A campaign report with fifteen different metrics, colourful charts and English acronyms can look rigorous without actually being so, or it can hide a real problem behind a nice presentation. Most business owners who receive this kind of report every month do the same thing: look at the biggest, most prominent number, nod, and move on. The problem is that the most prominent number isn't always the one that matters.
You don't need to be a marketing specialist to read a report critically. What you need is to know what to ask and in what order, because a report's metrics form a chain, and the meaning of each one depends on the one before it.
The right order: from the general down to what actually pays the bills
Ad platforms, and many agencies, tend to lead with the metrics that are easiest to grow and least tied to business outcomes: impressions (how many times the ad was shown) and reach (how many different people). These are numbers that almost always climb over time and look impressive on a chart, but on their own say nothing about whether the business is making money. The reading order that actually matters is the reverse: start with the business result and only work your way up to the more superficial metrics if something doesn't add up.
First: cost per result and volume of results
The first question to always ask is: how many conversions (sales, leads, bookings, whatever is relevant to your business) has the campaign generated, and at what cost each? If that number improves month over month or holds steady within a reasonable range, the campaign is working, regardless of everything else. If it worsens, you need to keep moving down the chain of metrics to understand why.
Second: conversion rate (what share of people who arrive do what you want)
If cost per result rises, the next question is where those people are being lost: is less traffic coming in, or is the same traffic coming in but converting worse? Conversion rate (number of conversions divided by number of clicks or visits) answers this. If it drops, the problem is probably on the landing page, in the offer, or in less qualified traffic arriving, not in the ad itself.
Third: cost per click and the ad's click-through rate (CTR)
If the conversion rate holds steady but cost per result still rises, you need to look at how much each click is costing and what share of people who see the ad click on it (CTR, click-through rate). A CTR that drops over time is usually the clearest sign of ad fatigue: the same people have seen the ad too many times and stopped reacting. That's the moment to refresh creatives, not to touch targeting.
Fourth (and last, in most cases): impressions and reach
These metrics genuinely matter only for campaigns focused on brand awareness, where the explicit goal is being seen by as many people as possible. For every other objective (sales, leads, bookings), they're useful contextual data, but secondary: they help you understand whether a slump comes from fewer people being available to see the ad (a saturated market) or whether the problem lies elsewhere in the chain.
The questions worth asking your agency every month
Instead of asking "how are the campaigns doing?", which invites a generic, reassuring answer, it's worth asking specific things: "what was the cost per sale this month compared with last month?", "which campaign or creative generated the most sales, and which the fewest?", "what are you going to test next month and why?". These questions force an answer backed by data, not by vibes, and let you quickly spot if someone is dressing up a bad result with filler metrics.
A warning sign worth watching for
If a report changes which metric it highlights from one month to the next (one month it's "record reach", the next "excellent CTR", the next "lowest cost per click of the quarter") without ever mentioning the final cost per result or how sales or leads are trending, that's a reasonable sign something isn't going well and is being papered over with whichever metric looks best that month. It isn't always bad faith (sometimes it's just a poorly designed report), but it's worth asking directly for the missing figure.
An example of a report read badly and read well
A monthly report showed a business owner the following headline figures in large type: "1.2 million impressions, reach of 300,000 people, CTR of 2.1%". With those numbers, the owner felt reasonably satisfied. Scrolling down the report to the conversions section, smaller and less visually prominent, cost per sale had been climbing steadily for three months, from around 22 euros to over 38 euros, while total sales volume had dropped 40% over the same period. Neither of those two figures appeared in the report's headline summary.
Asking directly about that trend, the explanation was that ad fatigue had been affecting the main creatives for weeks without anyone refreshing them, and the team had preferred to focus the report on the metrics that were actually improving. It wasn't a case of deliberate bad faith, but it was a report designed to reassure rather than inform accurately, and reading it in the right order (starting with cost per result, not reach) would have raised the alarm two months earlier.
How to ask for a better report without having to redesign it yourself
You don't need to know report design to ask for a better one. It's enough to ask that the first three lines of any monthly report always answer the same three questions, in this order: how many conversions were generated this month and how does that compare with last month, what was the cost per conversion and how does that compare, and what specific change will be tested next month based on that data. A report that clearly answers those three questions, even without a single chart, is more useful than one with fifteen visualisations that don't answer them as clearly.
A mental template for the monthly results meeting
For businesses that receive a report but don't have time to analyse it in depth every month, a simple mental template helps make the most of it in ten minutes: jot down this month's and last month's cost per result in the corner of a page, calculate the percentage change, and decide before reading anything else whether that change seems acceptable to you based on your own business (not on what the report says). Only after having that figure clear should you review the rest of the report looking for why it went up or down, not the other way round.
This reversed order (result first, explanation after) avoids the common bias of getting talked into a nice-sounding narrative before checking whether the numbers actually back it up. A well-made report should hold up just as well starting from the end as from the beginning.
What to ask when a report uses a term you don't recognise
Faced with an unfamiliar term in a report (CPA, CPM, frequency, bounce rate), the most common reaction is to nod along without asking for fear of looking uninformed. A simple way to solve this without discomfort: ask that, alongside each new technical term appearing in the report, a short explanation in plain language be included in brackets. Any reasonable agency or manager accepts this request without issue, and over time the report itself becomes a learning tool as well as a record of results.
Common mistakes when interpreting period-to-period comparisons
A frequent mistake when reading a report is comparing one month with the immediately preceding one without accounting for the business's own seasonality, which can lead to wrong conclusions: a cost per sale that rises from November to December can look like a warning sign, when it's actually an expected pattern if December is a month of higher advertising competition in that sector. The more reliable comparison isn't always "this month versus last month", but "this month versus the same month last year", which removes much of the seasonal noise and shows more clearly whether real performance is genuinely improving or worsening.
Another common mistake is comparing periods of different lengths without normalising the data: comparing a 28-day month with a 31-day one, or a campaign that ran for three weeks with one that ran for five, without adjusting the figures to a comparable daily basis, can make it look like performance changed when really only the time available to generate results changed.
Frequently asked questions
What is ROAS and why does it appear in almost every report?
ROAS (return on ad spend) measures how many euros of sales each euro invested in advertising generates. A ROAS of 4 means every euro spent generates 4 euros in sales. It's a useful metric, but it needs to be viewed alongside the product's real margin: a high ROAS on a low-margin product can still be unprofitable.
How do I know if my campaign's CTR is good or bad?
It depends a lot on the sector and the platform, but as a general reference, a CTR below 0.5% on Google search usually counts as low, while on social media a CTR below 1% usually indicates a creative that isn't working well. These figures are indicative, not fixed rules.
Should I worry if impressions drop from one month to the next?
Not necessarily, especially if cost per result holds steady or improves. Sometimes fewer impressions mean the system is being more selective about who it shows the ad to, which can be a positive, not a negative.
What does it mean for a campaign to be "in the learning phase"?
It's the initial period (or the one following a major change to the campaign) during which the algorithm is still gathering enough data to optimise well. Performance during this phase is usually worse than normal, and it's reasonable to wait a week or two before drawing final conclusions.
How often should I receive a report on my campaigns?
A monthly report with the key data is usually enough for most businesses, complemented by more frequent reviews (weekly or fortnightly) if the budget is high or the campaign is in a critical phase, like a launch or a major seasonal push.
What if I don't understand a metric in the report and I'm embarrassed to ask?
Ask anyway. An agency or manager who genuinely understands what's happening should be able to translate any technical term into a simple sentence without trouble. If they can't or won't do that patiently, that's a warning sign in itself worth taking into account.