A marketing measurement audit is not a technical review. It is a business review that happens to involve data.
I do these audits regularly. The pattern is always the same. The company thinks they have a tracking problem. They usually do not. They have an alignment problem. Marketing defines a conversion one way. Sales defines it another. Finance defines it a third way. The tracking is fine. The definitions are broken.
So the first step in any audit is not opening GA4. It is opening a document and writing down what the company means by “lead,” “qualified lead,” “opportunity,” and “customer.” If you cannot write these down in a way that sales, marketing, and finance all agree on, no amount of technical fixes will help.
The five things I check
Once the definitions are clear, I look at five things in order.
First, the source of truth. Where does each number come from? If the revenue number comes from the CRM but the pipeline number comes from a spreadsheet someone maintains manually, there is already a reconciliation problem. Both numbers might be correct in isolation. They are not correct together.
Second, the tracking gap. I compare what the dashboard says against what the bank account says. If the dashboard says you spent fifty thousand on Google Ads last month but the invoice says forty-seven thousand, that three-thousand-dollar gap tells you something about how tracking is configured. Usually it means conversion counting is wrong, or the currency conversion is off, or someone is counting clicks instead of cost.
Third, the attribution logic. Most companies use last-click attribution because it is the default. Last-click attribution tells you which touchpoint got the credit. It does not tell you which touchpoint did the work. If you are making budget decisions based on last-click, you are probably over-investing in branded search and under-investing in awareness. That is expensive.
Fourth, the ownership. Who is responsible when a number looks wrong? If the answer is “I do not know,” that is the problem. Every number needs a person who will notice when it breaks. Without that person, the number drifts slowly until nobody trusts it.
Fifth, the decision test. If this number changed by twenty percent, would you change a decision? If not, it is decorative. Stop tracking it. Every number you track that does not change a decision is a number that wastes someone’s time and creates a false sense of precision.
What the audit usually finds
In most audits, I find that seventy percent of the tracked events are never used in any report. They were set up for a campaign that ended. Or they were set up speculatively “in case we need them later.” Later never came. The events are still firing, still costing money in your Google Ads budget if you are using event-based bidding, and still confusing anyone who looks at the raw data.
The audit usually ends with two recommendations. Clean up the existing tracking before adding anything new. And agree on definitions before building new reports. Everything else is optimization.
The goal of a measurement audit is not perfect data. It is data that changes decisions. If the audit does not end with at least one decision being different, it was a waste of time.