Google just released its annual Ads Safety Report, laying out how many ad accounts it suspended last year, how many ads it disapproved, and some high-level claims about how it is enforcing policies across the ad network. There are some genuinely interesting numbers in there, and some that deserve a closer look.
At StubGroup, we have an entire team dedicated to Google Ads policy compliance, helping advertisers resolve suspensions, fix ad disapprovals, and stay on the right side of Google’s ever-evolving rules. That gives us a front-row seat to how enforcement actually plays out, not just how Google frames it in a report.
So let’s walk through Google’s claims, match them against the data, and compare both to what we see every day on the front lines.
Claim 1: Google Is Blocking Ads With Greater Precision
Google is leaning heavily on Gemini and AI in the 2025 report, crediting these tools for catching ads before they ever enter the system. According to Google, roughly 99% of policy-violating ads are now stopped at the creation stage, rather than being flagged hours or days after they’ve already started serving.
Historically, the pattern looked like this. You would create an ad, submit it for review, and then somewhere between a few hours and a day later Google would come back and say there was a problem. Sometimes the ad would even be eligible and start serving before getting flagged.
This claim actually lines up with what we’re seeing. Much more often now, when an advertiser tries to save an ad that Google has an issue with, the system blocks it on the spot. A good example: if Google believes a website has been compromised or hacked, trying to create an ad for that site will trigger an immediate disapproval at the point of entry, before the ad can even make it into the review queue.
So on this one, the data matches reality. Google is, in fact, catching more problems earlier.
Claim 2: Google Is Working Faster Than Scammers
Google also reports that it’s processing user feedback on problematic ads much more efficiently, claiming its teams took action on more than four times as many user reports in 2025 compared to the prior year.
I can’t verify that internally, obviously. I don’t have access to Google’s systems. What I can tell you is what we continue to see on our end, and it does not always line up with that narrative.
Here’s a live example. We have a client that sells kimonos. One of their competitors has been running Google Shopping ads for years with product titles that claim the products are silk. Click through to the actual product page, and you’ll find the item is not silk at all. It’s a high-quality polyester charmeuse that is designed to feel like silk, but is not, in fact, silk.
Why does this matter? Because the price point for polyester is much lower than real silk. So shoppers see what looks like a great deal on silk in Google Shopping, click through, and convert at a price that an actual silk seller cannot match.
We have reported this to Google many, many times. I’ve personally spoken with Google reps about it more than once. They agree it doesn’t look right. And yet no action has ever been taken.
I’m not trying to call out a specific advertiser here. I am calling Google out a bit, because when the same misrepresentation persists for years despite repeated, documented user reports, the “working faster than scammers” story starts to feel less complete than the headline number suggests.
Claim 3: Incorrect Advertiser Suspensions Are Down 80%
This is a big one. Google says it has reduced incorrect advertiser suspensions by 80%.
If that is accurate, it’s welcome news. Incorrect suspensions are one of the most painful experiences an advertiser can go through, especially when you’re a legitimate business that’s followed the rules and suddenly find yourself locked out of the platform.
From our vantage point, we are still seeing a steady stream of advertisers reaching out to us who have been wrongfully suspended. Businesses that are legitimate, not violating policies, trying to do everything right, and still get shut down for reasons that often require significant detective work to identify and resolve. We get those accounts reinstated regularly, but the volume has not noticeably dropped.
So I hope Google’s 80% number is real. From what we can see, there is still a long way to go.
The Numbers: Suspensions and Ads Blocked
Let’s get into the actual data Google reported.
In 2025, Google says it suspended more than 24.9 million advertiser accounts. Compare that to 2024’s report, where they reported suspending 39.2 million advertiser accounts. That is a significant year-over-year decrease.
On the ad side, 2024 showed over 5.1 billion ads removed and 9.1 billion restricted. In 2025, the number of ads blocked or removed jumped to 8.3 billion, while restricted ads dropped to 4.8 billion. So more ads are being fully removed rather than just restricted, which tracks with Google’s story about catching problems earlier with AI.
For additional historical context, you can also look at the 2023 Ads Safety Report to see how the trend lines have shifted over the past few years.
The Country-Level Data Gap
This is where things get curious.
Google broke out country-level data for 14 countries, including the US and the UK. I pulled all of it into a spreadsheet, because I love digging into numbers like this.
When you add up the suspensions reported across those 14 countries, you get a total of about 11,335,000 suspended accounts.
That is less than half of the 24.9 million total suspensions Google reported globally.
So where are the other 13 million plus accounts coming from? Which countries are they from, and why aren’t those broken out? The report doesn’t say. It’s a big gap, and it’s worth flagging because country-level enforcement patterns matter a lot to advertisers running international campaigns.
The US Numbers and a 2024 Data Oddity
For the US specifically, Google’s 2025 report shows 3.3 million ad accounts suspended and 1.7 billion ads removed.
I wanted to compare that to the US number from 2024, but there’s a strange inconsistency in last year’s report. In 2024, Google reported 39.2 million total global suspensions. When you go to the country-level enforcement insights section and look at the United States, the US number is also listed as 39.2 million. Meanwhile, the UK is listed separately at 1.1 million, and other countries have their own distinct numbers.
Something is clearly off with that 2024 data. Either the US number is wrong, or the global total is wrong. My best guess is that the US figure in the 2024 report was inaccurate and the actual US number was considerably lower than 39.2 million. But that’s speculation, because we don’t have access to Google’s internal data.
Either way, it does look like fewer ad accounts were suspended in 2025 than in 2024, and in theory that decrease is tied to Google’s improved AI enforcement.
One important caveat. 2024 itself saw a massive spike in suspensions compared to earlier years. So we’re looking at a graph that rose sharply in 2024, dipped a bit in 2025, and it will be interesting to see where 2026 lands.
The Bottom Line
The data in Google’s 2025 Ads Safety Report is fascinating, and I appreciate that Google continues to publish these numbers and invest in keeping its ad ecosystem clean.
At the same time, there is still a lot of work to do.
We still see plenty of incorrect suspensions. We still see legitimate advertisers struggling to navigate Google’s appeal systems. We still see reported policy violations that persist for years without enforcement. The AI improvements are real, and the numbers are moving in the right direction, but the gap between the report and the day-to-day experience of advertisers remains meaningful.
Every day is a new day, and I hope Google keeps pushing to make the platform safer and enforcement more accurate. In the meantime, advertisers still need a clear plan for staying compliant and a clear path to resolution when something goes wrong.
If you’re dealing with a Google Ads suspension, an ad disapproval you can’t explain, or you want to make sure your account doesn’t end up in hot water in the first place, we’d love to help.
Sources referenced in this post:
