Higher-rated Google ads win more impressions, not more conversions, study finds

Google Ads hands more impressions to the responsive search ad with the higher Ad Strength rating, but that ad often converts worse and costs more per conversion than the lower-rated ad running beside it. That is the finding from a new Adalysis study of more than 5,900 Google Ads accounts, published October 27 and taken up by Search Engine Journal this week — the clearest data yet that Google's four-tier Poor-to-Excellent label measures how an ad gets served, not how it performs.

The gap between the label and the result

Adalysis, the PPC platform run by longtime Google Ads trainer Brad Geddes, isolated ad groups running exactly two active responsive search ads with the same keywords, bids and budget but different Ad Strength scores. Ads had to be at least 30 days old with at least 250 impressions and one conversion; branded ads, statistical outliers and campaigns using final URL expansion were excluded from the sample.

The higher-rated ad won more impressions in 63% of ad groups — up from 56.8% when Adalysis ran the same comparison in 2022 — and the gap widens with distance between ratings: an Excellent ad out-impressed a Poor one 78% of the time. Impressions do not carry through to results. The higher-rated ad only had the better conversion rate in 45% of ad groups, and the lower-rated ad beat it on conversions per 1,000 impressions in 54% of ad groups and on cost per acquisition in 51.5%. Averaged across every ad group, higher-rated ads carried an 11.8% higher CPA — a gap Adalysis illustrates as a $100 CPA ad sitting next to a $111.80 one with the "better" score attached.

Pinning explains most of it

The study traces most of the gap to pinning — locking specific headlines or descriptions into fixed positions instead of letting Google rotate all assets freely. Pinning caps the achievable rating: lock down one or two asset slots and Good becomes the ceiling, regardless of how many headlines sit in the pool. Unpinned ads need at least 11 distinct headlines to reach Excellent, and only a third of them get there; even among ads running 15 or more headlines, just 44% hit Excellent. Fully pinned ads, meanwhile, often post the strongest conversion rates within their own rating tier, because an advertiser who hand-picks the exact headline shown usually knows more about buyer intent than Google's auto-assembled combination does.

"Often, the human ability to create compelling ads will outperform the machine's ability to serve the ads properly."

That is Brad Geddes's own reading of the data. Google maintains a different position. In Search Engine Journal's coverage this week, Google's Help Center and Ads Product Liaison Ginny Marvin are both cited saying Ad Strength "isn't used directly in the auction" and "doesn't directly influence" serving eligibility — a claim the impressions data makes hard to square at face value, even if the real mechanism sits somewhere adjacent to the auction rather than inside it.

What this means if you run Google Ads

Most SaaS marketing teams treat Ad Strength as a to-do list: get every RSA to Excellent, then move on. This data says that instinct trades performance for a label. A pinned ad stuck at Good because it carries three locked headlines may be out-converting the unpinned Excellent ad in the next ad group over, and a team optimizing purely for the badge would never notice, because Ad Strength does not surface CPA or conversion rate anywhere near the rating itself.

• Before changing an ad to chase a higher score, check whether it is pinned — pinned and unpinned ads should not be judged on the same rating scale
• On unpinned ads, build toward Excellent with genuine headline variation, not reworded duplicates — the study puts the practical minimum at 11 distinct headlines
• On pinned ads, treat a Good rating as a ceiling rather than unfinished work, and judge the ad on CPA and conversion rate instead
• Re-check any standing workflow that has your team "finishing" Ad Strength checklists — it may be optimizing for Google's impression share, not your pipeline

None of this makes Ad Strength useless. Google's own serving behavior shows it affects which ad gets shown, and impressions are a precondition for everything that follows. The problem is narrower and more fixable than that: a metric built to describe asset diversity is being read as a proxy for ad quality, and the accounts in this study show that proxy failing in more than half of cases once pinning enters the picture.