Quality Score is Google's keyword level diagnostic for Search campaigns, rated from 1 to 10. The displayed score is not an input in the ad auction.
Use its three component statuses to investigate the experience behind the keywords that matter commercially. Then judge the result by useful customers and acquisition economics, not by whether the number moved.
What is Google Ads Quality Score?
Google's Quality Score page describes it as a diagnostic tool. It's meant to give a sense of how well your ad quality compares with other advertisers. Google shows it at the keyword level for Search campaigns.
Quality Score has three components: expected clickthrough rate, ad relevance and landing page experience. Each is rated above average, average or below average. The yardstick is advertisers whose ads showed for the exact same searches over the last 90 days.
The same page says it is not an input in the ad auction. It is not a key performance indicator. It should not be optimised or aggregated with the rest of your data.
Four Google terms get collapsed into one grade. But they're four different things.
Concept | What it is | Used in the auction | Use it for |
|---|---|---|---|
Quality Score | Keyword level Search diagnostic from 1 to 10 | No | Finding which part of the experience to investigate |
Ad Rank | Auction time values deciding eligibility and placement | It is the auction mechanism | Nothing directly, there is no visible 1 to 10 grade |
Ad Strength | Feedback on a responsive search ad's supplied assets | No | Judging whether the asset mix is complete and varied |
Optimisation score | Estimate of account potential from open recommendations | No | Reading Google's recommendation list, not profit |
Google defines the other three on their own pages: Ad Rank, Ad Strength and optimisation score. Ad Strength isn't used to calculate Ad Rank, Quality Score or auction wins. Its own page says so plainly. The optimisation score runs from 0 to 100% and changes when recommendations are applied or dismissed.
How ad quality affects auctions and costs
Quality Score is a diagnostic rating for Search keywords, not an input in the ad auction. Google uses auction time ad quality alongside bids, competition, context, thresholds and asset impact to determine Ad Rank.
Improving the underlying experience can help performance. But Google publishes no discount by score. A higher displayed score doesn't guarantee a particular cost per click reduction.
Use the displayed Quality Score as a multiplier for the bid
Use the score to investigate experience. Recognise that Ad Rank uses auction time quality and other factors.
Google's Ad Rank page says Ad Rank is calculated every time someone searches. The inputs are:
the bid
the auction time quality of the ad and landing page
the Ad Rank thresholds
the competitiveness of the auction
the context of the search, such as location, device and time
the expected impact of assets and other formats
The "bid multiplied by Quality Score" formula describes an old simplification. No arrow runs from the displayed number into the auction.
Google's ad quality page does say higher ad quality generally leads to better performance. The page lists better positions and lower cost among the benefits. That is a reason to improve the experience behind a weak component.
It is not a claim that moving a 5 to an 8 buys a predictable discount. Auction conditions and your own bidding change at the same time. A real quality improvement can produce better results without reducing the observed average cost per click at all.
Device, user location and time of day shape the auction. The visible diagnostic doesn't fully capture them. That's why a keyword scoring 9 can still have a quality problem in one segment. A 4 can be the most profitable thing in the account.
Click costs and the tactics that bring them down are covered separately. See our guide to Google Ads cost per click and how to lower it.
Find and interpret your Quality Score data
In Search keyword reporting, add the Quality Score column and the three component columns. When you are investigating a change, add their historical versions too: Quality Score (hist.), Exp. CTR (hist.), Ad relevance (hist.) and Landing page exper. (hist.).
Google's page says the historical columns show past values for the reporting period. It also says segmenting the table by day shows the change in daily scores. The plain columns show the current status. That means changing the date range doesn't turn them into a period average.
So segment by day and find the date the component moved. Put that date beside the ad, page and account edits in the change history. Do that before you believe any explanation.
Three reporting quirks save wasted effort.
First, the components compare the same exact searches over the previous 90 days. That is the comparison window. It doesn't promise that a fix needs 90 days to show, or that it will refresh overnight.
Second, a dash means too few searches have exactly matched the keyword to calculate a score. It does not mean zero quality. Google's page suggests broadening the match type to enter more auctions. Nobody should buy traffic just to obtain a number.
Third, the score is based on historical impressions for exact searches of the keyword. Changing the match type does not change the diagnostic. A broad match keyword with no impressions of its own can display a score. That happens when the exact match keyword in the same ad group supplies the history. In that case both report the same score.
That explains an apparent inconsistency between a keyword's score and its own impression count. It is not a reason to duplicate every keyword.
Improve the experience behind each component
Each component is a clue about one part of the experience between the search and the page. The method is the same for all three.
Expected clickthrough rate
The component predicts how likely the ad is to be clicked for that search. It is not the observed clicks divided by impressions in the account. Read the two side by side rather than treating one as the other.
Google's page of five ways to use Quality Score says to make the offer more compelling. It also says to be more specific and to use clear calls to action. But it adds the caveat that matters commercially. A more specific ad can lead to a lower clickthrough rate and a higher conversion rate.
A more qualifying ad that turns away the wrong clicks is often the right trade.
For example, a Melbourne electrician's vague "electrical services" ad answers a search for commercial fit out work. The hypothesis is that the ad does not speak to commercial intent. The test is an ad with accurate commercial service detail, never an emergency promise the business can't keep.
Judge it on enquiries that are actually commercial work, not on clicks. Our guide on how to write Google Ads that get the right clicks covers the craft.
Ad relevance
The component asks whether the ad matches the intent of the search. Google's advice is to group keywords into themes and match the ad's language to the searches. That's about honesty, not mechanics. Can one ad serve every keyword in the group without misleading anyone?
Separate a theme when the message needs to change, not every time the wording changes. Do not fall for one keyword per ad group, or for repeating every keyword in every headline.
For example, an Adelaide accountant's "business accountant adelaide" and "personal tax return adelaide" keywords share one ad. The hypothesis is that business advisory and personal tax searchers need different messages. Confirm the mismatch in the search terms and the ad before restructuring anything.
If the search terms confirm it, split the group. Our guide to Google Ads account structure shows you how. Searches that never suited the business don't need a new ad group at all. They need Google Ads negative keywords.
Landing page experience
The component asks whether the page is relevant and useful to the person who clicked. Google's page says to give people what they're looking for. Keep the message consistent from ad to page. Make navigation easy, make the site work on mobile and improve loading speed.
It also says two things readers misremember. The word for word search phrase does not need to appear on the page. The conversion rate does not affect the landing page status. Even so, it is a good way to measure. Our guide to what a good conversion rate is shows what to compare it with.
The same page still links Google's Mobile Friendly Test. Google Search Central's notice marks it as retired in December 2023. Test on the devices that carry the traffic and use current performance tools instead.
Start with the keyword that carries the most spend. Our audit sample is 180 Google Ads account audits from August 2021 to September 2026. We measured spend over the 90 days before each audit. In about 1 in 3 audited accounts, the top spending keyword showed a below average landing page experience. When it does, check where the click lands before you test the page's speed.
For example, an online retailer sends "hiking boots" searches to the general store homepage. The hypothesis is that the stocked hiking boot category page serves the intent better. Verify the stock, the filters and the buying route before switching. An empty category is worse than a homepage.
Our guide to Google Ads landing page vs homepage covers which page should get the click. If a speed test shows a real loading problem, start with why your website is slow.
Reading the components together
The components also combine. A combination is a hypothesis to investigate rather than a proven cause.
Observed combination | First hypothesis to investigate, not a proven cause |
|---|---|
Weak expected CTR and relevance | Check whether the ads answer the actual commercial intent and whether themes are mixed |
Strong expected CTR and weak page experience | Check the promise, the actual destination, usefulness and usability, not speed alone |
Strong relevance and weak expected CTR | Inspect offer clarity and useful differentiation without chasing clickbait |
All components average | There may still be an improvement worth testing, since average is not fault free |
All components weak | Confirm the intended search audience fits the business before rewriting every asset |
Assets affect Ad Rank through their expected impact without becoming a fourth component. Our guide to Google Ads assets covers their role.
Which low scores deserve your attention?
Google supplies no universal good score and no target average. Google's page on using Quality Score to guide optimisations calls it a helpful diagnostic tool, not a KPI. It says to focus your effort on high value areas where you can change something.
Our rough bands are an optional Elev8d triage habit, not Google benchmarks. No band is an automatic action. On a keyword that carries spend or valuable work:
a score of 1 to 3 gets looked at this week
a 4 or 5 gets its components checked against the search terms
a 6 or 7 gets attention only where one component is below average
an 8 to 10 gets no score work while we watch its outcomes
The matrix below decides the order.
Situation | What it suggests | Action |
|---|---|---|
High spend, a weak component | The costliest experience in the account has a visible clue | Investigate first |
Low volume, high value service, weak component | Few searches, but each one is worth a lot | Investigate, do not deprioritise by volume |
Low score, profitable outcomes | The diagnostic disagrees with the bank | Monitor, never pause for the rating alone |
High score, poor outcomes | Something the score does not see is failing | Investigate targeting, intent, page and auction segments |
Brand keywords scoring high | Brand searches score well almost by definition | Do not use as the standard for generic terms |
Competitor keywords scoring low | Expected, and a separate legal and commercial question | Decide on economics and advice, not on the score |
When the larger verified failure is somewhere else, it goes first:
tracking that records nothing
a broken enquiry route
a wrong promise
a campaign that is not serving
Our guide for when Google Ads is not working sets out that order, check by check.
In about 1 in 5 accounts we audited, at least one paused keyword was doing its job. Its cost per enquiry or sale was acceptable. The only recorded reason for pausing it was a low Quality Score. That's an expensive way to misread a diagnostic.
A high score doesn't prove a keyword pays its way either. Here is the arithmetic that keeps the score in its place.
Cost per lead is the cost per click divided by the enquiry rate as a decimal. So $8 over 0.01 is $800 and $18 over 0.12 is $150. The lower scoring keyword has the lower media cost per lead. That is all the example shows.
Neither keyword is proven profitable, because qualification, close rate, customer value and delivery costs are missing. There is no calculated saving anywhere in it from changing a Quality Score.
Whether $150 is good depends on your margins. Our guide to a good cost per lead for Google Ads shows how to work that out. For planning ranges by trade and profession, see Google Ads cost per lead by industry. The missing profit inputs are in our guide to Google Ads ROI. It also covers how to turn them into a return.
Run a focused review and measure the result
Pick a few commercially important keywords, not the whole account. For each one:
read the components and the actual queries behind the keyword
open the ad and the real destination the click reaches
write down the hypothesis
make one proportionate change
set the evaluation window from the sales cycle and the conversion lag
Then compare equivalent periods on the outcomes that matter. That means useful leads or customers, the acquisition cost and relevant auction metrics such as impression share. Concurrent targeting, bid and market changes limit what the comparison can prove. Use a controlled comparison where the volume allows one.
Keyword commercial role and period: [why this keyword matters and the dates]
Component clue and actual query evidence: [the statuses and the searches behind them]
Ad or destination hypothesis: [what the evidence suggests needs work]
Change owner and other concurrent changes: [who changed what, and what else moved]
Useful outcome and evaluation window: [the outcome measured and the time allowed]
Result uncertainty and next action: [what moved, what could explain it, what happens next]
Three rules apply when you read the result. A higher score with worse qualified outcomes is not success. A better business result with an unchanged score is still worth having. And no 15 minute screening establishes a cause. So the record carries the uncertainty rather than a verdict.
Which numbers belong in your regular report is a bigger question. Our guide to Google Ads metrics that matter covers it. If someone else runs the account, what you pay should cover a focused review like this. The monthly Google Ads management page shows what ours includes.
We'll look at the keywords carrying your spend, their components and where the clicks land. Then we'll tell you what we'd fix first.
What we recommend at Elev8d
Read Quality Score last, not first. Before we open a single component, we confirm the account records real enquiries. If yours doesn't, set up Google Ads conversion tracking before you touch anything else in this guide. Without tracking, you can't tell a better ad from a worse one.
Then we work down the keywords that carry the spend. On each one we look at the destination, then the ad, then the grouping. A specific search sent to a general page is the fault we rule out first. That's because no headline can fix it. Regrouping comes last, because it's the most disruptive change and the easiest to get wrong.
We leave the score alone on a keyword that pays for itself. We also don't hold generic terms to a brand term's score. And we never broaden a match type just to turn a dash into a number. That's buying traffic to feed a diagnostic.
Every change gets judged on qualified enquiries and what they cost. If the score rises and the enquiries don't, nothing is fixed. If the enquiries improve and the score sits still, we take the win.
FAQ: questions that need a little more context
How long does a change take to show in the score?
Google publishes no refresh deadline. A change shows as impressions accumulate under the new ad or page. So a keyword with few exact searches can take longer to move. Watch it in the historical columns segmented by day. In the meantime, judge the change on outcomes.
How do I judge quality in Performance Max or Shopping?
Without a score. There is no Quality Score for a Shopping product or a Performance Max asset group. Inventing one by analogy leads to the wrong decisions. Auction time quality still applies to those ads. Judge them on the searches they match, the pages they send people to and their qualified outcomes.
Can our own tools calculate Quality Score?
No tool outside Google can. Our audit scorecard asks whether your ads and landing pages line up. Our copy grader checks that ad text is complete and within its limits. Both are useful. Neither can calculate Google's Quality Score, its predicted clickthrough rate or an auction outcome.
Next steps
Add the columns and sort by cost
Add the Quality Score and component columns to the Search keywords that carry the spend. Include their historical versions too. Cost first, score second.
Form one hypothesis per keyword
Read the components with the actual queries and open the real destination. Write down what the evidence suggests before you change anything.
Measure on customers, not on the number
Make one change and set the window by the sales cycle. Read the result on qualified outcomes and acquisition cost, with the uncertainty recorded.
