Google Ads Quality Score: What It Is, How to Raise It

Google Ads Quality Score is a 1-10 diagnostic Google assigns to each keyword from the last 90 days of data — not a multiplier on your bid. Most guides describe it as a formula, but Google’s own help page says the opposite: “Quality Score is not an input in the ad auction.”

Every percentage in this piece was measured on 8 September 2026, across the 15 of my managed accounts with Search spend in the last 30 days, from 19,745 active keywords. No “weight” or “formula” here that Google hasn’t published. The order for raising it is the same in most accounts: landing page first (that’s where the problem sits for 89.1% of the keywords I measured), then ad group structure, then copy — details below.

What Quality Score is, and isn’t

Quality Score is a backward-looking measurement that summarises, on a 1-10 scale, how your ad quality compares to other advertisers shown for the same search over the last 90 days. What actually runs the auction is Ad Rank — a separate calculation built from your bid, your ad quality at that moment, and a handful of contextual factors. There is no “bid × Quality Score” multiplication anywhere in Google’s documentation; raising your bid can raise your rank, but it never changes your score — the score is calculated purely from the three components. A keyword showing 3 in the panel can still produce a perfectly good quality signal for one specific search — say, mobile, midday. The score is a 90-day summary; the auction is that search’s own account. Read the score like a thermometer: it measures the fever, it doesn’t cause it. The distinction: the score you see in the panel looks backward, the signal used in the auction looks at that moment.

Three more limits, and most guides skip all of them. The score is reported only for Search campaigns and only at the keyword level — Google’s own page title is literally “About Quality Score for Search campaigns.” There’s no such column in Performance Max asset groups, Shopping product groups, or Display placements. The window is 90 days, so a change you make today feeds in slowly. And if there weren’t enough exact-match searches in the last 90 days, you see ”—” instead of a score — I’ll show exactly how common that is in the panel section below.

The three components

Quality Score is made of three components: expected click-through rate (CTR), ad relevance, and landing page experience. Google labels each one “below average,” “average,” or “above average,” comparing you against other advertisers shown for the same search over the last 90 days.

Expected CTR is how likely your ad is to get clicked when someone searches that keyword. Ad relevance is how closely the keyword matches your ad copy. Landing page experience is the speed, mobile-friendliness, and promise-match of the page a click lands on. The three labels matter more than the number itself — Google Ads optimisation is where that shows up in practice.

The common misconceptions about these components are addressed on Google’s own page. The user’s device is taken into account, but the panel doesn’t split the score by device — testing the mobile experience separately is on you. Account structure doesn’t change the score unless it changes what the user actually experiences. Performance on the Display Network and search partners doesn’t affect your score on Google.com. And ranking high doesn’t raise expected CTR either — Google’s own words: a top position on the page doesn’t increase the expected CTR rating of your ads.

How to find it in the Google Ads panel, and what ”—” means

The columns are off by default, so most accounts have never turned them on — and once you do, ”—” is what shows for 95.6% of my 19,745 keywords. In the keywords view, open the column-picker icon top right, expand “Quality Score,” and add all eight columns by hand: the score, the three components, and each one’s “history.” This is the first thing I do in an account audit.

The “history” columns do separate work. The regular column shows today’s score. Google describes the historical columns as showing past Quality Score statistics for the reporting period you’re viewing, and once you segment the report by day you get the day-by-day change. That’s the only way to see whether a landing-page change or new ad copy actually moved the score.

In my 8 September 2026 measurement, only 866 of 19,745 active keywords (4.4%) show a score at all — 18,879 (95.6%) show ”—.” That’s not a bug. Google only calculates the score once there have been enough exact-match searches in the last 90 days, and 99% of the ”—” keywords got zero impressions in the last 30 days. So those keywords aren’t bad — they’re not being searched. In one hotel account, 628 of 637 active keywords are exactly this. Trying to fix a score-less keyword is wasted effort. Either wait, or clean it out if it’s dead weight bloating the account — most of an 18,000-keyword pile was added years ago on a “might get searched” guess and never revisited since.

How I raise Quality Score, component by component

If landing page is “below average,” I start there first, because the fix sits outside the account and takes weeks — relevance and expected CTR both wrap up the same day, inside the account. Here’s the order:

1. Narrow the table first — look only at keywords with a visible score that are also spending money.
2. If landing page is “below average,” fix that first.
3. Check ad relevance.
4. Check expected click-through rate.
5. If it’s still low, question the keyword itself.

Narrow the table first: look only at keywords with a visible score that are also spending money. In my 8 September measurement, 455 of the 866 scored keywords had spend in the last 30 days — a few dozen per account. The 18,000 score-less keywords sit outside this table entirely.

Start with the landing page. In the measurement, 89.1% of the 866 keywords are “below average” on this component, only 1.7% are above average, and 95.8% of spend goes to keywords that are “below average” here. That’s why it comes first. See which page a keyword goes to from the Final URL column in the panel. If there’s no page for that service, make one — the promise in the ad headline, one clear call to action, and three paragraphs explaining the service is enough. Checklist:

• Does the promise in the ad headline show up verbatim on the page? If the ad says “same-day delivery,” the first screen of the page should say it too.
• Does the user land on the page for that specific service or product, rather than the homepage?
• Does the page fall in the “good” Core Web Vitals band on mobile, per PageSpeed Insights? Let Google’s own tool set the threshold — the second-figures quoted across guides don’t agree with each other.
• Is there one clear action on the page — a form or a phone number visible on the first screen?
• Does the keyword from the ad appear in the page copy? Google doesn’t require an exact match, but the user should see what they searched for on the page.

If the page is slow, WordPress speed is a separate job. I cover the content and structure side under landing page optimisation.

Same day, inside the account: ad relevance. If the label is “below average,” the keyword and the ad copy don’t match. Narrow your ad groups. The rule in one sentence: keywords that share intent belong in one group; a keyword that can’t appear verbatim in a headline doesn’t belong in that group. An example from a florist e-commerce account in the measurement: if “gift flowers,” “boutique flowers,” and “order flowers online” sit in one group, three different intents don’t fit one headline. “Gift” and “order” go into separate groups; if “boutique” is a positioning word, it goes into a separate ad instead. In Responsive Search Ads, make sure the keyword appears verbatim in at least one headline. I cover the setup on responsive search ads; grouping itself is part of keyword optimisation. This is already the healthiest component in my portfolio — 58.3% above average — so if it’s fixed, move on. One trap: dynamic keyword insertion (DKI) auto-matches the relevance label to the keyword, but if you forget the page, the user still lands on a generic one — DKI matches the headline, not the page.

A note specific to non-English markets: Google treats close spelling variants of the same keyword as a single near-variant, not two separate signals. In this Turkish portfolio, that meant Google counted words like “ısıtma” and “isitma” (a diacritic-free typo of the same word) together. Adding both separately gains nothing — keep the correct spelling as one keyword. In my 8 September measurement, 15 accounts had 432 such near-duplicate pairs; in 89.6% the second spelling never got an impression (”—”), but in 10.4% both variants accumulated data and earned separate scores — rare, but it happens, and it means the data gets split across two keywords. Use the correct spelling in your ad copy too.

Expected CTR is also same-day. If relevance is “average” or “above” but expected CTR is “below average,” the copy is saying the right thing and users still aren’t clicking. Start with the search terms report: an ad shown for irrelevant searches that don’t get clicked drags the keyword’s expected CTR down. Negative keyword management is a score problem as much as a budget one for this reason. Then the copy: the headline should reflect what the user searched for, and the promise should be concrete — not “cheap flowers,” but “same-day delivery, at your door in two hours.” Ad assets (sitelinks, descriptions, price and promotion extensions) don’t feed the score directly — they feed the “expected impact of assets” factor in Ad Rank instead, since Google calculates expected CTR with the visibility boost from assets stripped out. Add them anyway; they raise your actual click-through rate.

Finally, question the keyword itself. If all three fixes are done and the score is still stuck at 1-2, the keyword is probably a weak fit for your business. Keep it if it converts, pause it if it doesn’t.

Whether a fix actually worked shows in the “Quality Score (history)” column, but the real measure is money. Compare the average cost per click and cost per conversion for the same keyword in the 30 days before and after the fix — if cost drops even while the score sits still, the fix worked.

Don’t rush a new account. The window is 90 days; in the first weeks the score you see is either ”—” or a rough estimate from thin data. Google doesn’t start new keywords from zero — it seeds them from the history of similar keywords already in the account. Changing structure in the first month based on a score that hasn’t settled yet is deciding on a forecast, not a measurement.

A low-scoring keyword that converts well

Don’t touch a low-scoring keyword that converts — Google’s own line that the score “shouldn’t be treated as a key performance indicator” was written for exactly this case. If cost per conversion sits inside your target, a 2-point keyword can be worth more than an 8-point one. Pausing that keyword to raise your score means closing something that makes you money just so the report looks tidier.

One condition: if landing page is “below average,” fix it anyway — the page fix lifts conversion rate independently of the score. Don’t touch copy that’s working for relevance; it’s already converting.

What should the target score be

Half (49.8%) of the 866 scored keywords sit in the 1-3 band, 28.9% in 4-6, 21.4% in 7-10; the most common single score is 3, and exactly one keyword scores 9. If your account shows 3s, that’s normal against this distribution. Weighting by spend changes the picture: the 1-3 band is half the keywords but only 24.5% of spend, the 4-6 band takes 44.8%. In 13 of 15 accounts, the share of keywords with a “below average” landing page component runs 77% to 100% — not account-specific, a repeating pattern.

There’s no single target score — Google never says “above X is good,” and the “7-10 good, 4-6 fair, 1-3 poor” bands you see in guides are the authors’ own line, not Google’s. Instead of a band, look at your account’s spend-weighted average and the type of keyword.

Why a weighted average: a plain average counts a zero-spend keyword the same as one eating half your budget. Google doesn’t score at the account level at all — there’s no such thing as a Quality Score at the ad group, campaign, or account level. Here’s how I calculated the weighted average in this piece: pulled the keywords report with the Quality Score and Cost columns, dropped the score-less rows, then in a spreadsheet: =SUMPRODUCT(score, cost) / SUM(cost). Across the 15 accounts, the plain average is 4.18, the weighted average 4.95. The gap comes from most of the 431 keywords in the 1-3 band spending very little.

Expectations differ by keyword type. Two accounts from the 8 September table: in one hotel account, only nine of 637 active keywords have a score at all, and all nine are the hotel’s own name and variants. The spend-weighted score is 9.71 — landing page and relevance are above average for all nine, expected CTR for seven. That’s not an achievement, it’s expected: the user is already searching for that specific hotel, clicks, and lands on the hotel’s own page. In the florist account above, the weighted score is 2.45, and 86% of keywords sit in the 1-3 band. The highest-spending keyword, “gift flowers,” scores 2. The biggest spenders in the 1-3 band are “boutique flowers” and “order flowers online,” both scoring 1. All three are “below average” on both landing page and expected CTR. For a florist, the target isn’t “get to 7” — it’s pulling the landing-page label on those three keywords from “below average” to “average.” The order in that account is clear: page first, then group structure — not done yet. Most long-tail keywords will show ”—” regardless; don’t set a target for those.

How the score actually affects cost per click

A high Quality Score is generally seen alongside a lower cost per click. Google’s own sentence goes this far: higher ad quality generally leads to better performance, including better ad positions and lower cost. Google says “generally,” not a percentage. The table showing −50% at a score of 10 and +400% at a score of 1 is from 2013, it’s WordStream’s, and Google doesn’t endorse it. The direction is right, but the table is 13 years old — not something to hand a client as “the percentage you’ll gain.”

My 8 September measurement found a more modest gap. Within each account, I split keywords into score bands and compared average cost per click; 10 accounts had at least 30 clicks in each of the two bands being compared. In seven, the higher-scoring band was clearly cheaper (ratio 0.74–0.97, a 3–26% gap); in two there was no real difference (1.00 and 1.02); in one it went the other way (2.07). No account came close to WordStream’s 50%.

A third account from the 8 September table, the one that went the other way: a business buying damaged cars had a core keyword scoring 6, over 1,500 clicks, close to half the account’s spend, and its cheapest click. The 7-scoring niche keywords (“kazalı araç alımı,” “motor arızalı araç alımı” — Turkish for “buying crashed cars” and “buying cars with engine damage”) were clearly pricier than the core keyword, by 4.1x and 2.9x respectively, and 17 of 18 keywords in that group are “below average” on landing page. The score is a within-account comparison; cost per click is that specific query’s own auction. This is a correlation, not causation — for the same reason, don’t compare across accounts either: put an 8-scoring keyword from the hotel account next to a 3-scoring one from a law firm and you’re measuring industry competitiveness, not quality. I cover what actually sets cost per click in a separate piece.

A 2026 note: broad match, Smart Bidding, and AI Max

On broad match, Quality Score measures only the searches that match the keyword exactly, not the dozens of search terms broad match brings in — so on a broad-match-heavy account, the score tells you less about the real quality of the traffic you’re getting. Google puts it this way: changing keyword match types doesn’t affect Quality Score, because the score depends on historical impressions for the keyword’s exact-match searches; and a broad match keyword can have a Quality Score even with zero impressions of its own, as long as a corresponding exact-match keyword had impressions in the last 90 days. The only way to read the quality of the search terms a broad-match keyword is pulling in is the search terms report. Google doesn’t say this explicitly for AI Max. What follows from the exact-match rule: your keywords and their scores still exist, but searches arriving through match-type-free expansion aren’t credited to any keyword’s score.

The score still matters under Smart Bidding, but not because “your score gets multiplied by the bid.” Google sets the bid, Google expands the keywords. What’s left in your hands is ad copy and the landing page — and the three components measure exactly those two things. The 1-10 scale and the three components haven’t changed in years.

If you’re curious about your own account’s distribution, I can pull the Quality Score columns and work out your spend-weighted average and the “below average” share for each component — the same table this piece is built on, for your account. I send the findings back in priority order, and the rest is up to you. Reach me through the contact page.

Frequently Asked Questions

Why does Quality Score show ”—”?

Because there weren’t enough exact-match searches for that keyword in the last 90 days. In my 8 September 2026 measurement, 95.6% of active keywords were in this state — see the “How to find it in the panel” section above for what to do about it.

How often does Quality Score update?

As of September 2026, Google doesn’t publish a fixed refresh schedule. The score is calculated over a rolling 90-day window and shifts as that window moves. To see the change, add the “Quality Score (history)” column and segment the report by day — comparing today’s score to what you remember from before isn’t a real measurement.

Does Quality Score exist for Performance Max or Shopping campaigns?

Not in the panel. Google reports Quality Score only for Search campaigns, at the keyword level — I cover the limits in the “what it isn’t” section above.

Is Quality Score the same thing as optimization score?

No. Quality Score is a 1-10 diagnostic at the keyword level. Optimization score is a 0-100% figure at the account and campaign level that measures how many of Google’s own recommendations you’ve applied. Google’s optimization score page says plainly that the score isn’t used by Quality Score. Taking your optimization score to 100% doesn’t raise your Quality Score.

What does “rarely shown due to low Quality Score” mean, and what should I do?

It means the keyword’s Quality Score has fallen below the level Google looks for to show it — the ad isn’t paused, it’s just rarely entering auctions. Check the three components first; the cause is usually a mismatch between the ad copy and the landing page. If the keyword still isn’t converting after a fix, pausing it is cheaper than forcing it with a higher bid.

How long does it take to raise the score?

Ad copy and ad group changes are usually visible within a few weeks; full reflection takes longer because the window is 90 days. For landing-page-driven drops, it depends on how long the page fix takes.

I made the fix — did it work? What do I check?

The “Quality Score (history)” column shows whether the score changed, but the real measure is money. Compare the average cost per click and cost per conversion for the same keyword in the 30 days before the fix against the 30 days after. If cost dropped even while the score held steady, the fix worked.