The Meta ads metrics that actually matter (and the ones to ignore)
One of the most consistent things I see when I take over a Meta account is a brand drowning in the wrong numbers. The reporting dashboard has thirty metrics on it. The monthly reports are dense with figures. And almost none of it is connected to whether the business is actually making money.
After more than a decade running Meta for ecommerce and service brands, I've learned that the skill isn't tracking more metrics. It's knowing which handful actually tell you the truth, and ignoring the rest with discipline. Here's how I think about it.
The metrics that actually matter
Contribution margin, not ROAS in isolation
ROAS is the metric every brand quotes, and on its own it's almost meaningless. A 3x ROAS could be wildly profitable or quietly loss-making depending on your margins, your shipping costs, your discounting, and your cost of goods.
What I actually care about is what the campaign contributes after the real costs are stripped out. Two brands can both run a 3x ROAS. One is printing money, the other is going backwards. ROAS alone can't tell you which is which. Contribution margin can.
If you only change one thing about how you read your account, stop looking at ROAS as a standalone number and start asking what each campaign actually contributes to the business after costs.
Blended CAC, not platform-reported CAC
Meta will tell you what it costs to acquire a customer according to Meta. Google will tell you according to Google. Both are marking their own homework, and both over-claim, because modern attribution is messy and every platform takes credit for conversions it influenced loosely or not at all.
The number that doesn't lie is blended CAC: total marketing spend across everything, divided by total new customers. It doesn't care which platform takes credit. It tells you what growth is actually costing you. When I assess whether a brand's acquisition is healthy, blended CAC is the anchor, and the platform-reported numbers are context, not truth.
New customer acquisition cost specifically
Most accounts blend new and returning customers in their reporting, which flatters the numbers. Returning customers are cheap to convert, so mixing them in makes acquisition look more efficient than it is.
The metric that matters for growth is what it costs to acquire a genuinely new customer. That's the number that tells you whether you can actually scale. If new-customer CAC is healthy, you have room to grow. If it's only the blended-with-returning number that looks good, you may have a business that looks fine and can't actually scale.
Contribution after acquiring the customer, over time
A customer isn't worth what they spend on the first order. They're worth what they spend over the lifetime of the relationship. A brand with strong repeat purchase can afford a higher acquisition cost because the customer pays back over time. A brand with weak retention has to make all its margin on order one.
I look at what a customer is worth over time against what they cost to acquire. That ratio tells you whether the growth is healthy or whether the brand is running to stand still.
The metrics to mostly ignore
Reach and impressions
Reach and impressions tell you the platform showed your ad to people. They tell you nothing about whether the business made money. They go in the "vanity" pile. I almost never look at them when assessing account health.
Engagement metrics on conversion campaigns
Likes, comments, shares, post engagement. On a brand awareness play these might matter slightly. On a conversion campaign for an ecommerce brand, they're noise. An ad can have huge engagement and sell nothing. An ad can have modest engagement and be your best performer. Engagement and revenue are not the same thing, and conflating them leads brands to keep creative that feels good and sells poorly.
Click-through rate as a standalone signal
CTR is mildly useful as a creative diagnostic, a very low CTR can flag a weak hook. But a high CTR means nothing if those clicks don't convert. I've seen brands celebrate a high-CTR ad that was quietly losing money because the clicks were curiosity, not intent. CTR is a minor supporting signal, never a headline metric.
Frequency, in isolation
Frequency gets treated as a problem the moment it rises. Sometimes high frequency genuinely signals fatigue. Sometimes it's completely fine because the audience is small and qualified and converting well. Frequency only means something read alongside performance. On its own it's not a number to panic over.
Why this matters
The reason brands track the wrong metrics isn't stupidity. It's that the wrong metrics are easier, more available, and more flattering. Reach is always going up. Engagement feels good. ROAS as a clean number is simple to quote in a meeting.
The metrics that actually matter are harder. They require knowing your real margins. They require blending data across platforms. They require thinking about a customer over time, not just at the point of first sale. They're less flattering and more honest.
When I run a diagnostic on a Meta account, the first thing I do is throw out most of the dashboard and rebuild the reporting around the four or five numbers that actually connect to whether the business is making money. Almost every brand I work with has been making decisions on the comfortable metrics instead of the true ones.
If your Meta reporting has thirty numbers on it and you couldn't quickly say what each campaign contributes to the business after real costs, your reporting is working against you. Strip it back. Track what's true, not what's available. The brands that read their accounts honestly make better decisions than the brands drowning in numbers that feel like progress.
If you want a senior diagnostic of what your Meta account is actually telling you, you can book a discovery call with me here. 15 minutes, no pitch, no pressure.