The three components, and which one you control
Google reports one number from 1 to 10, plus three component ratings that are each just Below average, Average or Above average. The three ratings are where the instruction lives — the composite number only tells you whether to care.
Expected click-through rate predicts how often your ad gets clicked for that keyword, normalised for position. Driven by ad copy and by how well the keyword matches what the searcher meant.
Ad relevance measures how closely your ad matches the intent behind the keyword. Below average here nearly always means one ad group is carrying keywords about three different things.
Landing page experience covers what happens after the click: relevance to the query, useful content, transparency about who you are, navigability, mobile load speed. This is the component with the widest gap between how much it matters and how much attention it gets — because fixing it means talking to whoever owns the website, and rewriting ad copy doesn't.
What the score is worth in rupees
Actual cost per click on Google Search is, roughly, the Ad Rank of the advertiser below you divided by your Quality Score, plus a paisa. Hold the competitor constant at Ad Rank 160 and change only your score.
| Quality Score | Actual CPC | Clicks for ₹1,00,000 |
|---|---|---|
| 4 | ₹40.00 | 2,500 |
| 6 | ₹26.67 | 3,750 |
| 8 | ₹20.00 | 5,000 |
| 10 | ₹16.00 | 6,250 |
The number you see is not the number being used
The Quality Score in your keyword table is a historical diagnostic, rounded to a whole number and updated periodically. The auction computes quality fresh for every single search, factoring in the actual query, device, location and context.
So a keyword sitting at 7 can behave like a 9 on a tightly matched query and a 4 on a loose one. Broad match makes it worse: one keyword fired at dozens of intents, and you see the blended average.
Two rules follow. Optimise the three components, not the displayed number. And don't compare scores across accounts or industries — it's relative to other advertisers on the same keyword, so a 6 in insurance and a 6 in handicrafts describe different situations entirely.
Responsive search ads changed how ad relevance works
Expanded text ads were retired in June 2022. Responsive search ads are now the only search ad you can build, and that quietly rewrote ad relevance.
You no longer write one ad and measure it. You supply up to 15 headlines and 4 descriptions and Google assembles combinations, so relevance is assessed across whatever it chooses to serve. A weak headline can get paired with your keyword on a query where it makes no sense, and you'll never see it happen.
In practice: pin your two or three non-negotiable headlines so the keyword or offer always appears, keep the rest genuinely varied rather than fifteen rewordings of one claim, and split ad groups by intent harder than you did in the ETA era. Ad Strength rates asset variety only — it is not Quality Score, despite sitting next to things that are.
Where to find it, and what to fix first
Open a campaign, go to the search keywords table, then Columns — modify columns — Quality Score. Add all four columns plus their historical versions, which are the only way to see whether last month's fix did anything. A dash means the keyword hasn't accumulated enough impressions: no score, not a bad one.
Then sort by cost, descending, and work down. A keyword spending ₹22,000 a month at a Quality Score of 4 deserves a fortnight. The one spending ₹300 at a 3 deserves deleting.
Start with landing page experience, every time. It's the component most likely to be Below average, and unlike ad relevance it improves your cost per click and your cost per acquisition at once. Fast page, message matching the ad, one obvious action, real contact details, no popup inside three seconds. Not sophisticated — just work somebody has to schedule.