Artem Smirnov has spent 10 years in user acquisition and collaborates with Bïrch as an expert. After we published our guide to Meta's account scoring, we asked him to tell us more about the new scoring system, beyond what our guide covers. Here's his account.
This year, in my view, Meta rolled out one of the key metrics every advertiser is going to be watching. There's still a lot of speculation about how to read it, since the metric is hard to pin down with numbers, but one thing is clear — Meta wants to judge advertisers by how much they frustrate customers, both in their ads and in the product experience that follows. In this article I want to break down what the ACE score is, where it came from, and what to do so Meta doesn't block you as an advertiser over a low score.
What Meta's ACE score is (not to be confused with Opportunity Score)
In 2025 Meta started scoring all advertisers with an algorithm and penalizing the bad ones with higher CPMs. Before that, it was only used for e-com advertisers — Meta ranked them with it, raising CPMs for those who frustrate customers and lowering them for those who sell honestly. The only thing that counted was whether the customer stayed happy.
How was customer satisfaction measured?
Between 14 and 60 days after a customer buys something through an ad, a survey appears in their feed: "How satisfied are you with your purchase?" Next to it is the logo of the Facebook Page they bought from.
The whole questionnaire is framed negatively — the negative answer options come first, and the questions read like "How much of a scam is this business?" and "Does the business give you refunds?"
Customers' answers are combined into a score, and since last autumn that score has been used to rate every single advertiser, not just e-com. Bad products see their CPMs go up straight away, and good ones see them go down.
Mobile app businesses land in the "bad" category by default, because:
- subscription businesses are always fraud-positive by nature
- we sell recurring payments, and customers almost never realize it
- our upsells are one-click purchases with no explicit "Buy" CTA, so tapping the button charges the customer (which they don't expect)
- many in the industry still don't have a clear cancel flow
The score itself reflects how much you frustrate customers. The lowest band is the bottom 5% of all advertisers, the highest is the top 30%.
Until autumn 2026 the ACE score only affected CPMs — bad advertisers paid more. But mobile app companies in Eastern Europe have now been told that the score will also decide how willing Meta is to work with a company at all. For the worst 5% of advertisers, Meta may take away the products big players rely on — from a credit line all the way to deleting the Business Manager. In my view, Meta will make this call before the end of 2026, and we'll see the first decisions play out. This is a serious signal from Meta, so it's worth working on this metric now rather than getting caught when it rolls out.
Right now, to see any score at all, you need to be pouring millions of dollars a month into the platform. That doesn't mean smaller advertisers don't have a Customer Experience Score, or CES, though. As far as I know, CES now kicks in once you're spending around $300k a month on Meta.
How to improve your Meta ACE score
Before getting into the fixes, one question needs answering: how do you work out your ACE score if you don't have a report that shows it?
To do that, collect all the feedback from every place where customers usually review you — support chats and emails, app stores, Trustpilot, Reddit and comments under your ads (ad frequency is above 2, so customers are quite active in posting negative comments under ads).
As you go through the reviews, the job is to find patterns in what customers are unhappy about and sort them into groups. From what I've seen, when you do this, the average product comes up against two problems customers keep raising:
- They were charged when they didn't expect it.
- They complain about product quality (which includes getting access and the cancel/refund flow).
Based on this, I've put together a set of universal fixes that every product in our industry should make to raise its score. They're aimed at removing customer frustration and pointing to the areas worth working on to improve your ACE score.
⚠️ Important disclaimer: the most important thing any developer should be doing is working on product quality and improving retention.
Sign-in after purchase
A customer buys a subscription, and then you show them lots of upsells. There can be so many upsell screens that the customer simply closes the funnel. Or, even without upsells, the share of buyers who actually sign in to the product after purchase may be fairly low. These customers are quite aggressive, because they never actually got the product, and they'll leave a lot of negative feedback. And they most likely won't stop at one place: they'll post it everywhere they can, all at once.
This metric (sign-in rate) definitely needs attention. I'd put the benchmark at 90%, but I think you should aim for 95%. The closer you get to 100%, the less negativity you'll see.
To get this right, the industry usually relies on a couple of things:
- One-tap sign-in using data passed through a token.
- Follow-up emails for customers who haven't signed in to the product.
Keep in mind that if your business uses quiz funnels, customers go through them in the in-app browser of a social network. If they close it and there's no way back to the same screen, you can end up with customers who paid but never managed to sign in, so it's important to send them an email stating "We can see your transaction, and here's your sign-in link."
The next very angry group: customers who were misled
Misleading a customer is very easy — all it takes is running creatives that show features the product doesn't have. Add a funnel that picks up where the creatives left off and amplifies the promise, and you get a customer whose expectations you'll never be able to meet. Naturally, these customers leave negative feedback too. You need fairly strict guardrails here, and campaigns like these shouldn't be scaled past them.
The next sore point: the customer didn't realize they'd be charged for a subscription
In the subscription apps industry, we often don't say on the paywall and checkout screens that we're selling a subscription, or we say it vaguely. Most customers have no idea they're buying something with recurring charges rather than a one-off product at a discount.
A variation: a very cheap subscription (under $10) that rolls into a longer one after 7 days. All of this hurts your ACE score, because the customer doesn't understand how much they'll end up paying, or for what period. A very cheap 7-day trial that converts into a monthly or three-month plan causes the most frustration of all, because nobody finds its terms obvious.
A common fix: drop this kind of subscription altogether and sell plans of a single length, at a discount. Give 50% off the first billing period and state clearly that after that the price will be XXX/[period]. Another good tactic is to push longer plans (3 months and up), since the ACE survey reaches customers within 14–60 days.
The next thing to work on: upsells and how we sell them
Case #1: one-tap payment upsells. The button usually doesn't say "Buy", so customers who tap it don't expect to be charged on the spot.
Case #2: the skip button on an upsell screen may not be visible at all.
Case #3: you have more than three upsell screens. The customer has no idea when this will end, or whether it ever will. If there's nothing like a progress bar showing what comes next (a sign-in screen or a sign-up flow after these screens), the customer may think they've landed in a scam funnel that will just keep throwing new offers at them. So they drop out of the funnel and start posting negative reviews on different platforms.
The last thing to work on: the cancel flow and, essentially, the refund flow
If your product has no profile tab with cancel and refund buttons (where the customer can tap and get confirmation that their subscription is cancelled), that's where your ACE score can grow. If customers have to fight to cancel or to get their money back, they end up leaving a lot of negative feedback on different platforms, because they feel misled.
These are the main areas where your ACE score can grow. Fix them, and I think you'll get out of the bottom 5% straight away. At the same time, these fixes can affect your economics, so it's all about balance. But the direction regulators are moving in shows you'd better reach the point where you frustrate customers as little as possible, and soon.
