Few Google Ads metrics have produced as much bad optimisation as Quality Score. For years it was treated as a grade to raise, with whole account rebuilds justified by one extra point. It is worth starting from what Google itself says, because it is blunter than most people remember: the score is a diagnostic tool meant to give a sense of how your ad quality compares to other advertisers, and it is not an input in the ad auction.
In plain terms: the 1-10 number is not fed into any formula when the auction decides who shows and at what price. What does count in the auction are the real-time quality estimates that the score summarises, coarsely and with a lag.
The three components, read for what they are
| Component | What it estimates | What actually moves it |
|---|---|---|
| Expected CTR | How likely the ad is to be clicked when it shows | Keyword history, clarity of the promise, match with intent |
| Ad relevance | How closely the ad answers the intent behind the search | Coherence between ad group theme and copy; assets on one subject |
| Landing page experience | How useful and relevant the page is to people who click | Consistency with the ad's promise, speed, navigation, content |
Each reports a status: above average, average, below average. That is the real diagnostic. A 6/10 tells you nothing actionable; a 6/10 with "landing page experience: below average" tells you exactly where to look.
The working rule. Never read the number on its own. Sort keywords by cost, filter for any component below average, and work from the top down: those are the places where you are paying a quality premium on volume that matters. Everything else is cosmetics.
What to do for each component
Ad relevance below average
The quickest to fix, and nearly always a symptom of one thing: the ad group contains more than one theme. If "accounting software" and "invoicing software" live in the same group, no ad can be relevant to both. The fix is not writing vaguer ads — it is splitting the themes so the query appears naturally in a headline.
Beware one common mistake with responsive search ads: adding fifteen interchangeable headlines degrades relevance instead of improving it. Fewer headlines, all coherent with the group's intent, plus pinned headlines where the promise cannot change.
Expected CTR below average
Here the lever is differentiation, not keyword repetition. When the top four ads on the page say the same thing, CTR goes to whoever states something checkable: a price, a timeframe, a condition, a proof point. Assets — sitelinks, callouts, structured snippets — increase the ad's surface area and with it the probability of a click.
Landing page experience below average
The slowest to fix and the one with the biggest economic effect, because it also moves conversion rate. The two usual causes are a broken promise — the ad says one thing, the page shows another — and mobile slowness. We go through it in detail in landing page experience and conversion rate.
Quality Score in a broad match account
Here is the 2026 problem: Quality Score lives at keyword level, but a growing share of traffic no longer originates from a keyword in any strict sense. Between broad match, query expansion and AI Max in Search campaigns, the keyword is increasingly a steer and decreasingly the text people actually typed.
Two practical consequences:
- The score describes a partial slice of the account. A keyword scoring 9 can collect traffic from queries that look nothing like it. Complete the diagnosis in the search terms report, not in the score column alone.
- Relevance is played on themes, not exact matches. Ad groups built around a coherent intent survive expansion; ad groups built around lists of variants do not.
Policing the boundaries — which queries must not trigger your ads — becomes as much a part of quality as the copy. That is the work described in governing negative keywords and search terms.
What is not worth doing
- Chasing 10/10. It is not a business objective and not an auction input. A 7 on a keyword that brings customers beats a 10 on a marginal one.
- Pausing every low-scoring keyword. Check conversions and cost first: if they convert well, the low score is a false positive, not a diagnosis.
- Rebuilding the account into single-keyword ad groups. An exact-match-era tactic that today fragments data and slows bidding down.
- Averaging Quality Score across the account. It means nothing: it blends keywords with ten clicks a year and keywords carrying half the spend.
The quarterly review, in five steps
- Export keywords with the three status columns, sorted by cost descending.
- Isolate the ones absorbing 80% of spend: the only block worth acting on.
- Group the problems by component — it is usually one cause repeated, not twenty different ones.
- Fix relevance first (days), then the page (weeks).
- Re-measure at 30 days on average CPC and conversion rate, not on the score. That is where you see whether the work paid.
Quality Score remains a useful instrument, provided you use it the way Google describes it: a diagnosis to start from, not a grade to raise.
If you want to know where your account is paying a quality premium, the free audit starts from the keywords concentrating your spend. You can also see how we work and which services cover this ground.