Google Ads Quality Score, Deconstructed: The Exact Math Behind Your CPC

September 29, 2026 · 15 min read

Most advertisers treat Quality Score as a report card. It is more useful to treat it as a pricing mechanism. A four-point swing in Quality Score can cut cost per click by a third on the same keyword, or hand the top position to a competitor bidding half your max CPC. This guide works through the official mechanics from Google Ads Help, the three component diagnostics, and the exact arithmetic step by step, so every optimization decision traces back to a number.

What Quality Score is, and what it is not

Quality Score is Google’s estimate of the quality of your ads, keywords, and landing pages, reported as a number from 1 to 10 for each keyword. Higher scores mean Google considers your ad and landing page more relevant and useful to someone searching that keyword. You can surface it in the Google Ads interface by adding the “Qual. Score” column to the Keywords view, alongside its three component columns: “Exp. CTR”, “Ad relevance”, and “Landing page exp.”

The reported Quality Score is a diagnostic summary shown in your account so you can spot weak keywords. It is not the number Google plugs into the auction. At auction time, Google calculates quality fresh for every query, using more granular signals than a single digit can hold. Per Google’s Ads Help documentation, the auction weighs your bid against auction-time quality (expected clickthrough rate, ad relevance, and landing page experience), plus Ad Rank thresholds, search context, and the expected impact of ad assets.

Two keywords can both report a QS 7 and behave differently in the auction, because the auction-time calculation sees detail the report collapses. Optimizing purely to move the reported number is aiming at the thermometer instead of the temperature; the components underneath are the actual levers.

Three more things Quality Score is not:

• It is not a bid input. Raising max CPC does not raise Quality Score; the components contain no bid variable.
• It is not account-wide. It is per keyword, although landing page assessments apply at page and domain level, and account history feeds the auction-time signals.
• It is not a goal in itself. A keyword at 7 with strong economics beats a keyword at 10 that never converts; profit is the scoreboard that counts.

Google evaluates each keyword’s recent performance and assigns one of three statuses per component: Above average, Average, or Below average, relative to other advertisers on similar keywords. A “Below average” flag names exactly which component drags the score down, and each flag has a different fix, which is how the rest of this guide is organized.

Component 1: Expected CTR

Expected CTR is Google’s prediction of how likely your ad is to get clicked when shown, normalized for ad position. An ad in position 1 naturally earns a higher raw CTR than one in position 4, so Google adjusts for that. The real question: for this position and query type, does this ad get clicked more or less often than expected?

The signal draws on historical CTR adjusted for position, plus auction-time context: a mobile query at 9 PM can carry a different expected CTR than the same query from a desktop at noon.

What drives it

• Message-to-query match. Ads that mirror the searcher’s language earn clicks. An ad headlined “Emergency Plumber, 60-Minute Response” will out-click a generic “Quality Plumbing Services” on the query “emergency plumber near me”, and Google notices the gap.
• Ad asset coverage. Sitelinks, callouts, structured snippets, call assets, and image assets expand the ad and give searchers more reasons to click. Expected asset impact is a formal Ad Rank input, and broader coverage consistently lifts CTR.
• Brand vs. non-brand intent. Brand keywords almost always show “Above average” because searchers click your ad. Non-brand competitive terms are where the flag turns “Average” or “Below average”, and that is where the work belongs.
• Match type dilution. Broad match showing one ad to loosely related queries mixes intents and depresses CTR; the component status will tell you about it.

How to diagnose it

Filter the Keywords view to “Below average” expected CTR and sort by cost descending: weak CTR on expensive keywords is a daily tax. Then segment CTR by device and hour to separate structural weakness (the ad itself) from contextual weakness (bids or scheduling), and check the RSA asset report for per-headline performance ratings.

The fix playbook

• Write to the query, not the business. Ad copy should read like an answer to the keywords in the group, with the core keyword or its closest paraphrase in at least two headlines.
• Fill all 15 headlines and 4 descriptions. Thin RSAs give Google’s system little to test.
• Pin with restraint. Pinning guarantees message control but shrinks testable combinations, which can depress CTR. Pin only what must be fixed, and test pinned vs. unpinned.
• Stack ad assets. Add 4-plus sitelinks, callouts, structured snippets, and call and image assets where relevant.
• Test hooks systematically. Rotate headline angles: speed (“Same-Day Service”), proof (“4.9 Stars, 2,300 Reviews”), offer (“Free Quote in 60 Seconds”), and specificity (“Licensed Plumbers in Austin”). Keep winners, replace losers monthly.
• Tighten match types where dilution is visible. The search terms report will show irrelevant queries; add negatives aggressively or split the theme into its own ad group.

Component 2: Ad Relevance

Ad relevance measures how closely your ad matches the intent behind the keyword. Google compares the keyword’s meaning against your ad copy and assigns the same three statuses. This is the most mechanical component and usually the fastest to fix, because it responds to structure rather than persuasion.

Keyword grouping: the real lever

Ad relevance is won or lost in keyword grouping. One ad group containing “divorce lawyer”, “child custody attorney”, “prenuptial agreement cost”, and “pro bono legal aid” cannot produce one ad relevant to all four. The fix is grouping by intent: each ad group holds keywords sharing one search intent, with its own responsive search ad written for that intent.

A note on SKAGs (single keyword ad groups), since legacy advice still circulates: in the era of expanded text ads, SKAGs were the standard prescription for perfect relevance. With responsive search ads as the only search ad format, Google’s guidance has moved to intent-themed ad groups of roughly 5 to 20 closely related keywords: tight enough that one ad serves the whole group, broad enough to accumulate meaningful data. For a lawyer PPC account, that means separate ad groups for “divorce”, “custody”, and “prenup” queries rather than one “family law” bucket.

Diagnosing weak ad relevance

Filter to “Below average” ad relevance and inspect at ad group level. The pattern is usually obvious: 30-plus keywords spanning multiple intents, or ad copy written for a different theme than half the keywords. Also check the search terms report: if the queries actually triggering your ads drift from the group’s theme, relevance suffers even when the keyword list looks clean.

The fix playbook

• Split by intent, not by volume. “emergency plumber”, “water heater installation”, and “drain cleaning” are three ad groups with dedicated copy. A plumber PPC campaign structured this way typically reaches “Above average” ad relevance within weeks.
• Put the theme in headlines and paths. The keyword or its core concept belongs in headlines and the display URL path (e.g., /emergency-plumber).
• Use ad customizers for scale. When dozens of ad groups share a template but differ by service or city, customizers insert the right detail per group.
• Prune stragglers. Keywords the group’s ad cannot serve well move to their own group or get paused.
• Mind dynamic keyword insertion. DKI can lift relevance mechanically, but on broad match it produces broken headlines. Use it on exact and phrase match only, with sensible default text.

Component 3: Landing Page Experience

Landing page experience evaluates the page people reach after clicking: relevance and originality of content, business transparency, and ease of navigation including load time. Unlike the other components, it is assessed at page and domain level, then surfaced per keyword, so two keywords pointing at the same page generally share its assessment.

One fix therefore lifts every keyword pointing at that page. It also means a domain with a history of thin or deceptive pages carries a handicap no ad copy tweak removes.

What Google actually measures

• Message match. Does the page deliver what the ad promised? A search for “dental implants cost” landing on a generic dentistry homepage is a relevance failure. The H1, opening copy, and offer should mirror the query and the ad: implant queries need implant pages with pricing information, not a homepage. Dental PPC campaigns live or die on this.
• Usefulness and originality. Substantive, specific content outperforms thin pages; boilerplate duplicated across fifty location pages reads as low value.
• Transparency and trust. Clear business name, address or service area, phone number, privacy policy, honest claims. Pages hiding who is behind them or making unsubstantiated promises (“#1 guaranteed”) score poorly.
• Navigability. Clear headings, a visible CTA, working forms, no broken elements. On mobile, which drives most clicks in local verticals, this is decisive.
• Speed. Load time feeds the assessment directly, and Core Web Vitals (Largest Contentful Paint, Interaction to Next Paint, Cumulative Layout Shift) are the practical yardstick. A page taking six seconds to become interactive will struggle to hold “Above average” regardless of copy.

The fix playbook

• Build dedicated pages per intent: the highest-ROI move in most accounts. Each major ad group gets a page whose H1 matches the query theme, whose copy answers the specific questions (cost, timeline, process), and whose CTA matches the intent.
• Fix speed before copy. Compress images, defer non-critical scripts, strip bloated elements, and aim for good Core Web Vitals on mobile, not just desktop.
• Answer the money questions above the fold: pricing ranges, service areas, response times, credentials. Visitors who bounce in eight seconds because the page opened with a mission statement are voting against your landing page experience.
• Add trust signals visibly. Reviews, licenses, guarantees, real team photos, and complete contact details in the footer.
• Keep the scent consistent: the ad’s promise, the page’s H1, and the page’s offer form one thread. If the ad says “Free Quote in 60 Seconds”, the page should say the same words and deliver a 60-second form.

The math: how Quality Score changes what you actually pay

Your position is decided by Ad Rank; your price is derived from the Ad Rank beneath you divided by your own quality. The two formulas below are the standard teaching model (PPC Hero’s formulation). One honest caveat: Google’s full Ad Rank also weighs thresholds, search context, and expected asset impact. The simplified version isolates the Quality Score effect, which is exactly what we want to study.

The second formula contains the entire business case: Quality Score sits in the denominator, so a larger score means a smaller price. Two advertisers can bid identically on the same keyword and pay very different prices purely on quality. The worked examples below are illustrative, using round numbers for transparent arithmetic.

Before the examples, note the account-level signals feeding this math in the background. The reported score is keyword-level, but auction-time quality draws on broader history:

• Domain and landing page history. A domain with a track record of relevant, fast, trustworthy pages starts each auction with a head start over a brand-new or historically thin domain.
• Account history. Long-running accounts with consistently relevant ads and good engagement carry positive signal into every auction. This is why established advertisers often find new keywords earning respectable scores faster than a new account would.
• Policy history. Repeated disapprovals erode the trust layer quality calculations rest on.

None of these appears as a column in the interface. They explain why identical campaigns in different accounts price differently, and why rebuilding an account to “reset” Quality Score backfires: you discard the positive history with the negative.

Worked example 1: QS 4 vs QS 8 at identical bids

Same keyword (“emergency plumber near me”), same max CPC bid of $5.00, three competing advertisers. The only difference is Quality Score. Watch what happens to the price each one pays.

Step 1: rank by Ad Rank. Apex leads at 40, Budget follows at 20, QuickFix trails at 15. Identical bids from Apex and Budget produce different positions, because 5.00 x 8 beats 5.00 x 4.

Step 2: compute what each pays. Apply the actual CPC formula, working top down. Each advertiser’s price is set by the Ad Rank directly below them.

• Apex Plumbing (position 1): (20 / 8) + 0.01 = 2.50 + 0.01 = $2.51
• Budget Drains (position 2): (15 / 4) + 0.01 = 3.75 + 0.01 = $3.76
• QuickFix Co (position 3): with nobody below, they pay near the minimum threshold (illustrative: assume roughly $1.50).

Step 3: compare. Apex pays $2.51 per click; Budget pays $3.76, a $1.25 gap. Apex pays 33.2% less (1.25 / 3.76) for identical traffic, while Budget pays 49.8% more (1.25 / 2.51) for a lower position. At 1,000 clicks a month, that is $2,510 vs $3,760: a $1,250 monthly penalty for four points of Quality Score, on one keyword.

Most accounts do not have one weak keyword; they have dozens, and each QS 4 or 5 against competitors at 7 or 8 is a daily surcharge that compounds. This is also why quality and impression share interact: Google estimates a first-page bid per keyword, and low quality raises that estimate. When the estimate exceeds your max CPC, the keyword stops showing, so weak quality shrinks eligible impressions as well as raising CPC. Fix the quality and impression share recovers without spending more on bids.

Worked example 2: outranking a higher bidder on quality alone

This is the scenario that surprises experienced advertisers most. A competitor bids nearly double your max CPC and still loses the top position, because quality outweighs bid in the Ad Rank multiplication. Illustrative numbers again, kept round for clarity.

Step 1: rank by Ad Rank. Cornerstone’s 4.50 x 7 = 31.5 beats Titan’s 8.00 x 3 = 24. Cornerstone bids 43.8% less yet takes position 1: 7 is more than double 3, and multiplication does not care which input is larger.

Step 2: compute actual CPCs.

• Cornerstone Law (position 1): (24 / 7) + 0.01 = 3.4286 + 0.01 = $3.44
• Titan Legal (position 2): (12 / 3) + 0.01 = 4.00 + 0.01 = $4.01

Step 3: read the result. Titan bids $8.00, ranks second, pays $4.01 per click. Cornerstone bids $4.50, ranks first, pays $3.44. The higher bidder pays about 17% more per click for a worse position: the auction is pricing Titan’s QS of 3 into every click.

Titan’s rational response is not a $10 bid: at QS 3 that lifts Ad Rank to 30, still below 31.5, while raising every click’s price. The rational response is a component audit, since a QS of 3 almost always means two or three “Below average” flags. Quality is the only auction input that improves position and lowers price simultaneously.

Five Quality Score myths worth dropping

Myth 1: The 1-to-10 score is the number used in the auction

The 1-to-10 Quality Score in your Keywords tab is a reported diagnostic, refreshed periodically. At auction time, Google computes quality fresh for each query with far more granular signals than a single digit. Two keywords showing QS 7 can therefore clear different Ad Ranks in the same auction. Treat the reported score as a triage tool: it tells you where to look, not what the auction saw. Optimize the components and the auction-time signals, and the reported number follows.

Myth 2: Raising your bid raises your Quality Score

Bid amount appears nowhere in the Quality Score components. Expected CTR, ad relevance, and landing page experience contain no bid variable. Advertisers who “fix” weak quality by bidding higher pay the weakness tax at a larger scale instead of removing it.

Myth 3: Deleting and re-adding a keyword resets its Quality Score

A re-added keyword briefly shows no score, then Google recalculates from current performance. The ad, the grouping, and the landing page have not changed, so the score lands back where it was, and you destroyed the history helping auction-time signals learn. The shortcut does not exist.

Myth 4: Quality Score only affects CPC, not position

The Ad Rank formula puts Quality Score on equal footing with bid. As worked example 2 showed, quality differences routinely decide who ranks first, not just who pays what. Ignoring quality while managing bids is like tuning half the engine. This also explains a common mystery: impression share lost to rank on keywords with competitive bids. The bids are fine; the quality is losing the Ad Rank multiplication.

Myth 5: Every keyword should be pushed to a perfect 10

Moving a keyword from 4 to 7 delivers the bulk of the CPC savings; moving from 8 to 10 buys a small incremental discount at a large effort cost. The professional standard: eliminate every “Below average” status on keywords with meaningful spend, and get important keywords to 7 or above. Beyond that, hours are better spent on offer testing, conversion rate, or expanding coverage.

The 30-day improvement playbook

Quality Score work succeeds as a four-week sprint with a fixed order: measure, restructure, rewrite, rebuild pages. Structural fixes move the needle more than copy tweaks, and copy needs clean structure to be measurable.

Week 1: Audit and prioritize

• Add the four quality columns (Qual. Score, Exp. CTR, Ad relevance, Landing page exp.) to the Keywords view and export 30 days with cost and conversions.
• Sort by cost descending and flag every meaningful spender with a “Below average” component. These are the only keywords that matter this month.
• Group flags by component: forty expensive keywords sharing “Below average” landing page experience on one URL is one page problem, not forty keyword problems.
• Record baseline CPCs and impression share lost to rank for the flagged set to compare in week 4.

Week 2: Fix ad relevance (structure)

• Split every ad group where one ad serves more than one intent. Target 5 to 20 tightly themed keywords per group.
• Write a dedicated responsive search ad for each new group, with the theme in at least two headlines and in the display path.
• Move straggler keywords to their own groups or pause them.
• Review the search terms report per group and add negatives for query themes the ad cannot serve.

Week 3: Fix expected CTR (copy and assets)

• Fill every RSA to 15 headlines and 4 descriptions. Remove “poor” rated assets from the asset report and replace them with new angles.
• Add the full asset stack: 4-plus sitelinks, callouts, structured snippets, and call and image assets where relevant. Verify they are actually serving.
• Launch one headline-angle test per major ad group (speed vs. proof vs. offer vs. specificity) and let it reach statistical relevance.
• Unpin headlines that do not need to be pinned, and test whether combination freedom lifts CTR.

Week 4: Fix landing page experience (pages)

• Take the top 5 landing pages by spend and verify message match: H1 mirrors the ad group’s theme, offer mirrors the ad’s promise, CTA mirrors the intent.
• Run PageSpeed Insights on mobile for each page. Fix the worst offender first: compress images, defer render-blocking scripts, strip bloated elements.
• Move the money content above the fold: pricing or pricing ranges, service area, response time, credentials, and a working form or click-to-call button.
• Add visible trust signals: reviews, licenses, guarantees, real team photos, complete contact details in the footer.
• Re-measure against the week 1 baseline. Quality Score updates as fresh data accumulates, so expect movement over the next two to three weeks, not overnight.

FAQ

Conclusion

Quality Score is three documented components, each with a defined fix, feeding an auction formula where quality sits in the denominator of your price. Long-term winners are rarely the biggest budgets; they are the advertisers whose ads match intent, whose copy earns the click, and whose pages deliver on the promise. Every point of quality is a permanent discount on every future click.

The 33% CPC gap in the first example and the outranked higher bidder in the second are the same mechanism operating in every auction, every day. The 30-day playbook above is the fastest honest path to capturing it.

For teams that would rather have specialists run the sprint, SCORSH manages Google Ads accounts where Quality Score improvement is part of the standard operating procedure: component audits, intent-based restructuring, RSA testing programs, and landing page rebuilds tied to paid traffic. If paid search is central to how the business gets customers, whether through lead generation funnels or direct response campaigns, the quality of the account is the quality of the economics. Talk to the SCORSH team about a Quality Score audit and find out exactly what each point is costing.

Rishabh

Rishabh is the founder of SCORSH, a performance marketing agency working with businesses across India and the US. With 14 years of experience, he writes about SEO, paid media, and the math behind growth.

LinkedIn | Facebook