Quality Score: the link between SEO and cheap ads.

Quality Score is the 1–10 number Google Ads shows next to each keyword, and it is the most misread figure in the platform. It is not the number the auction uses, it is not a target worth chasing for its own sake, and it does not move the morning after you fix something. What it does do is tell you which of three things Google thinks you are getting wrong — which makes it the cheapest diagnostic in the account, provided you know what each of those three things is worth. Two of the three are sentences inside your account. The third is your website — and it is the one that sets the ceiling the other two have to live under. That is the short version of this article: the SEO work on your site is what makes your ads cheap.

What each part of Quality Score is worth

Shares are of the 9 points above the floor: every scored keyword starts at 1, and each part adds its points as Google rates it average and then above average. The expected-CTR and ad-relevance weights are observed directly in that account; the landing-page weight is inferred from the top of the scale — the derivation is below.

What Quality Score actually is

It is a diagnostic, computed per keyword, out of the history of that keyword in your account. Google reports it as a whole number from 1 to 10 alongside three components — expected click-through rate, ad relevance and landing page experience — each rated only as below average, average or above average. Those three ratings are the useful part. The number is a summary of them.

The score is not what the auction uses. Google is explicit that ad quality is recalculated every time an ad becomes eligible for a search, taking the actual query, the device, the location and the time of day into account. Quality Score is the historical, aggregated version of that judgement, surfaced in the interface so there is something to act on. Treating it as the price you pay per click is a simplification, not an equation.

It also explains the dashes. A keyword needs enough impressions of its own before Google will score it at all: in the account we fitted this to, 130 of 149 keywords have no score, and the 19 that do are the ones carrying the traffic.

The three parts, and what each is worth

Google publishes the three parts and has never published their weights. Every percentage you will read on the subject, including the ones in this article, is somebody’s estimate. The only thing that separates them is whether the estimate was fitted to real scores or repeated from another article.

The weighting that came out is simple enough to do in your head. A scored keyword starts at 1. Expected CTR adds up to 4 points, landing page experience up to 3, ad relevance up to 2 — nine points on top of the floor, which is how a keyword reaches 10. As shares of those nine: expected CTR 44.4%, landing page 33.3%, ad relevance 22.2%.

The ordering matters more than the decimals. Expected CTR is worth twice what ad relevance is worth, so a keyword whose ad nobody clicks cannot be rescued by rewriting the ad to mention the keyword more often. The two move together — better copy usually lifts both — but only one of them is worth 4 points.

Landing page experience: the third you own outright

Google defines this part in a single line — how relevant and useful your landing page is to someone who clicks your ad — and then, on a separate page, names four ways to improve it: give people what they are looking for, keep the messaging consistent from ad to landing page, make the site mobile friendly, improve loading speed. Read that list again. Three of the four are not advertising decisions at all. They are how the website is built.

Two of the three parts of Quality Score are sentences in your account; this one is a property of your website. Expected CTR and ad relevance are worth 6 of the 9 points between them, and both are words somebody can rewrite this afternoon. The remaining 3 are earned by the page itself — its speed, its first screen, its one obvious next step — and no amount of rewriting an ad will produce them. That is why this part takes longest, and why it is the part most accounts never get to.

It is also the only part that keeps paying after the campaign stops. Money spent on expected CTR buys cheaper clicks for exactly as long as you keep buying clicks. Money spent on the page improves what every visitor sees — the ones from organic search, the ones who typed the address, the ones who clicked a link in an email — because it is the same page, judged on the same things. We measured that across forty agency websites, and the spread was wider than this industry admits.

We should say plainly that this is the part we build for a living, so read the emphasis as interested. The arithmetic behind it is still Google’s: a Landing Page built for the campaign it serves is usually cheaper than retrofitting one built for something else, and it is the only one of the three parts you can have finished before the first click is paid for.

How we worked out the weighting

One of our own accounts handed us an unusually clean experiment, by accident. Landing page experience is rated below average on all 19 scored keywords, which makes it a constant: it cannot explain why one keyword scores 7 and another scores 1. Every bit of the variation comes from the other two parts, and that makes them solvable.

Six distinct combinations of those two appear in the account, and each maps to exactly one Quality Score — above and above scores 7, average and above scores 5, below and above scores 3, below and average scores 2, below and below scores 1. Fit a linear weighting to that and one answer falls out: score = 1 + 2×(expected CTR) + 1×(ad relevance) + 1.5×(landing page), counting below average as 0, average as 1 and above average as 2. Total error across all 19 keywords: zero.

The landing-page weight is the one number here we did not measure. Because that rating never varies in that account, its weight is pinned only by the requirement that a keyword rated above average on all three can reach 10 — which forces 1.5 per level, or 3 points. That is arithmetic, not observation, and it is the figure to treat with the most suspicion. Whether average on the landing page is worth 1 of those 3 points or 2 is invisible in the data.

One corroboration is worth a line: the 22.2% we get for ad relevance is the figure third-party analyses have quoted for years. We arrived at it from a different direction, which is mildly reassuring rather than conclusive.

What a high Quality Score actually buys you

Three things, in descending order of how much anyone should care. Cheaper clicks for the same position, because ad quality is one of the inputs to Ad Rank and a higher-quality ad can hold ground a lower-quality one would have to outbid it to take. Eligibility, because some formats and assets only appear above certain Ad Rank thresholds, so poor quality removes options quietly rather than announcing it. And the diagnostic itself.

The diagnostic is the reason to look at the score at all. A 3 with expected CTR below average and ad relevance above it is not a grade — it is an instruction. It says the ad is written well and nobody wants what it offers, which is a keyword problem, and no amount of rewriting will fix it. The number without its three components is nearly useless; the components without the number are nearly sufficient.

Why good SEO is what makes ads cheap

A third of Quality Score is your landing page, and the landing page is not an advertising asset. It is the same page organic search judges, on the same things: how fast it renders, whether its copy answers the query, whether the next step is obvious. Work done for SEO is therefore work done for Quality Score — and it is the only part of the score you can improve before spending a single click.

Google spells the mechanism out on its own Ad Rank page. Ad Rank is built from six things, and one of them is the quality of your ads and landing page — Google’s own phrasing, which puts the page inside the auction’s inputs rather than beside them. The same page states the consequence just as plainly: higher quality ads can often lead to lower CPCs, and you pay less per click when your ads are higher quality. Put the two together and the chain is short. A better page earns a better landing page experience, which earns better ad quality, which earns better Ad Rank, which costs you less money for the same click.

With the landing page rated below average, the highest score any keyword can reach is 7. One point of floor, four for a perfect expected CTR, two for perfect ad relevance, and nothing at all for the page. One of our own accounts is the proof: its best keyword scores exactly 7, with expected CTR and ad relevance both above average. Three points were unavailable no matter how the ads were written.

That is the argument for doing the SEO work first, or at least alongside. Ads buy position inside a ceiling your site sets. A campaign pointed at a slow, thin page is not a campaign with a landing-page problem to fix later — it is a campaign paying a surcharge on every click for as long as it runs. It is also why our SEO Audit exists as something you can buy before any advertising: the findings are the same findings, whichever channel ends up paying for them.

The overlap runs the other way too. Expected CTR is judged partly on how well your listing answers the query, which is the same discipline that earns an organic click. Sites that do well organically tend to do well on expected CTR, because the work behind both is identical: know the query, answer it plainly, do not make people hunt. Which is the whole argument in one line: SEO is not an alternative to running ads. It is the discount on them.

How to move each part, in the order that pays

Start with expected CTR, because it is worth 4 points and because the cheapest fix is subtraction. Pull the search-terms report and add negative keywords for every query you never wanted; each one removes impressions that were never going to convert into a click. Then split ad groups until one theme sits in each, so the ad is allowed to name the query.

The landing page comes second on the list and first in effect, worth 3 points, because it is the slowest to change and the one fix that outlives the campaign. Make it fast, make the first screen repeat the words in the ad, give it one obvious action — then check all three on a phone rather than on the machine it was built on.

Ad relevance last, worth 2 and the easiest of the three — put the keyword in a headline, keep one theme per group, stop there. It is also the part most accounts start with, because it is the part that looks like work.

Then expect the score to lag. It is built from history, so a fix appears as new history accumulates, not the next morning. If you would rather not watch it week by week, that is what Ads Administration is for.

The trap: it is a diagnostic, not a target

Google says this plainly and it still needs repeating: Quality Score is a diagnostic, not a key performance indicator. An account can raise its average Quality Score by pausing every low-volume keyword and be no better off, because the average Google reports is unweighted — a keyword with 2 impressions counts the same as one with 2,000.

We know because an automated notice quoted one of our own accounts back at us and offered a feature as the fix. Read today, that account averages 3.0 unweighted, 3.68 weighted by impressions, and 4.63 weighted by cost. All three are arithmetically correct, and only the last two describe anything the auction or a customer experienced. Fix the components on the keywords that spend money, and let the average land where it lands.

Sources and method

The weighting in this article was fitted to the 19 scored keywords in one of our own Google Ads Search accounts, read through the Google Ads API on 8 September 2026. Everything else is Google’s own documentation.

• About Quality Score for Search campaigns ↗

  Google’s own page: the three components, the below/average/above ratings, and the statement that Quality Score is a diagnostic rather than a number used in the auction.

• Quality Score: Definition ↗

  The 1–10 scale, and that it is estimated from the historical performance of the keyword rather than measured live.

• About Ad Rank ↗

  Where ad quality actually enters the auction, and why position and price are recalculated per search instead of read off a stored score.

• Best practices for responsive search ads ↗

  Google’s guidance on the ad-side inputs sitting behind expected CTR and ad relevance.

• 5 ways to use Quality Score to improve your performance ↗

  Where Google names its four landing-page recommendations: give people what they are looking for, keep messaging consistent from ad to landing page, make the site mobile friendly, improve loading speed.

• Web Vitals ↗

  The speed metrics landing page experience is judged on — the same ones organic search uses, which is why the work counts twice.