We run paid, organic, email, and content as one engine for established brands, and we have watched a lot of clients arrive from another agency clutching a beautiful dashboard that never once told them whether the work paid for itself. So this is the grown-up version: the metrics worth reporting, the vanity metrics to walk away from, and how to tell the difference in five minutes. If you are earlier in the process, our companion pieces on how to choose a marketing agency you can actually trust and the questions to ask before you sign cover the steps before the first report lands.
What is a marketing report actually for?
A marketing report exists to answer one question: is the marketing working, and is it worth the money? Everything else is supporting evidence. A good report ties spend to outcomes, shows the cost of each outcome, and tells you what the agency will change next month and why. If you finish reading and still cannot say whether the work paid off, the report has failed at its only job.
There is a quiet assumption baked into a bad report: that volume of activity equals value. Ten posts went out. Reach was up. The agency was busy. None of that tells you whether a single customer arrived because of it. A report is a decision tool, not a status update: you should be able to read it and decide to keep spending here, stop spending there, or change something. If a report cannot drive a decision, it is theatre.
What should a marketing agency report on?
A marketing agency should report on the metrics that connect spend to business outcomes: qualified leads or sales, pipeline value, revenue, cost per acquisition (CPA), return on ad spend (ROAS), customer lifetime value (LTV) where the data exists, and retention. It should show the target, the actual, the variance, and the cost of each outcome, then state what changes next and why.
Here is the shortlist, in order of how close it sits to your bank account.
Qualified leads or sales. Not "leads". Qualified leads, by a definition you both agreed. A form fill from someone who will never buy is not a result.
Pipeline and revenue. For a longer sales cycle, the honest leading indicator is pipeline value created, then revenue closed. For e-commerce, it is revenue and orders. This is the line your finance person cares about, so it belongs near the top.
Cost per acquisition (CPA). What did it cost to get one customer or one qualified lead? CPA is the most clarifying number in marketing, because it forces the question every founder should ask: is this channel paying for itself?
Return on ad spend (ROAS). For paid media, ROAS tells you what every pound of ad spend returned. It is honest only when segmented by channel and read alongside CPA and margin (a 3x ROAS on a thin-margin product can still lose money).
Lifetime value and retention. Acquisition is half the story. If you sell anything repeatable, LTV and retention decide whether the model works. A brand that already sells should expect its agency to care about keeping customers, not just buying new ones.
Cost, plainly stated. Spend, by channel, against budget. No surprises, no rounding in the agency’s favour. A report that hides what was spent is hiding the only thing that makes the rest of the numbers mean anything.
The format matters as much as the metrics. Every line should show four things: the target, the actual, the variance, and the cost. A number with no target is trivia. A number with no cost is a half-truth.
Which marketing metrics are vanity metrics?
Vanity metrics are numbers that go up without telling you whether the business benefited: impressions, reach, follower growth, total likes, engagement rate in isolation, video views, and ad frequency presented as a win. They are not useless for diagnosis, but they are dangerous as headlines, because they let an agency look busy and successful while the revenue line stays flat.
Walk away from a report that leads with these:
Impressions and reach. The default fluff metric. Impressions count how many times an ad or post was served. They cost nothing to inflate and they correlate with nothing you can bank. Reach is the same metric wearing a different coat.
Follower count and growth. A bigger audience is only worth something if it converts, and most of it never will. Buying followers is trivial; turning them into customers is the actual job. A follower chart as a headline KPI is a tell.
Total likes, comments, and "engagement". Engagement is a useful diagnostic for working out which message lands, but as a top-line success metric it is hollow. A viral post that sells nothing is a cost, not a win.
Video views. Three-second views are almost free and almost meaningless. Watch-through and what the viewer did next can be useful. "1.2M views" on its own is built to impress, not to inform.
Open rates and click rates, reported alone. For email, opens and clicks are mid-funnel signals. They matter, but only on the way to revenue per email and per subscriber. A report that stops at open rate stops one step before the part that pays.
None of these are forbidden. The problem is hierarchy. When a vanity metric is the headline and the outcome metric is missing, the report is built to flatter, not to inform.
How can you tell good reporting from fluff in five minutes?
You can spot a fluff report fast: it leads with impressions or reach, it has no target to measure against, it never states cost per outcome, it never connects activity to revenue or pipeline, and it ends without saying what will change next month. A good report does the opposite on all five. If three or more of those tells are present, the reporting is decoration.
Run any monthly report past these five questions.
- Is there a target? Every headline metric should sit next to the number you agreed to hit. No target means no accountability.
- Is cost per outcome stated? CPA or cost per qualified lead, plainly. If you cannot find it, ask why.
- Does it connect to revenue or pipeline? Trace the line from activity to money. If it stops at "engagement", it stops too early.
- Is the variance explained? Up or down, the report should say why in plain language, not just colour the cell green or red.
- Does it say what changes next? A report with no next action is a scoreboard with no game plan.
A report that passes all five is doing its job. One that fails three or more is managing your perception.
What does good reporting actually change about the work?
Good reporting changes the incentive structure. When an agency reports on cost per outcome and revenue, it is on the hook for the same number you are, so it optimises for results instead of activity. Vanity-metric reporting lets an agency look successful while the business stalls, which is exactly how the gap between what was promised and what was delivered opens up.
A report is not neutral. It quietly decides what the agency will optimise for next month. Report on impressions, and you get more impressions: louder posting, broader targeting, cheaper reach. Report on cost per qualified lead, and the work bends toward the lead, even when that means doing less and killing what does not convert.
That is why reporting is a trust issue, not an admin one. The agency that volunteers cost per outcome before you ask is comfortable being measured. We build our reporting around a simple promise: we deliver what we said we would, faster than you expected, and the report is where we prove it or admit we did not. Run by an artist, operated like an engineer. A confident agency will tell you, in the report, what did not work this month and what it is changing. The fluff exists to avoid exactly that conversation.
No reporting is perfect, and any agency claiming flawless attribution is selling you something. Platforms over-claim credit and journeys cross channels, so good reporting shows what can be measured cleanly, flags what is directional, and resists dressing up an estimate as a fact. The Social Target has spent nine years and 600+ clients learning that the brands that scale are the ones whose agency tells them the unglamorous truth on a schedule, even when it is not good news.
If this is the standard of reporting you want from the people running your marketing, tell us about your business. We will tell you what we would measure, what we would change, and what we would not promise.
↳ Frequently asked
01What metrics should a marketing agency report on every month?
At minimum: qualified leads or sales, pipeline or revenue, cost per acquisition, return on ad spend for paid channels, spend against budget, and the variance against an agreed target. Each metric should show the target, the actual, the cost, and a plain note on what changes next.
02Are impressions and reach completely useless?
No. They are fine as diagnostic context, for example to explain why results moved. They become a problem only when they are the headline and the outcome metrics (cost per lead, revenue) are missing. Activity metrics support the story; they should never be the story.
03Is ROAS a vanity metric?
Not when it is honest. ROAS is a real performance metric for paid media, but it misleads if reported without margin context or blended across channels to hide a weak one. A high ROAS on a thin-margin product can still lose money, so read it alongside CPA and margin.
04What is the fastest way to spot a bad marketing report?
Look at what it leads with and what it omits. If it opens with impressions or follower growth, has no agreed target, never states cost per outcome, and ends without a next action, it is built to reassure you rather than inform you. Three or more of those tells means decoration.