Every Google Ads advertiser has stared at the same number: a Quality Score of 4 on the keyword that matters most, next to a competitor who pays less for the same click. The instinct is to fix the number. Google's own documentation says that instinct is wrong — Quality Score "is not an input in the ad auction" and "is not a key performance indicator." The number is a diagnostic; the auction runs on Ad Rank. This guide works the auction the way Google describes it, with the arithmetic exposed so you can price quality against bids before spending a dollar.
What Quality Score is, in Google's words
Quality Score is a diagnostic tool that gives a sense of how ad quality compares to other advertisers, measured on a scale from 1 to 10 and available at the keyword level. A higher score means the ad and landing page are more relevant and useful to someone searching for the keyword, compared to other advertisers. It is calculated from the combined performance of three components:
- Expected clickthrough rate (CTR): the likelihood that the ad will be clicked when shown.
- Ad relevance: how closely the ad matches the intent behind a user's search.
- Landing page experience: how relevant and useful the landing page is to people who click the ad.
Each component is evaluated with a status of "Above average," "Average," or "Below average," based on a comparison with other advertisers whose ads showed for the exact same search over the last 90 days. The score is a trailing 90-day read, so a fix does not show up overnight, and a broad-match keyword may carry a score with no impressions if a corresponding exact-match keyword with impressions sits in the same ad group. Changing match types does not change the calculation. A dash means not enough exact-match searches exist to determine a score.
To see it, open Keywords in a search campaign, open the columns icon and the Quality Score section, and add Quality Score, Landing Page Exp., Exp. CTR, and Ad Relevance. Historical versions let a score be segmented by day.
Ad Rank: the six factors that actually run the auction
Ad Rank is a set of values used to determine whether ads are eligible to show and, if eligible, where on the page they show relative to other advertisers' ads. Google runs different auctions for each ad location — top ads are selected by a different auction from ads that show elsewhere — and Ad Rank is calculated first to determine eligibility and a second time to determine ranking. At a high level, Ad Rank scores are based on six factors:
- Your bid: the maximum you will pay for a click; what you actually pay is often less.
- The quality of your ads and landing page: usefulness and relevance of the ad and landing page, expectations set by the ad creative, ease of navigation, and other search-quality signals — summarized in the Quality Score you monitor.
- The Ad Rank thresholds: minimum thresholds an ad must achieve to show, set to keep low-quality ads off the page.
- The competitiveness of an auction: when two ads have similar Ad Ranks, each has a similar opportunity to win; as the gap grows, the higher-ranking ad is more likely to win but may pay a higher cost per click for the increased certainty.
- The context of the person's search: the search terms entered, the person's location, device type, time of search, the nature of the search terms, other ads and search results on the page, and other user signals.
- The expected impact from your ad assets and other ad formats: sitelinks, phone numbers, and other additions, judged on relevance, expected clickthrough rates, and prominence.
Google's worked example: seven advertisers at Ad Ranks 80, 50, 30, 10, −10, −30, −35. With a minimum of 0 to show anywhere, the last three are ineligible. With 40 needed for the top location, only the first two show above the results; 30 and 10 show lower down. The advertiser at 80 takes one top and one absolute-top impression; at 50, one top impression and none at absolute top. Position is a threshold game before it is a bidding game: an ad below the top threshold cannot buy its way above the results at any price.
What you actually pay: the arithmetic
Actual cost-per-click is the final amount charged, often much less than the maximum CPC bid: you pay only what is minimally required to clear the Ad Rank thresholds and beat the Ad Rank of the competitor immediately below you. With no competitor below you, you pay only the reserve price. Actual CPC may exceed the maximum CPC with enhanced CPC or bid adjustments, and top-location CPCs typically run higher than below-results CPCs because the top thresholds are higher.
The working formula is the arithmetic expression of that rule. Ad Rank is, in simplified form, bid multiplied by quality; the click price that produces exactly the Ad Rank of the advertiser below you is that Ad Rank divided by your own quality, rounded up to the nearest billable unit — $0.01 in the United States:
Google's actual-CPC example: five advertisers at Ad Ranks 80, 50, 30, 10, and 5, competing for four positions above the results with a minimum of 40. Only the first two exceed it: 80 pays just enough to beat 50, and 50 pays just enough to beat 40. Below the results, with a minimum of 8, the advertisers at 30 and 10 show, paying just enough to beat 10 and 8; the advertiser at 5 does not show.
Put the formula to work on the numbers that decide a budget. A competitor sits below you with an Ad Rank of 50:
Doubling quality from 4 to 8 halves the price of the same position — the economic case for fixing landing pages instead of raising bids: the bid buys eligibility, but quality sets the price of every click after it.
The bid-versus-quality break-even
The simplified Ad Rank proxy — maximum CPC multiplied by Quality Score — shows when quality beats money outright. Bidder A bids $4.00 with a Quality Score of 5, for a proxy Ad Rank of 20. Bidder B bids $2.50 with a Quality Score of 10, for a proxy Ad Rank of 25. B ranks above A despite bidding 37.5% less, and B's actual CPC is the Ad Rank below divided by B's quality plus a cent: $20 ÷ 10 + $0.01 = $2.01. B wins the position and pays less than the bid that lost.
| Bidder | Max CPC bid | Quality Score | Proxy Ad Rank | Rank | Actual CPC |
|---|---|---|---|---|---|
| A | $4.00 | 5 | 20 | 2nd | Threshold / reserve only |
| B | $2.50 | 10 | 25 | 1st | $2.01 |
| Authority | Google Ads Help, About Ad Rank | Google Ads Help, About Quality Score | Simplified proxy: bid × quality | Ad Rank thresholds | Actual CPC definition, answer/6297 |
The break-even runs the other way too. To match B's Ad Rank of 25 with a Quality Score of 5, A must bid $5.00 — double B's bid — and would pay $5.01 per click against B's $2.01. At a Quality Score of 3, matching 25 costs an $8.33 bid and an $8.34 click.
Fixing each component, in order
Fix the component rated Below average first; the other two wait until it is Average or better. The component ratings, not the headline number, are the work order.
- Expected CTR is Below average — rewrite the ad. The ad gives the searcher no reason to click relative to competitors on the same query. Write headlines to the specific keyword intent, test variations, and remove generic copy that could belong to any advertiser.
- Ad relevance is Below average — restructure the ad group. This is almost always a structure problem: an ad group holding 40 loosely related keywords pointing at one generic ad will score Below average here. Split ad groups by intent, not by product, so each ad's text matches the intent behind the search that triggered it.
- Landing page experience is Below average — fix the page, not the ad. Google assesses page-keyword relevance, loading speed, mobile usability, navigability, and whether the page delivers on the ad's promise. A generic homepage, a slow load, or a page that never mentions the offer are the usual failures; a dedicated landing page per ad-group theme is the fix.
Quality also gates more than price. Ad quality helps set the Ad Rank thresholds and whether the ad qualifies to appear at all, and Ad Rank decides eligibility for ad assets such as sitelinks. A low-quality ad can lose the sitelinks, lose eligibility, and disappear from the top while a cheaper, better ad takes its place.
Founder & CEO Insight — Enow A. Jovial
Running the score at scale
A ten-keyword account can be managed in the web interface. A B2B SaaS account with 40,000 keywords cannot. The working pattern is the Google Ads API: pull Quality Score and the three component ratings nightly, export the keyword report as CSV for the weekly review, and warehouse the history in SQL so a rating can be plotted against the bid and the actual CPC it produced. That history is what turns the 90-day window from a waiting game into a measurement: the week a landing page ships, the component rating should start moving, and the CPC on that keyword should fall in the same window. If the rating moves and the CPC does not, the bid is the constraint, not the quality.
The operating sequence
- Add the four columns. In the Keywords view, add Quality Score, Exp. CTR, Ad Relevance, and Landing Page Exp., plus the historical versions, so every keyword carries its own diagnosis.
- Sort by spend, not by score. Work the keywords that consume the budget first; a Quality Score of 3 on a $4-a-month keyword is a rounding error.
- Read the components, not the number. For each high-spend keyword, find the component rated Below average and apply only that component's fix.
- Wait out the 90 days. A fix made today moves the rating gradually; judging it after a week measures noise.
- Re-price the bid after the fix lands. A keyword that moved from Quality Score 4 to 8 needs roughly half the bid to hold the same Ad Rank — lower the bid and bank the saving instead of buying more clicks at the old price.
- [x] Four columns added and historical versions enabled: every keyword shows its score and all three component ratings, segmented by day.
- [x] High-spend keywords triaged by component: each keyword above the spend threshold has exactly one Below-average component named and one fix assigned.
- [x] Landing pages matched to ad groups: no high-spend ad group points at a generic homepage; each theme has a dedicated page that names the offer in the ad.
- [x] Bids re-priced after quality moves: any keyword that gained two or more Quality Score points has its max CPC bid re-tested downward.
Who this does not help
This guide does not help an advertiser whose problem is demand rather than auction mechanics: if nobody searches the keyword, no Quality Score exists to fix, and the dash in the column is the honest answer. It does not help a fully automated Smart Bidding campaign run as a black box — auction-time quality signals still apply, but the stored 1–10 score is not what sets those bids. And it does not help anyone who treats the score as a grade to be chased for its own sake: a Quality Score of 10 on a keyword that never converts is a perfect score on a worthless auction.
The Microsoft Ads vs Google Ads search pricing playbook prices the other search auction for comparison, and the LinkedIn Ads vs Google Ads B2B lead cost playbook shows what the same budget buys when the auction is not search at all. The Google Ads vs Meta Ads small business ROAS playbook puts the search auction next to the social auction. Model the revenue side of any bid change with the ad revenue estimator before raising a max CPC — a cheaper click that never converts is still a wasted click.
