What Quality Score is and is not
The documentation is unusually direct on this. Quality Score is described as a diagnostic tool meant to give a sense of how well ad quality compares to other advertisers, reported at keyword level, and not a key performance indicator to be optimised for its own sake. It is an estimate aggregated over time, which is why it moves slowly, and it is calculated for the keyword rather than for a particular search, device or location.
It is not an auction input. Google states that Quality Score is not a factor in Ad RankThe value Google Ads calculates at each auction to decide if and where an ad shows.; what the auction uses is an auction-time assessment of ad quality, built from the same three ingredients but calculated for the specific query, device, location and moment. The reported score is a summary of how those assessments have gone recently. The relationship is like that between a credit score and a lending decision: correlated, built from the same history, but the score is not consulted when the decision is made.
Three further limits follow. The score does not account for auction-time factors such as the actual search term or the device, so it cannot tell you why one search was expensive and another cheap. It is reported for keywords only, so ad groups, campaigns and accounts have no documented score; averages that tools compute are weighted inventions. And a keyword with too few recent impressions shows no score at all, which is the normal state of most long-tail keywords in most accounts. Treat the score as a map of where quality is weak, read the components beneath it, and act on those.
The three components and how they are rated
Google reports each component as Above average, Average or Below average, alongside the overall score, in the keyword columns of the Google Ads interface. The three are documented as follows.
| Component | What Google says it measures | What it does not account for |
|---|---|---|
| Expected click-through rate | How likely the ad is to be clicked when shown for this keyword, based on past performance and excluding the effects of ad position, assets and other formats | Position: a keyword that has only ever shown low on the page is not penalised for the clicks that position costs |
| Ad relevance | How closely the keyword matches the message in the ads of its ad group | Whether the ad is persuasive; relevance is about match, not quality of writing |
| Landing page experience | How relevant and useful the landing page is to people who click the ad: relevance, transparency, ease of navigation, and, in practice, speed and mobile usability | Conversion rate as such; the component is about the page as an experience of the click, not a sales result |
Historical columns are also available, showing how the score and each component stood on past days, which is how a change is dated. The ratings are relative: Average and Above average are comparisons with other advertisers' ads on the same searches, so a component can slip without anything in the account changing, because competitors improved. The components are not weighted equally in the documentation, and Google does not publish how they combine into the score; industry practice treats expected click-through as the one with the largest effect, but that is inference from observation rather than documented behaviour.
Quality Score, Ad Rank and what you pay
Ad Rank decides whether an ad shows, where it shows and which assets show with it. Google documents its inputs as the bid, the auction-time quality of the ad (expected click-through, ad relevance and landing page experience, assessed for that search), the Ad Rank thresholds that set the minimum quality and bid to show at all, the competitiveness of the auction, the context of the search, and the expected impact of the ad's assets and formats.
The cost of a click (CPCAverage cost of each click.) follows from the same auction. Google describes the price as what is needed to beat the Ad Rank of the next competitor below, adjusted for the ad's own quality, and it is often less than the maximum bid. The effect of quality is therefore two-sided: an ad with stronger quality factors can rank above a higher bid, and it pays less for the position it gets. A keyword with weak quality needs a higher bid to reach the same position and pays more for each click once there.
This is why the score is worth attention even though it is not an auction factor: it is the visible trace of the factors that set the price. A keyword showing Below average on two components is paying a premium on every click that the fixes below would remove. The CPC calculator can show what a change in the average price does to a budget's click volume, and lost Impression ShareShare of eligible impressions you received. due to rank is the other symptom of the same problem: the ad is not reaching the auctions it could because its Ad Rank, quality included, is too low to clear the thresholds. The Smart Bidding strategies guide covers the bid side of the same equation; this page stays on the quality side. For a programme that also buys feeds and programmatic display, how ad auctions set the price of performance media sets this search auction beside the other two.
Diagnosing a low score
Start from the component, not the number. The table maps each Below average rating to its likely causes and the fix that addresses them.
| Component rated Below average | Likely cause | Fix |
|---|---|---|
| Expected click-through | Ads that do not reflect the query; generic headlines; no assets; a keyword too broad for its ad group; competitors with stronger offers | Rewrite the ad pool around the keyword theme; add assets; move the keyword to a tighter ad group; sharpen the offer |
| Expected click-through (on a keyword that converts well) | Low-intent variants matching the keyword and never clicking | Add negatives for the variants; narrow the match type |
| Ad relevance | Ad group holds keywords with different meanings served by one ad; headlines never mention the keyword's subject | Split the ad group by theme; add keyword-reflecting headlines; check that the right ad group is winning the keyword |
| Landing page experience | Page sends every ad group to one destination; thin or off-topic content; slow or unusable on mobile; intrusive interstitials; unclear who the business is | Build or choose a page per theme; match the page to the query's intent; fix speed and mobile usability; add contact and policy information |
| All three Average with a middling score | Nothing specifically wrong; competitors are as good | Decide whether the keyword matters enough to invest in; otherwise leave it |
| No score shown | Too few impressions for a rating | Nothing to fix; wait, or widen the match type if the keyword should be serving |
Two readings prevent wasted effort. First, check the keyword's results before its score: a keyword with a low score and an acceptable cost per conversion has a price problem worth fixing but not an emergency. Second, check when the score changed, using the historical columns, and match the date to what changed in the account: a new ad, a landing page redesign, a match typeSettings that decide how closely a search must match a keyword for the ad to enter the auction. change, a competitor's arrival. The cause is usually in the change log.
Expected CTR: what moves it
Expected click-through (CTRShare of impressions that result in a click.) is a prediction about this keyword's ads, built from how they have performed on this keyword, normalised for position and without the lift that assets give. What moves it is anything that makes the ad more likely to be clicked by the people who search the keyword.
The strongest lever is the match between the query and the headline. People click ads that say what they searched for, and a responsive search adA Google Ads search ad built from multiple headlines and descriptions combined at auction time. with several headlines that reflect the keyword's theme lets the system show that match; the responsive search ads guide covers how to write the pool and this page will not repeat it. The second lever is the offer: a specific proof or benefit outperforms a generic claim. The third is the keyword's own scope. A broad match keyword, or a phrase keyword matching many loosely related variants, collects impressions from searches the ad was never written for, and those impressions are counted; the match types guide and the negative keywords that trim what a match type lets in are expected click-through tools as much as cost tools.
Because the component is relative to competitors on the same searches, it can fall when a stronger advertiser arrives. The response is the same, better match and a better offer, but the ceiling may have moved.
Ad relevance: what moves it
Ad relevance is the closest of the three to a structural check. It asks whether the ads in the keyword's ad group talk about what the keyword means. An ad group that holds "boiler repair", "boiler installation" and "boiler finance" and serves one ad will be relevant to none of them well; split it, and each keyword gets ads that mention its subject.
What moves it is therefore ad group structure first and headline writing second. Keywords grouped by meaning and intentThe goal behind a search query: to learn, to find a site, to compare or to act., with a pool whose headlines reflect that meaning in several forms, rate well. Keywords grouped by product line or by whatever the old account did, with a generic pool, rate poorly. Dynamic keyword insertion, which places the matched keyword into a headline, can raise relevance for tightly themed ad groups, and in practice it produces awkward headlines in loosely themed ones, which then fail on click-through instead.
Relevance is not persuasion. A perfectly relevant ad can be dull and rate well on this component while failing on expected click-through; the two components separate "does the ad talk about the keyword" from "do people click it".
Landing page experience: what moves it
Google documents the landing page component in terms of relevance to the keyword and ad, originality and usefulness of content, transparency about the business and how it handles visitors' information, and ease of navigation, with speed and mobile-friendliness part of the experience in practice. It is assessed by Google's systems rather than by a human, and it is the component least often fixed because the fix usually lives outside the Google Ads account.
What moves it is, first, the choice of page. The page should answer the query the keyword represents, which is the same question organic search asks of a page; the search intent guide explains how to read what a query wants from the results page, and the same reading applies to a paid landing page. Sending ad groups with different intents to one page is the most common cause of a Below average rating. Second, the content: enough of it, specific to the subject, with the business's name, contact details and policies visible. Third, the mechanics: a page that loads quickly on a phone, with no interstitials that hide the content and no navigation dead ends.
Landing page experience also connects to results that the score does not report. A page that rates well on this component tends to convert better, because the properties Google assesses are the properties people want. That is the practical reason to treat it as the first fix when the budget for improvement is limited.
Diagnosing one keyword in a furniture retailer's account
A furniture retailer's keyword "oak dining table" has a Quality Score of 4, with expected click-through Average, ad relevance Average and landing page experience Below average. It costs £1.60 per click and converts at 1.1%, against the campaign's 2.4%. The ad sends clicks to the general dining room category, which lists 180 products of which 22 are oak tables. The retailer creates a filtered oak dining tables page with its own copy and a visible delivery promise, points the ad group at it, and adds two headlines naming oak tables. Six weeks later the landing page component reads Average, the score is 7, the click price has settled at £1.25 and conversion rate at 2.6%. On the same £900 monthly spend the keyword now produces 19 orders rather than 6. The retailer repeats the exercise on the next three keywords with a Below average landing page rating.
When not to chase Quality Score
The score is a diagnostic, and some diagnoses are not worth treating. Competitor keywords will rate poorly on relevance and expected click-through almost by design, because the ad cannot mention the competitor and most searchers wanted the competitor; the question is whether the clicks that do come convert at an acceptable cost. Keywords in a discovery campaign, run on broad match to find new searches, will carry low scores while they do their job. Brand keywords usually rate high without effort and gain nothing from attention.
More generally, the score is a means. If a keyword's cost per result is within target, improving its score is an efficiency gain to schedule, not a problem to fix today. If improving the score would mean removing the offer that converts, narrowing the ad group until it has too little traffic to learn, or building landing pages for keywords that bring a handful of clicks a month, the score is leading the account in the wrong direction. Accounts that report Quality Score as a headline figure, or set targets for an "average" of it, have confused the thermometer with the room; the Google Ads expertise page describes how the author reports on quality in practice without that confusion.
If the visible symptom is a rising price per click rather than a low score, the rising CPC diagnosis separates quality from competition, bid strategy, match-type drift and traffic mix. If it is a campaign whose click-through looks low rather than a keyword's rating, the low click-through rate diagnosis separates the expected lows from the causes worth fixing.
For campaigns where the click itself is the goal, the Google Ads setup for website traffic treats landing page relevance as the main lever on what each visit costs.
Quality Score is often confused with the ad strength rating on responsive search ads; the comparison of Quality Score and ad strength sets the two side by side and says which to act on when they disagree.
Common mistakes
- Treating the score as a ranking factor. It is a diagnostic; the auction uses real-time quality.
- Reporting an account-level average. No such score exists; the keyword-level components are the data.
- Fixing the number instead of the component. Each component has its own causes; read which is Below average first.
- Sending every ad group to the home page. The most common and most fixable landing page failure.
- Pausing keywords for a low score alone. Judge on cost per result; improve the keyword's ads and page.
- Expecting fast movement. The score aggregates history; use the historical columns and allow weeks.
- Chasing the score on competitor and discovery keywords. Their job is not to rate well.
- Ignoring competitor movement. The components are relative; a rating can fall because a rival improved, and the response is still better match and a better offer.
