Google Ads Quality Score: How to Improve It and Measure the Effect, Step by Step

Ad rank comes from your bid and your quality. With a better score you pay less for the same position, and in industries where clicks are expensive, improving quality is one of the few ways to bring cost per click down without shrinking reach. Google rates each keyword from 1 to 10 and builds that rating from three components: expected click-through rate, ad relevance, and landing page experience. Each one comes back as above average, average, or below average.

Treat the 1-to-10 number as a diagnostic. The auction calculates quality in real time for every impression, so the score tells you where to work, and turning it into a dollar figure has no real basis. It's also calculated on searches that exactly match the keyword, which means switching match types won't move it.

All you need is the Google Ads UI with the Quality Score columns enabled, a spreadsheet, and a browser. Someone following these steps by hand gets the same result. The recording above is one run on a real store's account, with the industry and names masked.

1. Follow the money before the score

Sort keywords by Quality Score and the top of the list fills up with keywords that barely spend. A keyword with a score of 2 and twelve dollars of spend last quarter doesn't deserve an hour of your time. A keyword with a score of 5 and a few thousand in spend does. Rank by the spend of keywords with at least one below-average component, then total that spend per ad group. You fix things at the ad group level, since each group has its own ad and its own landing page.

Keep only enabled keywords in enabled ad groups and campaigns with traffic in the past 90 days. Inside each group, look at keywords carrying at least a few percent of its spend; a weak score on a keyword that costs pennies doesn't justify a change. Work in batches, say twenty groups at a time from the top. The ranking gives you an order of work and can't promise savings, because Google doesn't publish how the components are weighted.

Before going group by group, look for account-wide patterns. If the same headlines run in ads across three or more ad groups, fix the template once for all of them. If below-average landing page ratings cover most of the spend across several different pages, the cause is usually the whole site rather than one page (more on that in step 5).

A dash where the score should be means the keyword hasn't had enough searches that exactly match it. Google hasn't rated it yet, so there's nothing to diagnose.

2. Find the component that's dragging the score down

Every component has its own cause and its own fix, so start by checking which of the three is below average. Google compares your ad with the advertisers who showed for exactly the same search in the past 90 days. That's how a single page can rate above average for a product name and below average for a broad category keyword: on broad searches, competitors have better-matched pages.

Comparisons inside the account help. When the same keyword has a better component rating in another ad group, whatever differs between the two groups, the page or the ad, usually points to the cause. Sometimes the ad is doing fine and the score still sits low. With an 11% CTR and 80% of impressions at the top of the page, expected CTR comes back average and the overall score stalls at 5. In that case the landing page is the bottleneck.

When two components are weak, start with the one tied to more spend. If they carry about the same weight, the order is fixed: ad relevance first, since the fix is quick and entirely in your hands, then landing page experience, then expected CTR.

3. Ad relevance: the keyword in headline one

Ad relevance is the quickest component to improve, and it usually comes down to whether headline one contains the keyword you're bidding on. The clearest warning sign is a keyword that doesn't appear in any headline of the ad actually getting the traffic. In a responsive search ad, pin the headline with the keyword to position one. On the accounts we manage, a keyword in an unpinned headline often performs almost as well, but pinning guarantees it shows on every impression. Pinning an existing generic headline isn't enough, though. Replace it with one that carries the keyword.

One fixed headline can't cover a group with several keywords. In that case, make headline one keyword insertion, written as {KeyWord:Default text}, and pin it to position one. The default text appears when the keyword runs past 30 characters, so a keyword longer than that needs its own headline or a shorter form. You can pin both versions, the fixed one and the insertion, to position one and let Google compare them. Insertion usually wins. Leave headlines two through fifteen unpinned. Free rotation is what makes them useful, and an ad with all three positions pinned behaves like an old expanded text ad with three headlines. Add the keyword's words to the display path as well.

Fix the ad that's getting the traffic. A well-written ad holding a few percent of the group's spend won't lift the score, and adding a third ad next to it won't change anything. Also look for headlines with the keyword in square brackets. Brackets display literally, and they were usually meant to be keyword insertion.

Skip keyword insertion on keywords that contain a competitor's name, or that name ends up in your headline. Skip it in brand groups too. Brand searches typed as a domain or misspelled naturally get lower relevance, and a fixed brand headline pinned to position one works better there. Low scores on competitor keywords are expected, and whether to keep bidding on them is a business call.

New ads take a day or two to get reviewed, so a change made on Friday afternoon may not go live until after the weekend.

4. The ad group: one topic, one page

If ad relevance stays below average with the keyword already pinned in headline one, the usual cause is a mixed ad group, where keywords on several topics share one ad. Split the group by topic and give each new group its own ad and its own landing page. A good rule of thumb for the whole account: as many ad groups as landing pages.

Search terms show you where to split. When a large share of clicks comes from searches with a word that isn't in any of the group's keywords (a material, a model, a product type) and the site has a page for that variant, move it into its own group with that page. Keywords you move keep their match types and bids. When a broad category keyword pulls in searches that break down into subcategories with their own pages, split the group by subcategory and leave the broad keyword where it is. Check first whether another ad group already catches those searches. If one does, you don't need a new group.

5. Landing page experience: a better URL first, then a brief

The landing page is the hardest component to move, because it takes work on the website. Google looks at whether the page gives people what they searched for: whether the topic matches the search, whether the ad's promises are backed up, whether the mobile version is easy to use, and whether the page loads fast. The keyword doesn't have to appear word for word, but putting it in the page title, the H1, and the copy is the simplest way to make the topic obvious. No headline will rescue a page built around a different topic.

Before you ask anyone to change the site, check whether a better page already exists. Below-average landing page ratings show up far more often when the URL lacks the keyword's words, and on short broad keywords of one or two words sent to a single product page. A category page with the keyword in its URL, title, and H1 often solves the problem with a URL change in the ad and no developer time. The homepage is the culprit less often than people assume, since brand and broad searches tend to rate fine there. It becomes a problem when the keyword is about something the homepage doesn't show at all. In local campaigns that send traffic to a store location page, keep that URL. Switching it to the online store would cut off the campaign's goal.

When no better page exists, write a brief for whoever runs the website. Read the page first, because a brief built on guesses won't help anyone. List specific gaps: the keyword missing from the title and H1, ad promises with nothing to back them up on the page, outdated content (an expired promotion, an old brand name), title tags and meta descriptions duplicated across pages, and slow loading on mobile (measure it with PageSpeed Insights or Lighthouse in the browser). When one weak page serves several ad groups, a single brief lifts all of them, so that page goes to the front of the line.

Watch out for pages that load their content with JavaScript. A crawler or the page source will show only an empty shell, while Google runs the scripts and sees more. An empty shell just means you don't know yet what's on the page. Open it in a browser, and use the URL Inspection tool in Search Console to see what Google sees.

When weak landing page ratings cover most of the account's spend across several pages, the cause is usually sitewide: speed, the mobile version, the page template, or thin category pages. One brief for the whole site covers it.

6. Expected CTR: give people a reason to click

Expected CTR estimates how likely people are to click your ad compared with competitors on the same search. Better ads raise it. A bigger budget only buys reach. Weak CTR most often shows up in ads with fewer than eleven headlines or with no numbers in any headline.

Fill the ad out to fifteen headlines and add specifics the landing page confirms: a discount, the number of models, the number of reviews. Use price and delivery time only when they genuinely set the offer apart. A fact in a headline should come from the seller, never from customer reviews, and should hold for everything the group sells. Everything the ad promises has to appear on the landing page, or a CTR gain turns into a landing page loss. When a group runs two ads with at least a few hundred impressions each and the weaker one's CTR trails by more than a quarter, pause it or move what works in the stronger ad into it.

7. What to leave alone

Ad strength in the Google Ads UI is a separate metric with no link to Quality Score. Google's recommendations often push you to unpin headlines because that raises ad strength. Unpinning headline one also lowers ad relevance, which is the rating that drives what you actually pay.

Negative keywords save money, but they don't raise a keyword's score, because Google only rates searches that exactly match the keyword. That's work for your regular account review.

Be careful in campaigns with a running experiment or with AI Max turned on. An ad edit in an experiment carries over to both arms and spoils the comparison. AI Max can rewrite text and swap the landing page URL, so a pinned headline or a new URL won't always show, so read any results with caution. In regular Search campaigns, we turn AI Max off.

8. Log every change

Without a log there's nothing to measure. For each fix, record the date and time, the ad groups, the component it targeted, and exactly what changed. Log each website brief with the date you sent it and the date the change went live. That second date starts the measurement clock. Note other account changes from the same period, too: new campaigns, new conversion goals, new assets. They'll explain part of the difference later.

9. Measure in score history and cost per click

Google stores Quality Score history day by day, component by component, which is what makes the effect measurable. Run the first check after about fourteen days and the second after thirty. Before that, the score hasn't had time to update. Read the fourteen-day result as an early signal. Components look back 90 days, so a landing page rating that hasn't moved yet doesn't mean the fix failed.

Compare the component the change targeted: ad relevance and CTR after an ad fix, landing page experience after a URL or page change. For "before," take the last day with a score before the change. For "now," take the keyword's current score. The most recent entry in the history may be days old, since days without impressions aren't recorded. A keyword with no score in the measurement window simply had too few searches, and that's different from "no change."

Then compare cost per click, CTR, and impression share across two windows of equal length, one before the change and one after. Impression share under 10% in both windows is Google's reporting floor, so that metric can't show you anything here. "The score went up and the cost stayed flat" is a legitimate result, and you should log it that way.

A lower cost per click proves nothing on its own, because the whole account may have gotten cheaper over the same weeks. Measure every change against a control group: weak ad groups in the same campaign that nobody touched. The effect of your fix is the gap between the changed groups and the control. Keep seasonality in mind as well. At peak season the set of competitors changes, while the score still looks back 90 days.

In the recording, pinning the keyword in headline one lifted ad relevance the very next day, and Quality Score on the two main keywords went from 3 to 5. The landing page rating didn't move, as expected, since the change touched only the ads. Cost per click fell 16.1% in the changed groups and 11.5% in the control group, so about 4.6 percentage points can be attributed to the fix. The rest reflects a drop across the whole account.

The rules in short

• Rank by the spend of keywords with a weak component, totaled per ad group. Enabled items with traffic in the past 90 days only.
• Look for account-wide patterns first: fix a shared ad template once, and cover weak pages across the site with one brief.
• A dash instead of a score means missing data. Leave those keywords alone.
• Find the weak component first. At equal weight: relevance, then landing page, then CTR.
• Relevance: the keyword or {KeyWord:Default text} pinned to position one, other headlines free, the keyword in the display path. Fix the ad that gets the traffic. No keyword insertion for competitor or brand keywords.
• Relevance still weak with the keyword in headline one: split the group by topic. As many ad groups as landing pages.
• Landing page: look for a better existing URL first, then write a brief covering the keyword in title, H1, and copy, backing for every promise, fresh content, and mobile speed. Read the page before writing, and check an empty shell in a browser.
• CTR: fifteen headlines and specifics the page confirms. Budget doesn't raise CTR.
• Ignore ad strength. Negative keywords don't change a keyword's score. Experiments and AI Max call for caution.
• Every change and every brief goes into the log with its date and component.
• Measure after 14 and 30 days: the targeted component from score history, then cost per click and impression share in equal windows, always against a control group from the same campaign.