Every monthly agency report is written by the party it evaluates. The agency picks the settings, picks the questions, and picks which answers make the deck.
That’s not an accusation. Most agency reports aren’t dishonest, they’re curated, which is a different problem and a harder one to see. Somebody chose which metrics to show, which time windows to compare, and which definitions to apply. Green arrows are easy to manufacture when you control all three.
If you run a business somewhere between $5M and $50M and you’ve got an agency or two, you’ve probably felt the gap. The report says everything is up. Your revenue doesn’t obviously agree. Most owners respond by either trying to become a marketing expert themselves or quietly losing confidence in the relationship and doing nothing about it for a year.
There’s a third option, which is to audit the report on a schedule. Seven questions, asked every month, each one connecting a claimed win back to a system you own — your CRM, your bank account, your pipeline. That’s the whole discipline. Below: the questions, the patterns that give a curated report away, what a report built for accountability actually contains, and how to run this without wrecking a relationship you’d rather keep.
Why does an agency report need auditing at all?
Because good faith doesn’t fix a structural problem. Every choice in that report — which metric leads, which window gets used, how a lead is defined — was made by someone with an incentive to renew the contract. That’s true of an agency you love.
Picture how the thing actually got made. Someone opened a handful of dashboards, exported some numbers, drew some comparisons, and wrote a narrative around them. Organic traffic dipped? Lead with engagement rate. Cost per lead climbed? Compare year-over-year instead of month-over-month. Nobody lied at any point in that process. Building a report is selection from start to finish, and selection is where the argument lives. (Same structural problem as every attribution dashboard grading its own homework — the settings belong to the party being measured.)
Treat this as a governance question rather than a character question. You wouldn’t take a financial statement prepared by a vendor and file it without reconciling against your own books. Same principle. The report is the agency’s account of its own performance, and the audit is how you check that account against what your systems recorded.
The owners who handle this well haven’t learned attribution modeling or studied ad platforms. They ask one question and keep asking it: does this result show up anywhere I can check independently? Everything else comes out of that.
What are the seven questions?
Ask these of every report, every month, roughly in this order. None of them require marketing expertise. They just require you to keep asking.
1. Which of these numbers appear in my systems?
Start here. For every headline metric, find the corresponding entry in your CRM, your pipeline, or your bank account. Claimed leads should be records you can open. Attributed revenue should reconcile against deals you can name out loud. A number that lives only inside the agency’s dashboard is a claim you haven’t checked yet, and it shouldn’t carry weight until you have.
2. What is this metric a proxy for, and did the real thing move?
Every marketing metric stands in for something you actually care about. Impressions stand in for attention, clicks for interest, marketing-qualified leads (MQLs) for pipeline. So check whether the thing being stood in for moved. If impressions doubled and qualified conversations didn’t budge, the proxy stopped working, and a report celebrating the proxy anyway isn’t telling you much. Make the agency name what each metric is supposed to predict. Then check whether it predicted. (Which metrics actually predict revenue is its own conversation, and worth having with your team first.)
3. What decisions came out of this?
A report full of activity (posts published, campaigns launched, emails sent) describes effort. You’re paying for judgment. So ask what the results caused: what got scaled up, what got killed, which assumption got thrown out. If the answer is that they’ll keep monitoring, you’re funding observation. A team that owns its decisions can always tell you the last one it made and why.
4. What did you try that didn’t work?
A report with no failures in it is a brochure. Marketing involves tests, tests fail, and an agency that never reports a miss is either not running any or not telling you about them. Ask straight out: what did you try this month that underperformed, and what did it teach you? How well they answer that tells you more about their process than any green arrow on the page.
5. What would you cut from your own scope if this were your money?
Agencies expand scope by default. No monthly report has ever recommended shrinking the retainer. But an agency that’s genuinely thinking about your business can tell you which line item in its own scope is producing the least, because it has already asked itself. If nobody can name one, nobody’s looked.
6. Which of these results would have happened without you?
Some portion of your leads and revenue was already arriving before the agency showed up, from referrals and existing brand and sales relationships and plain seasonal demand. A report that claims the whole number is claiming your baseline as its own work. Ask them to separate what they caused from what they inherited. Getting that precisely right is genuinely hard. Attempting it honestly isn’t. Watch which one you get.
7. What are you measuring next month that you can’t measure yet?
Measurement should get better over time. Attribution improves, tracking tightens, definitions sharpen. An agency shipping the identical report template for eighteen months has stopped improving how it sees your business. Ask what visibility they’re building next. If there’s no answer, measurement is something they produce for you rather than something they use.
What are the tells of a curated report?
These patterns show up reliably enough to work as a checklist. Any one of them can be innocent. Two or three in the same deck is a pattern.
Metric switching between months. Clicks led last month, impressions lead this month. When the featured metric rotates, it usually rotates toward whatever went up. Fix the scorecard and the rotation stops working.
Percentages with no absolute numbers. “Engagement up 340%” means one thing at 50,000 interactions and something else entirely at 12. A percentage without its base doesn’t tell you anything. Ask for both, every time.
Activity lists dressed as outcomes. Twelve posts, four campaigns, three landing pages. That’s a timesheet. The report should account for what all of it produced, not just that it happened.
“Brand awareness” invoked whenever performance dips. Brand is real and worth investing in. But if brand awareness only comes up in the months when measurable performance falls, it’s being used as cover. Real brand investment gets declared up front and carries its own goals. It doesn’t get discovered retroactively in a bad quarter.
Screenshots instead of shared dashboard access. A screenshot is a moment the agency picked, cropped the way they wanted it. You should have direct login access to the ad accounts and the analytics property. If that access is somehow always about to get sorted out, ask why. Keep asking.
What should a good report contain?
Flip each of those and you have the spec. Five things, all of which you can ask for in a single conversation.
The same metrics every month, against agreed targets. Agree on the scorecard once — which metrics, which definitions, which targets — and then leave it alone. Consistency is the only thing that makes month-over-month comparison mean anything. If the scorecard needs to change, it changes by agreement, before the month starts.
Absolute numbers alongside every percentage. The base and the change, together, every time. That single requirement closes most of the room for cosmetic reporting.
A failures section. What got tested, what missed, what the miss taught. If your reporting doesn’t have a standing place for this, add one. It’s the fastest upgrade available to you, because it changes what the agency prepares for and not just what they show you.
Spend against the plan. What was budgeted, what was actually spent, and where the variance went. Money is the one number in the whole report that isn’t a proxy for something else.
A “decisions needed from you” section. A real partner surfaces the choices only you can make: budget shifts, positioning calls, approvals that are holding up work. If nothing has needed your decision in six months, either the work isn’t touching anything strategic or somebody else has been making those calls. You want to know which.
How do you run the audit without poisoning the relationship?
Announce it as standing practice, apply it to every vendor the same way, and let the responses sort themselves out. The framing does more work here than the questions do. This is how you review external partners now. It isn’t a referendum on one of them.
Say it plainly at the start of an engagement, or at the next monthly review of one you’re already in: “We reconcile every vendor report against our own systems. Here are the seven questions we’ll be asking each month. Telling you now so none of it is a surprise.” Announced in advance and applied to everyone, that’s governance. Sprung on one vendor after a bad quarter, it’s an ambush, and all you’ll get back is defensiveness you can’t read.
Then watch the response, because how an agency responds to the audit is the audit.
A good agency will welcome it. Serious operators want clients who look at the numbers, because scrutiny is what separates them from the competitors coasting on pretty dashboards. They’ll answer the failure question without flinching, and you’ll have dashboard access before you finish asking for it. The scope question will land as interesting rather than threatening, because they’ve already asked it of themselves.
Defensiveness is information too, though not a verdict. Some of it is just surprise, and a fair process gives a partner a full cycle to adjust. But three months in, if the failure question is still getting deflected and screenshots are still standing in for access and the metrics are still rotating after you fixed the scorecard, you’ve learned something no single report was going to tell you. And you learned it without becoming a marketing expert.
None of this is really about catching anyone out. It’s about getting you and your agency operating from the same set of facts, which means your facts, recorded in your systems, tied to your revenue. Clear ownership of decisions, measurement that stays consistent, honest accounting of what worked. That’s just how the work should run.
Where this leads
Ask these consistently for a couple of quarters and something shifts. You stop evaluating marketing by how the report makes you feel and start evaluating it by what your own systems confirm. Weak vendors surface fast. Strong ones get better, because they’re finally working for a client who can tell the difference.
The questions do have a ceiling. They’ll tell you whether the current work is real. They won’t tell you whether it’s the right work: whether the strategy behind the reports maps to where the business actually needs to go, or whether the budget is split sensibly across channels in the first place. That’s a bigger question, and answering it usually takes someone who has sat on both sides of the reporting table, built these decks, reviewed these proposals, and knows where the curation hides. Someone working alongside you rather than instead of you. If the audit turns up gaps you can’t close from where you sit, that’s the next conversation.