deposits is normal. A 300 percent spread means something is double-counting, and you have been making budget decisions on fiction.

Then check four things:

• Is there one agreed source of truth for pipeline, or does each team bring its own number to the meeting?
• Is "qualified lead" defined in writing, in one sentence, with sales having signed off?
• Does conversion tracking fire once per event, deduplicated across platforms?
• Can someone name, without looking, the single biggest driver of last quarter's result?

The CMO Survey has asked marketing leaders for years whether they can prove the short-term impact of their spending quantitatively. Only about a third say yes. That is the real state of the discipline, and it is why measurement goes first in every audit we run.

If you fail this step, stop auditing performance. You cannot grade work you cannot see. Fix measurement, give it a quarter, and come back. We walk through the mechanics in Marketing Attribution, Explained Simply.

Step 3: Audit the Money

Now the spend. Three cuts, in this order.

By channel against output. Put twelve months of spend next to qualified opportunities and closed revenue from the same channel. Some channels will look worse than their dashboards claim. Those are the ones to examine, not the ones to cut on sight.

By cost per qualified opportunity. Not cost per lead. The gap between those two numbers is where most budgets quietly leak, because volume is easy to buy and quality is not.

By fees against working media. Separate what you pay people from what actually reaches a buyer. The ANA's programmatic media transparency research found that only about 36 cents of every dollar entering a demand-side platform reaches a consumer, with roughly a quarter of open-web programmatic spend going to waste outright. If a line item on your invoice says "media," ask how much of it is media.

One more cut worth taking. Gartner's 2026 CMO Spend Survey found budgets flat at 7.8 percent of revenue, with 56 percent of CMOs saying they lacked the budget to deliver their own 2026 strategy. In the same research, awareness and conversion grew to 62.6 percent of media spend while loyalty and retention fell 29 percent since 2024. Check whether your retention budget got quietly reallocated to acquisition. It usually did, and nobody decided it on purpose.

Step 4: Audit the Path a Buyer Takes

Walk your own funnel as a stranger would. Fill in the form. Request the demo. Start the trial. Time every step.

Measure stage-to-stage conversion, not just the top and the bottom. A program that generates 400 leads and converts 3 to opportunity does not have a traffic problem. Check how long first contact takes after a form submission, and check what happens on day four when nobody has replied.

Most of what an audit finds here is unglamorous: a broken form on mobile, a routing rule sending enterprise leads to a junior rep, a nurture sequence that stopped sending in March. The dull findings are usually the profitable ones.

Step 5: Audit Visibility

Two questions now, where there used to be one.

Does search find you? Organic sessions, indexed pages, the handful of commercial queries that matter, and whether the pages ranking for them are the pages you would choose.

Do AI assistants cite you? Ask ChatGPT, Perplexity, and Google's AI mode the ten questions your buyers actually type. Write down who gets named. If a competitor appears in eight of ten answers and you appear in none, that is a finding with a dollar figure attached, and it will not show up anywhere in your analytics. How to Get Cited by ChatGPT, Perplexity & AI Overviews covers the fix.

Step 6: Audit the Stack and the Ownership

List every tool, its annual cost, its renewal date, and the person whose job breaks if it disappears. Any tool without a name in that last column is a cancellation candidate.

Gartner's 2025 marketing technology survey found teams using 49 percent of their martech capabilities, down from 56 percent the year before. Half of every software dollar buys nothing. Your audit will find the same thing, and the renewal dates are where you act on it.

Then do the same for outcomes. Pipeline, cost per acquisition, retention, brand — one named human per line. If a line has no name, or three names, you found why the number drifts. The Marketing Accountability Framework lays out the full version of this exercise.

Step 7: Rank the Findings and Write One Page

Rank by dollars at stake divided by effort to fix. Not by how annoying it is. Not by whose fault it is.

Each finding gets five fields: what you found, the evidence, what it costs per year, the recommended fix, and who owns it with a date. Five fields, one line each.

Then cut the list to your top three. A twelve-item action plan is a way of doing none of them. Three fixes with owners and dates will outperform a complete list every time, and in ninety days you will have evidence instead of opinions.

The Mistakes That Ruin Audits

• Starting with creative. It is the most fun part and the least diagnostic.
• Auditing without a baseline. If no number was committed before the period, you are grading a story written after the fact by the people being graded.
• Letting the audited party run the audit. Agencies do not find agency problems.
• Producing findings with no owner and no date. That is a reading document, not an audit.
• Confusing a measurement failure with a performance failure. The work may be fine and invisible. Those two problems look identical on a report and need opposite responses, which we unpack in How to Tell If Your Marketing Spend Is Working.

How Long Before the Audit Pays for Itself

The cancellations pay for it in month one. Unused software and misrouted spend are easy to find and easy to stop.

The rest takes longer. The CMO Survey's January 2026 edition puts the median expectation for marketing to show impact at six months, which is a fair benchmark for the structural fixes an audit surfaces. Measurement repairs show up inside a quarter. Visibility work takes two.

The Bottom Line

A marketing audit is two weeks of unglamorous work that ends in a one-page document with three fixes, three owners, and three dates.

Most companies skip it because it produces no campaign, no creative, and nothing to show a board. Then they spend another year funding channels nobody can defend. Run the audit. The expensive part is not doing it.

If you would rather have someone outside the building run it, that is what our marketing audit is for — no stake in the answer, and no scope of work waiting at the end of it.