Meta Ads account health score: how it works and how to use it
Updated 2026-10-06 by the Cofoundr team
A health score turns dozens of metrics into one number per account so you can see which clients need attention first. The useful kind compares an account with its own recent normal, not with a fixed industry target that may not fit.
The idea: compare today with the account's own average
Fixed targets fail because every business is different. A better score asks whether efficiency, volume and delivery today are in line with the selected period's daily average. If cost per result jumps while results fall, the score drops, even when the absolute numbers look fine on paper.
Three parts of the score
- Efficiency: cost metrics such as cost per result, compared with the baseline. Lower cost than usual scores well.
- Volume: results and reach compared with the baseline.
- Delivery: whether budget is being spent as expected and ads are reaching people.
Weights depend on the campaign type
An ecommerce account cares most about efficiency, an awareness account about volume. When an account mixes types, each type counts in proportion to its share of spend, so a small awareness campaign does not distort a large sales account.
Overrides for real trouble
Some situations should drop the score immediately: spend that suddenly falls to zero, a budget burning at more than double the pace, or a sharp cost per thousand impressions spike. New campaigns still in the learning phase and campaigns just resumed are given room, so they do not create false alarms.
Using it day to day
Sort clients by score each morning, open the lowest, read the one-line primary driver and act. Cofoundr calculates this for every connected account and shows the driver in plain language, with score bands of Excellent, Needs Attention and Critical.
Frequently asked questions
Is a health score the same as Meta's account quality?
No. It is a performance score derived from your own delivery and results data, not a Meta policy or quality rating.
Why does a quiet new day not tank the score?
When there is no data yet for today, the score reflects the period average and says so, instead of treating zero as a failure.
