HomeArticles › Meta ads: the numbers that decide, and the ones that reassure

Meta ads: the numbers that decide, and the ones that reassure

An ad account reports thirty columns. Three of them decide whether you make money. The rest exist to make you feel better when you do not.

By Gaëtan Chardon, fractional CMO · Updated August 30, 2026

In short

Three numbers decide whether a Meta account is profitable: cost per real acquisition, margin per customer over their lifetime, and how often your audience sees your ads. Cost per click, click-through rate and impressions explain things after the fact, but they decide nothing.

An ad account shows about thirty columns. That abundance feels like control, while it mostly produces activity. Three numbers decide whether you make money. The others explain, after the fact, why those three move.

The three numbers that decide

Cost per real acquisition. Not cost per lead, not cost per booked call: the cost of getting a customer who paid. As long as your ad platform does not receive the purchase event with its value, it optimises toward an intermediate goal, and you are flying blind on the only thing that matters.

Margin per customer over their lifetime. This number sets your bidding ceiling. A company that knows its twelve-month margin can pay three times more than a competitor thinking in first-purchase terms, and it buys the market without ever being smarter at advertising.

Exposure rate. How many times the same person sees your ads over a period. This number explains most of the performance drops that get blamed on the algorithm or the creative.

These three share one trait: none of them can be read inside the ad interface alone. They require advertising, payment and customer lifetime to be connected. That plumbing, not targeting finesse, is what separates profitable accounts from the rest.

The numbers that reassure

They are not wrong, they are misused. People look at them when the first three go badly, because they always offer a flattering reading somewhere.

Cost per click. It falls when you reach a broad, poorly qualified audience. A collapsing cost per click is more often a warning than a win.

Click-through rate. It measures the promise, not the sale. A hook that intrigues raises this rate and can sink conversion, because it attracts the curious rather than buyers.

Impressions and reach. They measure money spent, presented as a result.

The return on ad spend reported by the platform. It counts the sales the platform attributes to itself. On an account that also sells by call, by email or with a delay, the gap with your accounting is structural. Use it to compare two campaigns against each other, never to decide whether you are making money.

Structuring an account so the numbers become readable

An unreadable account is not fixed with better creative. It is fixed with fewer boxes.

The rule fits in one sentence: every ad set needs enough conversions to learn. Below that threshold, the platform spends without converging, and your averages mean nothing.

Three practical consequences. Consolidate audiences aiming at the same buyer with the same offer. Separate only what is genuinely different, meaning another language, another currency, another landing page or another product. And keep the budget at campaign level rather than splitting it by hand, unless you have a precise reason to force the split.

A well-structured account is recognisable by one sign. When a campaign drops, you know within five minutes whether the problem is the audience, the creative or the page. If the answer takes half a day of analysis, the structure is the culprit.

Exposure rate, the number people look at too late

It explains more unexplained drops than anything else. A finite audience sees your ads a certain number of times per month. Past a threshold, the same dollars buy the same eyes, and performance decays without any other column moving.

Read this number monthly rather than daily. Over seven days it stays low and reassuring. Over thirty it shows what your audience actually absorbs.

When it climbs you have two levers, and raising the budget is not one of them. Widen the audience, or refresh the angles. A new angle is a new ad attacking a different problem, not the same ad with a different banner colour.

What to check first on an account that stopped performing

Order matters, because each step makes the next one readable.

  1. Does the purchase event arrive, with its value? If not, everything else is commentary.
  2. How many conversions per week and per ad set? If the account is fragmented, consolidate before touching anything.
  3. What is the exposure rate over thirty days? If it is high, the problem is audience or refresh, not bidding.
  4. How many distinct angles are running right now? Usually one, repeated ten times.
  5. Does the page keep the promise of the ad? A high click rate with weak conversion points at the page, not the account.

In most accounts the blockage sits in the first two points. They are also the two nobody wants to handle, because they call for plumbing rather than creativity.

Budget is not a strategy

An account that works at five thousand dollars a month does not mechanically work at fifty thousand. Scaling changes three things at once: it exhausts the audience faster, it pushes cost per acquisition up, and it demands a creative production rhythm most teams do not hold.

Before raising spend, check that production keeps up. A simple field rule: if you cannot ship new angles at the rate your audience consumes them, your ceiling is not budgetary, it is editorial.

Frequently asked questions

What is the minimum budget to test on Meta?

The useful minimum is not an amount, it is a number of events. You need enough weekly conversions for the algorithm to leave its learning phase. If your target cost per acquisition is fifty dollars, a budget producing two sales a week teaches you nothing, whatever the headline figure.

How many creatives does a campaign need?

The right question is not how many, but how many distinct angles. Five variations of the same argument are worth less than three ads attacking three different problems. The account learns on angles, not on colour changes.

Should audiences be split into several ad sets?

Less and less. Multiplying ad sets divides your conversions across several learning phases, and none of them reaches the threshold. Consolidate, unless you have a real reason to separate, such as two countries with two currencies and two landing pages.

What do you do when frequency climbs?

Rising frequency means you are hitting the same people again. Two answers, not three: widen the audience or refresh the creatives. Raising the budget without doing either means paying more to tire the same people.

Want to go through your account together and decide what to cut and what to double down on?

Book a 45-minute session

You leave with three decisions to execute and the plan within 24 hours.

Read next : Meta ads consulting · What a fractional CMO does · What is a sales funnel