What Fixing Your Feedback Score Does to Your CPMs (Real Data, 2026)
Watch a store’s feedback score climb and something quiet happens in the background: its CPMs start to ease. Let it collapse and the opposite hits, CPMs spike even when nothing else about the account changed. We see this play out again and again across the accounts we manage. Below is real (anonymized) data from two of them, honest caveats included, so you can weigh the signal yourself.
There’s no public formula that says “feedback score X equals CPM Y.” Meta doesn’t publish one, and anyone promising you a guaranteed number is guessing. What we can hand you is first-party evidence from accounts we’ve actually worked on. The direction holds even when the exact size of the move doesn’t. So read this as a correlation with a strong mechanism underneath it, not a law of physics.
Why feedback score and CPM move together
Your feedback score is Meta’s post-purchase read on customer experience. It runs 0 to 5, built from surveys on product quality, shipping, and service. In our experience a score of 4+ is healthy. Drop below 2 and you tend to pick up a delivery penalty. Fall under 1 and you can lose the ability to advertise at all. Here’s the logic. Meta charges you partly on how good an experience it expects to send its users toward, so a weak score becomes a factor that can push up what you pay to reach people. A recovering score can pull it back down. That’s the mechanism. The data below is what the mechanism looks like once it’s running.
The metric we actually watch: CPM vs CPMr
Standard CPM (cost per 1,000 impressions) is noisy. The same person seeing your ad ten times inflates it. So internally we lean on CPMr, cost per 1,000 unique accounts reached, as a cleaner proxy for how much Meta “trusts” your ecosystem. It measures how expensive it is to reach each distinct user, not how many times you hit the same one. (CPMr is our own working metric, not an official Meta figure.) When CPMr falls while CPM barely budges, it usually means Meta is opening up reach to more unique users per dollar. That’s a trust signal.
Case 1: a DTC supplements brand, and what a trust penalty (and recovery) actually looks like
This account (data spring 2026, one account we manage) is the clearest illustration we have, because a crisis and a recovery landed back to back:
- Crisis: the main ad account got banned, payments started failing on the backup, and the Business Manager was downgraded a tier. CPMr spiked from about $99 to a peak of roughly $172. CTR collapsed from 3.8% to 1.65%. Weekly purchases fell from 222 to about 97.
- Recovery: we ran a “penalty cleanup” to reverse the structural downgrade, and Meta re-scored the account on its usual 2 to 3 week lag. About 17 days later CPMr dropped roughly 35% week over week ($181 to $116), frequency fell 30%, and weekly purchases scaled to 443. That’s a +339% jump versus the pre-boost baseline.
The detail that proves it was trust, not audience saturation
Here’s the part that rules out the tired “you just spent too much” explanation. During the crisis week we cut spend by 62%, from about $16.8k down to $6.3k. And CPMr nearly doubled anyway, $99 to $172. If high spend were driving the high cost, slashing spend should have brought it down. It did the opposite. That’s the fingerprint of a trust and quality penalty, not saturation.
Case 2: a wellness brand, CPM down about 37% as the score work ramped
A second account (data late spring 2026, anonymized) shows the calmer, everyday version of the same effect. As the feedback-score work ramped and post-purchase orders piled up, CPM on the main testing campaign slid steadily from $30.25 to $18.99, about 37% down. The more positive post-purchase signal accrued, the lower the cost to reach trended. No drama. Just a line heading in the right direction.
Case 3: a DTC brand, a shipping crisis, and a 35% CPM drop
This is the clearest one we have put on record. A DTC brand came to us mid-March with costs that had gone vertical. The account had been stable, then a crisis: internal operations broke, product issues and shipping delays stacked up, and the customer-experience signal turned against it. The CPMr peaked around 172. From the discussions with the client, the timing lined up exactly with the operational disruption, not with anything they changed in the ads.
Once the score work ramped, the numbers moved together the way they always do: CPM down about 35%, purchase volume up by more than 300%. The tell that this was a trust penalty and not audience saturation is simple: the spend increase alone did not explain the March CPMr spike, and the recovery tracked the score, not the budget. Performance turned roughly 70 days later, in line with the two to three week latency we see after the algorithm re-evaluates an account with a clean setup and sustained positive signals.
The honest part: a downgrade like that would usually take much longer to climb out of on its own. What accelerated it was fixing the underlying operations so the negative signals stopped, then making sure it would not happen again, no more failed payments, monitored metrics, a clean account. The feedback score did the pricing, the operational fix did the healing.
The honest caveats
We’d rather show you the limits than oversell the story:
- Correlation isn’t proof of causation. CPMs also move with auction seasonality, creative fatigue, and targeting. On the second account we flagged internally that the trend looked promising but needed a few more weeks before we’d call it durable.
- Tracking can lie. That same account had a pixel double-firing issue that made its ROAS numbers unreliable. Which is exactly why we lean on CPM and CPMr, a delivery-side signal, instead of trusting a ROAS figure when tracking is broken.
- No guaranteed formula. Meta doesn’t publish a feedback-score-to-CPM table, and we won’t invent one for you. The direction is consistent across accounts. The magnitude varies.
What this means for you
If your CPMs jumped suddenly and cutting spend didn’t help, don’t reach straight for saturation or bad creative. It’s often an account-trust problem, and your feedback score is the most common lever behind it. And if you’re scaling on a strong offer, protecting that score is what lets you actually deploy the margin. The practical move: audit your score, fix the customer-experience root cause, and give Meta 2 to 3 weeks to re-score. That lag is normal, not a sign it isn’t working. Start with how to fix and improve your feedback score.
Want to know what your feedback score is doing to your CPMs, and move it? That’s the core of what we do. We audit the signal, fix the root cause, and work with Meta’s re-scoring latency so the improvement holds.
FAQ
Does the Facebook feedback score really affect CPM?
In our experience across the accounts we manage, yes. A weak feedback score is a factor that tends to raise your cost to reach people, and a recovering score tends to lower it. Meta publishes no guaranteed formula, so treat it as a strong, consistent correlation with a clear mechanism (Meta pricing partly on expected user experience), not a fixed law.
How much can CPM drop when feedback score improves?
It varies by account and we won’t invent a universal number. In two anonymized accounts we managed, one saw cost-per-unique-reach fall about 35% week over week during recovery after a penalty cleanup, and another saw CPM fall about 37% as positive post-purchase orders accumulated. Direction is consistent; magnitude depends on the account.
How do I know my high CPM is a trust problem and not saturation?
A useful test: if you cut spend significantly and your cost-to-reach stays high or rises, saturation isn’t the cause. That’s the fingerprint of a trust or quality penalty. In one account we managed, spend was cut 62% during a crisis yet CPMr nearly doubled, which pointed squarely at an account-trust issue rather than audience saturation.
How long before fixing my feedback score lowers my CPM?
Expect a lag. Meta typically re-scores an account roughly 2 to 3 weeks after the underlying issue is cleaned up, so costs and reach normalize over that window rather than overnight. In one recovery we tracked, the first clear cost drop appeared about 17 days after the cleanup, right on Meta’s usual latency.
Written by Mouss, founder of Unlimited Scaling, an agency that has helped 1,000+ e-commerce brands scale and protect their Meta ad assets. Based in Bali, 8+ years inside Meta’s ad ecosystem. Follow him on Instagram @mouss_unlimitedscaling. Account figures are first-party, anonymized, and shared as directional evidence, not a guaranteed formula.
