Every month a nice PDF arrives. Colorful chart, arrow pointing up, a sentence saying the campaign performed well.
And you close the file without knowing whether you made or lost money.
That is not you being slow. It is how the document is designed. I received those reports for years, as the owner of a children's entertainment business inside shopping malls, and it took me a long time to understand what was bothering me. It was not missing information. It was that all the information answered a question that was not mine.
This article shows how to read what you already receive, what to ask about what is missing, and why there is a ceiling on how much that document can improve.
The report is not lying, it is just answering a different question
Start here, because it changes the tone of the conversation you are about to have.
In most cases, the numbers in the report are correct. The reach really was that. The impressions really happened. Nobody made anything up.
The report answers: "here is what we did last month". It is a proof of service document, and in that role it works.
Your question is a different one: "how many customers did this bring me, and at what cost". Those are different questions, and the second is harder to answer, because it requires connecting what happened in the ad account with what happened in your bank account. Connecting those takes work and, in many cases, exposes a bad result.
So the document settles into what is easy to show and hard to dispute.
Vanity metrics: reach, impressions, likes, CTR on its own
It is worth translating each one, because half the discomfort comes from the vocabulary.
Impressions. How many times your ad appeared on somebody's screen. Appearing is not being seen, and being seen is not sparking interest.
Reach. How many different people saw it. A bigger number than last month does not mean they were the right people.
Likes, comments and shares. Reactions on social media. They help build reputation over time, but nobody pays a bill with a like.
CTR. Short for click through rate: out of every hundred people who saw it, how many clicked. It is a useful metric, but only when compared against itself, across your own ads. On its own, it says nothing about money. An ad can get plenty of clicks and no sales, and that happens all the time.
Cost per click. How much you paid for each click. Also useful, also insufficient. Cheap clicks from the wrong audience are efficient waste.
Notice the pattern. All of these metrics measure what happens before the person reaches you. None of them crosses the border that matters, which is a person becoming a customer.
They are not wrong metrics. They serve the person operating the campaign, comparing one ad to another and deciding what to cut. What is wrong is those being the summary handed to the owner.
The four questions a useful report answers
If I could swap the whole PDF for four lines, these would be them.
1. How much was invested, in total? Ad budget and fees, separately. They are different things and they turn into one thing in a lot of people's heads.
2. How many people got in touch? Forms filled in, WhatsApp messages, phone calls. The raw number of new people talking to you.
3. How much did each of those people cost? Just divide the investment by the number of contacts. That is your cost per lead. Lead is the short way of saying "a person who showed interest and left a contact".
4. How many became customers, and what did each customer cost? Here is the line that decides everything. If ten people contacted you, two closed and you invested an amount X, the cost per customer is X divided by two.
With those four lines you can make any decision. Without them, none.
And it is worth noting: the agency has the first two. It can calculate the third. The fourth depends on you, because only you know who closed. No report solves the fourth on its own, and that is exactly why conversations about reporting tend to stall.
Cost per lead and cost per customer: the two missing lines
I want to press on those two, because they change the conversation from "is this good?" to "is this expensive?".
Cost per lead is easy and it lets you compare channels and campaigns against each other. Cost per customer is the real number, because that is the one you compare against what a customer is worth to you.
And then comes the calculation almost nobody does, and it is the simplest one in all of marketing: how much do you earn from a new customer, on average, counting everything they buy over time?
If you do not know that number, no cost per customer will look good or bad. It will just look like a number. With it in hand, reading becomes trivial: if the customer is worth more than they cost, scale. If they are worth less, cut or fix.
Do that math on paper today, even roughly. It is worth more than any report you will receive this year.
Why the number changes from month to month (and why that is a signal)
A detail that slips by: pay attention to whether the highlighted metrics change from one report to the next.
One month the headline is reach. The next month it is engagement. The month after that it is video views.
Sometimes it is a campaign change, and that makes sense. But when the headline metric changes every month without any change in strategy, what is usually being picked is the metric that went up.
The test is simple and you can run it yourself: take the last three reports and see whether there is one line present in all three, with the same name and the same calculation. If there is not, you have no historical series, and without a historical series you cannot tell whether things are improving.
Ask for one fixed line. Whichever it is, as long as it is always the same.
The three questions you can email today
Copy, paste, send. No accusatory tone, because the goal is information, not confrontation.
1. "How many new contacts, counting forms and messages, came from the campaigns last month, and what was the cost per contact given the budget invested?"
2. "Can I receive, every month, the same table with investment, contacts generated and cost per contact, so I can compare months side by side?"
3. "Are the pixel and the ad account in an account under my name or under yours? If they are under yours, what do I need to do to get administrator access?"
The third one looks out of place in an article about reporting. It is not. It is the most important of the three, and the next section explains why.
The report you build yourself once the data is in your account
There is a ceiling on how much the document you receive can improve, and the ceiling is not about goodwill. It is about architecture.
The report is a summary of data that lives somewhere else. As long as the pixel, the ad account and the tag sit inside another company's account, you depend on somebody to go and look. You receive an excerpt, in the format of whoever excerpts it, on the day that suits them.
When that same data is born in accounts under your name, the nature of the thing changes. You do not request a report. You open a screen.
Today I have my own dashboard with four numbers that I read in thirty seconds: visits, new contacts, cost per contact and sales. It is not an expensive system, it is a page that runs on my computer, built by an artificial intelligence from a request written in plain language. When I want to see something different, I ask and it changes.
The honest comparison is not between a bad report and a good report. It is between waiting for a document and opening a screen.
And the condition for that is not technical, it is ownership. If the pixel and the tag are in your name, the data reaches you. If they are not, none of this exists, however willing you may be.
Where to start
You do not have to decide anything about your agency today. Send the three questions and see what comes back. The answer, and above all how easily it arrives, already tells you a lot.
What you can do without depending on anyone's reply is check who owns the accounts that hold your history. It is quick and you do it on your own.
And there is one number the report almost never carries, the one you would need to judge everything else in it: what you pay today for each new customer. The free Cost Per Customer X-Ray, delivered by email, works it out with you. You paste the prompt into any free AI, answer a few plain questions about your spend and your new customers, and in about five minutes you have the figure. No report request, no waiting.
To keep going:
- How to install the Meta Pixel on your site and check that it is measuring, the practical part of having the data in your account.
- Domain, pixel and accounts: how to check what is really in your name, the twenty minute routine.