Google Ads Quality Score: What Matters and How to Improve It

Google Ads Quality Score is a keyword-level diagnostic for Search campaigns. It estimates, on a 1–10 scale, how expected CTR, ad relevance, and landing page experience compare with other advertisers for the same search. It is not a KPI and it is not an auction input. Use the component statuses to choose a repair, then judge that repair with CTR, CVR, CPA, lead quality, and CRM evidence. A higher score does not guarantee lower CPC, better position, leads, revenue, or profit.

What Quality Score actually is

Google’s Quality Score documentation is the controlling source: the score is available at keyword level for Search campaigns, uses a 1–10 scale, and is built from three component evaluations. A dash means Google lacks enough searches that exactly match the keyword to calculate the score. Do not turn a dash into zero, and do not manufacture a diagnosis from a tiny sample.

The useful question is not “How do I get 10/10?” It is “Which observable mismatch is expensive enough and important enough to repair?” A score with little spend and no strategic value can wait. A commercially important keyword with meaningful spend and a below-average component deserves investigation even when it still converts.

Fact	Official boundary	Operational decision	Unsafe claim
Keyword diagnostic	Search keyword reporting	Read components before editing	Account average is the KPI
Visible 1–10 estimate	Use only where Google reports it	Prioritize by business impact	Every keyword needs 10/10
Three components	Expected CTR, ad relevance, landing page experience	Assign one repair owner	Conversion data is in the formula
Dash	Insufficient exact-match search data	Wait or use other evidence	Dash means failure

The three components: expected CTR, ad relevance and landing page experience

Google’s component-led improvement guidance separates three jobs. Expected CTR asks whether the ad is likely to earn a click for the search. Ad relevance asks whether the ad answers the search intent. Landing page experience asks whether the destination is useful after the click. Each can be above average, average, or below average.

Read the statuses together. Below-average expected CTR with above-average relevance points toward offer framing or copy specificity. Below-average relevance with acceptable CTR often means curiosity is earning clicks while the ad group mixes intents. Below-average landing page experience with strong ad signals means the promise may break after the click. Mixed failures usually require structure and page work, not another headline variant.

Component	Status to inspect	Diagnostic question	Repair owner	Measurement limit
Expected CTR	Below average	Does the ad earn the right click for this intent?	Paid media copy owner	More clicks can mean worse leads
Ad relevance	Below average	Does one ad answer this query cluster?	Campaign structure owner	Keyword repetition is not intent match
Landing page experience	Below average	Can the visitor complete the promised task?	Paid media and page owner	Speed alone is not usefulness
Mixed failure	Multiple weak components	Is the query-to-page chain structurally wrong?	Channel lead	Multiple simultaneous edits blur learning
Dash / no score	Not enough exact-match search data	Is there enough evidence to diagnose?	Paid media analyst	A dash is not zero or failure
Low score, good result	Weak diagnostic with viable economics	Can the component improve without harming qualified demand?	Channel director	Do not pause a profitable keyword from score alone
High score, poor result	Strong diagnostic with weak economics	Is intent, offer, tracking, or margin the real problem?	Channel director	A high score does not prove business value

Research on CTR prediction, semantic relevance in sponsored search, and ad-text strength can explain why prediction and semantic fit matter in advertising systems generally. None of these papers documents Google’s current product formula, so they stay background context rather than repair authority.

What Quality Score is not

Quality Score is not Ad Rank, not a KPI, not a target to average across an account, and not proof that a keyword is profitable. Google explicitly says the score is not an auction input. Match-type changes do not directly change it because the diagnostic is based on historical impressions for exact searches of the keyword.

Not true	Safer wording	Why it matters
Quality Score is Ad Rank	Quality Score diagnoses; Ad Rank determines eligibility and placement	Prevents false auction explanations
Quality Score is the KPI	Qualified demand and economics remain the objective	Keeps reporting tied to the business
Changing match type fixes the score	Repair query, ad, and page alignment	Avoids structural busywork
Google uses conversions in the score	Use conversions to judge the repair, not define the score	Avoids an invented formula

How to find and segment Quality Score columns

Google’s column instructions list the current score, current component statuses, historical score, and historical component statuses. Add all eight fields at keyword level, segment by day when useful, then join the diagnosis to search terms, ads, landing pages, and business outcomes.

Step	Column or segment	Decision	Stop condition
1	Quality Score and historical score	Find visible, persistent issues	Dash or negligible data
2	Expected CTR and history	Isolate appeal or qualification risk	No meaningful impressions
3	Ad relevance and history	Isolate intent and structure mismatch	No stable query cluster
4	Landing page experience and history	Isolate post-click mismatch	Page changed without a log
5	Day segment and change history	Connect movement to a dated edit	Several changes share the same date
6	Search terms, ads, page, CRM	Build the full evidence chain	Tracking or consent gaps invalidate comparison

A one-pass practical Quality Score audit

Use this pass only after the column set is complete. It turns a score review into assigned work instead of a list of low numbers.

Columns to add	Filter condition	Component diagnosis	First repair	Owner	Change log	Metric to watch	Do not claim
Score + all current/historical components	Meaningful cost and impressions	Below-average expected CTR	Rewrite one ad around the dominant intent	Paid media copy owner	Date, keyword cluster, ad ID, old status	CTR, CVR, CPA, lead quality	Copy caused a CPC reduction
Same column set	Commercially relevant query cluster	Below-average ad relevance	Split only incompatible answer shapes	Campaign structure owner	Old/new group, queries, ads, page	Relevance, query fit, qualified conversion	Splitting guarantees improvement
Same column set	Page receives meaningful paid traffic	Below-average landing page experience	Repair promise, proof, form path, or destination	Page owner	URL, changed block, evidence, date	Page status, CVR, qualification	The page caused score movement
Same column set	Several components weak	Mixed structural failure	Sequence structure, copy, then page	Channel director	One change window per repair	Component history and business metrics	Score movement proves causality

How to prioritize low-score keywords

Do not sort by score alone. Start with keywords that combine meaningful cost, a clearly weak component, commercial relevance, and a fix you can execute cleanly. Use this editorial triage formula: repair_priority_score = monthly_cost × component_severity × business_relevance × fix_confidence. This is not Google’s Quality Score formula.

Input	Definition	Use it to	Do not use it to
monthly_cost	Recent cost in the chosen review window	Size economic exposure	Predict future spend
component_severity	Internal weight for average or below-average status	Put clearer failures first	Convert Google labels into a platform formula
business_relevance	Internal value of the query and offer	Protect strategic demand	Promise revenue
fix_confidence	Evidence that one repair addresses one mismatch	Choose executable work	Claim causality

Pause only when the keyword is a bad business fit, violates policy or claim boundaries, or spends without defensible value after enough evidence. A low score alone is not a pause rule. A profitable keyword may remain active while its weak component is repaired; a high-score keyword may still be commercially poor.

Expected CTR fixes without clickbait

Inspect search-term patterns before rewriting. Separate urgency, location, price, comparison, alternative, branded, product-specific, and research intent. Make the offer, scope, location, availability, or next step explicit. Avoid fake scarcity, unsupported superlatives, and broad curiosity hooks that win clicks from people the business cannot serve.

Google’s improvement guidance notes that more specific ads can sometimes receive lower CTR but higher conversion rates. That is the correct business boundary: an expected-CTR repair must be read beside CVR, CPA, qualified lead rate, and CRM notes.

Intent	Weak ad	Stronger ad	What to avoid	Measurement limit
Emergency	Fast help for everyone	24/7 boiler repair in Leeds; check service area	Unavailable response promises	Calls still require qualification
Price	Best prices guaranteed	See what the quote includes before booking	Unverified lowest-price claims	CTR does not prove margin
Comparison	We are number one	Compare implementation, support, and contract fit	Unsupported superiority	Engagement does not prove purchase
B2B research	Transform your business	See the finance approval workflow for multi-site teams	Vague transformation claims	Form volume can hide role mismatch

Ad relevance fixes by query intent

Use the search terms report to cluster tasks, not merely words. Split when emergency and comparison traffic need different response times, proof, and pages; when local and national searches have different service boundaries; or when branded support traffic is mixed with new-customer acquisition. Do not split when the same ad and page can answer the entire cluster without becoming generic.

Google advises grouping keywords into themes and separating ad groups that cannot be addressed by one ad. That is not permission to create one keyword per ad group. The practical unit is one coherent answer shape: one audience situation, one promise, one evidence set, and one destination.

Query intent	Split when	Do not split when	Ad answer shape	Landing page need
Emergency	Response and service-area proof differ	Same coverage and response path	Availability and eligibility	Emergency service page
Local	Locations have different coverage or offers	One verified area page answers all	Area, route, next step	Location page
Price	Quote logic differs from product purchase	One transparent pricing flow fits	Cost scope and exclusions	Pricing or quote page
Comparison/alternative	Decision criteria differ from buying intent	Same proof resolves both	Trade-offs and fit	Comparison page
Branded	Support, login, and sales are mixed	All branded queries seek purchase	Correct brand task	Task-specific brand page
Product-specific	Use cases require different proof	Same SKU and evidence apply	Product outcome and constraints	Product or collection page
Research	Education intent overwhelms commercial copy	Commercial next step is clear	Explain and qualify	Guide with relevant CTA

Landing page experience fixes beyond speed

Start above the fold: can a visitor confirm the offer, scope, location, eligibility, and next step without decoding a generic hero? Then inspect mobile usability, form path, navigation, load behavior, trust evidence, claim safety, and destination specificity. A homepage is acceptable only when it genuinely answers the intent; otherwise use a service, product, collection, comparison, or qualification page.

Google’s landing-page guidance emphasizes usefulness, message consistency, mobile friendliness, and speed. It also says conversion rate can be used as a proxy for optimization but does not affect the landing-page status. That distinction prevents teams from claiming a conversion change proves how Google calculated the component.

Landing page issue	Repair	Evidence needed	Metric after repair	Do not claim
Ad promise absent above the fold	State offer, scope, proof, and next step	Ad-to-page review	Task completion and CVR	A fixed score increase
Homepage used for specific intent	Choose or build the matching destination	Query cluster and page inventory	Qualified conversion path	Homepages always fail
Mobile form friction	Remove nonessential fields or stage qualification	Device QA and form analytics	Completion and lead quality	Speed solved everything
Weak trust evidence	Add approved proof near the decision	Evidence owner and policy review	Qualified contacts	Invented testimonials or results
Unsafe claim	Rewrite, substantiate, or remove	Legal and policy evidence	Approval and business outcome	Guaranteed result
Offer is unavailable	Stop paid traffic until promise is true	Operations confirmation	No spend while blocked	A page edit can fix fulfilment

Quality Score, ad quality, Ad Rank and CPC

Google’s Ad Rank definition says eligibility and placement use multiple factors: bid, ad and landing-page quality, thresholds, competition, search context, and expected asset impact. Google’s ad-quality guidance explains that quality can affect eligibility, position, CPC, and whether assets show. Quality Score remains the visible diagnostic, not the auction calculation.

Concept	Official boundary	Client-safe explanation	Prohibited promise
Quality Score	Keyword diagnostic, not auction input	Use it to locate repair work	The score prices every click
Ad quality	Broader auction-time quality signals	Better quality may improve auction outcomes	Quality guarantees position
Ad Rank	Bid, quality, thresholds, competition, context, assets	Eligibility and placement are auction-specific	Quality Score equals Ad Rank
CPC	Auction outcome with multiple inputs	A quality repair may help but must be measured	A score increase guarantees lower CPC

Performance Max, AI Max and automation boundaries

Quality Score is a Search keyword diagnostic. Do not invent the same workflow for Performance Max or AI Max for Search. For automation, inspect the controls and evidence the campaign type actually exposes: assets, search themes, landing pages, conversion goals, exclusions, change history, and available reporting.

Campaign type	Use this diagnosis	Do not import	Keep audit-only
Search keywords	Quality Score and component history	Account-average target	Causal CPC explanation
Performance Max	Assets, search themes, pages, goals, exclusions	Keyword Quality Score equivalent	Unpublished automation formula
AI Max for Search	Controls, themes, pages, queries, outcomes	A fictional AI score	Visibility or performance guarantee
API or Scripts	Documented fields and change logs	A field that documentation does not expose	Private benchmark without evidence

Measurement workflow without fake certainty

Use the interface or documented reporting paths such as the Google Ads API reporting guide and Google Ads Scripts overview when technical automation is justified. These sources describe access and automation; they do not prove that an edit will improve performance.

Workflow step	Record	Business check	Decision boundary
Baseline	Score, components, query cluster, spend, impressions	CVR, CPA, lead quality	Tracking must be trustworthy
Repair	One ad, structure, or page change and owner	Implementation QA	Avoid simultaneous unrelated edits
Observation	Current and historical component movement	CTR, CVR, CPA, CRM notes	Wait for meaningful data
Review	Keep, revert, expand, or investigate	Qualified pipeline or margin where available	Score movement is not causality

Weak setup versus stronger setup

Business	Weak setup	Stronger setup	Why stronger	Evidence needed	Do not claim
Service	Emergency, price, and comparison terms in one ad group to a homepage	Intent groups with service-area, quote-scope, and comparison pages	Each task receives a matching promise and path	Search terms, service rules, page QA	More leads are guaranteed
Ecommerce	Generic product ad sends every use case to a broad category	Use-case ads point to verified product or collection pages	Availability, fit, and proof stay continuous	Feed, stock, returns, page proof	CVR must rise
B2B/SaaS	Feature slogan sends every role to the homepage	Role-specific workflow copy points to proof and qualification	The buyer can validate fit before the form	CRM role data and approved proof	Pipeline will increase

Three concrete ad rewrites

Query	Weak ad	Stronger ad	Destination	Evidence needed	Do not claim
waterproof women’s hiking jacket	Best Outdoor Jackets — Shop Now	Women’s Waterproof Hiking Jackets — Compare Fit, Shell and Stock	Filtered collection with fit and availability	Feed, size guide, waterproof-spec approval	Best, fully waterproof, or higher CVR without proof
field service CRM alternative	The #1 CRM for Every Team	Field-Service CRM Alternative — Compare Dispatch, Mobile Notes and Setup	Role-specific comparison page	Approved feature matrix and implementation scope	Number-one status or guaranteed adoption
emergency plumber near me	Fast Plumber — Call Today	Emergency Plumber in Bristol — Check Coverage and Callout Availability	Service-area emergency page	Live coverage, hours, callout terms	Arrival time or availability not confirmed by operations

Detailed operational example

Assume a regional repair company has meaningful spend on three clusters: emergency callout, price estimate, and competitor comparison. Expected CTR is average, ad relevance is below average, and landing page experience is below average for the emergency cluster. One ad group uses “Trusted Repairs at Great Prices” and sends every query to the homepage. This is a hypothetical operating example, not a case study.

The repair plan creates three intent groups. Emergency copy states service area, availability check, and call path; the page confirms eligibility and what happens next. Price copy explains quote scope and exclusions; the destination shows the quote process. Comparison copy explains decision criteria without unsupported superiority; the page shows approved evidence and fit boundaries. The team changes structure first, logs the date and IDs, then changes the page in a separate observation window.

Stage	Exact action	Owner	Measure	Do not conclude
Diagnosis	Export components and cluster search terms	Paid media lead	Spend, impressions, statuses	Low score caused high CPC
Structure repair	Create three coherent intent groups	Paid media lead	Ad relevance, search-term fit	A split guarantees improvement
Copy repair	Write scope-specific claims and CTAs	Copy and offer owner	Expected CTR, CVR, lead quality	Higher CTR means better business
Page repair	Match promise, proof, form path, and service area	Page owner	Landing-page status and qualified conversion	Page speed alone caused movement
Review	Compare history and CRM notes after enough data	Channel director	CPA and qualified pipeline	Score movement proves causality

What to remove, rewrite or keep audit-only

Claim or asset	Remove	Rewrite	Keep audit-only
Quality Score guarantees lower CPC	Yes
Quality Score is the KPI		It is a diagnostic for choosing repairs
Low-score keyword must be paused		Check business value, spend, components, and fixability
Performance Max has keyword Quality Score	Yes	Use campaign-specific diagnostics
Anonymous before/after account result		Use only with verified first-party evidence	Keep private until approved
Academic model explains Google’s formula	Yes	Research is background context only	Model interpretation

Common Quality Score mistakes

• Averaging the score into an account objective.
• Editing bids before reading component statuses.
• Stuffing keywords into headlines instead of separating incompatible intent.
• Treating speed as the whole landing-page experience.
• Changing ads, structure, pages, and bids on the same day without a log.
• Pausing a valuable keyword because the visible number looks low.
• Reporting CPC, CPA, lead, ROAS, or revenue causality from score movement alone.

FAQ

Does Quality Score directly determine CPC?

No. Google’s Ad Rank definition lists bid, ad and landing-page quality, thresholds, competition, search context, and expected asset impact among the auction factors. Quality Score is a keyword diagnostic, not the auction calculation. Improving a weak component may improve ad quality, but a score increase does not guarantee a lower CPC, a particular position, or a cheaper conversion.

Should I chase 10/10 Quality Score?

No. Google’s Quality Score guidance says the score is not a KPI and should not be optimized or aggregated as an account objective. Prioritize commercially important keywords where a weak component reveals a repairable user-path problem. A perfect score on low-value demand is not the goal; qualified conversions, sustainable CPA, lead quality, and business value remain the decision criteria.

Does match type change Quality Score?

No, not directly. Google explains that Quality Score is based on historical impressions for searches that exactly match the keyword, so changing match type does not itself change the score. Use the search terms report to inspect intent, then repair the ad-group structure, ad answer, or landing page where the evidence points. Do not treat a match-type edit as a Quality Score shortcut.

Should every low-score keyword be paused?

No. Use Google’s component guidance to identify whether expected CTR, ad relevance, or landing page experience is weak, then combine that diagnosis with cost, query intent, conversions, and lead quality. Pause when the keyword is a poor business fit, creates policy risk, or remains economically indefensible after enough evidence. A low diagnostic score by itself is not a universal pause rule.

Does Performance Max use the same Quality Score workflow?

No. Quality Score is a keyword-level diagnostic for Search campaigns. Google’s Performance Max documentation describes a different campaign structure and set of controls. Review assets, search themes, landing pages, conversion goals, exclusions, change history, and available reporting instead. Do not invent a keyword Quality Score equivalent for Performance Max or use Search component statuses as if they were exposed there.

Can higher CTR hurt performance?

Yes. Google notes that a more specific ad can sometimes produce a lower CTR but a higher conversion rate. Broad, sensational, or discount-led copy may win extra clicks from people who are unlikely to buy or qualify. Evaluate CTR beside CVR, CPA, lead quality, margin, and CRM notes. The goal is the right click and a coherent customer path, not maximum click volume.

What I would fix first

• Add current and historical score and component columns.
• Filter to meaningful spend, impressions, and business relevance.
• Group keywords by the weak component and query intent.
• Choose one repair with a clear owner and evidence requirement.
• Log the change before editing anything else.
• Wait for enough data and compare component movement with business metrics.
• Remove or quarantine every unsupported performance claim.