Google Ads Quality Score: how it affects your costs and when to ignore it

Quality Score in Google Ads looks like a simple 1–10 grade on your keywords, but Google defines it as a narrow diagnostic, not a performance metric or auction input. If you are deciding what to fix in an account, you will get farther by understanding Ad Rank, costs, and search impression share than by chasing a higher Quality Score number.

What Quality Score is and where it lives

Google describes Quality Score as a diagnostic tool that gives you a sense of how your ad quality compares to other advertisers. It is reported on a 1–10 scale at the keyword level in search campaigns. Each keyword’s score is based on historical impressions for exact searches of that keyword over roughly the last 90 days, not on your match type choices.

The score is built from 3 components: expected clickthrough rate, ad relevance, and landing page experience. Each component is labeled Above average, Average, or Below average compared with other advertisers whose ads showed for the exact same search in that period. Google’s help content also notes that Quality Score does not reflect devices, user location, time of day, or assets such as sitelinks.

Because Quality Score is tied to exact-search history, changing a keyword from broad match to phrase or exact does not change the score by itself. The diagnostic will only move if the underlying behavior on those exact queries changes enough to shift expected CTR, ad relevance, or landing page experience.

Whether Quality Score is used in the ad auction

Google’s documentation is explicit that Quality Score is not an input in the ad auction. The help pages state that it is a diagnostic tool to identify how ads that show for certain keywords affect the user experience, and that it is not a key performance indicator and should not be optimized or aggregated with the rest of your data.

This means the visible 1–10 Quality Score number does not participate when Google calculates which ads show and what you pay. Optimizing for the number itself is different from optimizing for the real-time quality signals that affect performance. Treating Quality Score as the primary lever for cost control is a misconception, because the auction uses separate, auction-time measurements of quality.

What determines CPCs and how Ad Rank uses quality

Your cost per click and position are driven by Ad Rank, which Google defines as a value used to determine your ad position and whether your ads are eligible to show. Ad Rank combines your bid with auction-time measurements like expected CTR, ad relevance, and landing page experience, along with other factors. These quality evaluations are recalculated for every search and can differ by page position.

Google documents 6 Ad Rank factors: your bid, ad and landing page quality, Ad Rank thresholds, auction competitiveness, search context, and the expected impact of assets and formats. Search context includes the search terms, location, device, time, other results on the page, and user signals. The expected impact of assets and formats covers how extensions and formats might improve performance.

Per Google’s help content, higher quality ads can often lead to lower CPCs and better eligibility, because strong quality can help you meet or exceed Ad Rank thresholds and unlock more formats such as sitelinks. The key distinction is that the auction uses real-time versions of expected CTR, ad relevance, and landing page experience, while the visible Quality Score is a separate, offline diagnostic. The misconception that Quality Score directly determines CPCs confuses these two layers.

Why a well-run account can show mediocre Quality Scores

A low or middling Quality Score does not automatically mean an account is badly run. Google’s own description notes that the diagnostic omits several elements that affect performance, including devices, user location, time of day, and assets. An account that is carefully tuned by device, geography, schedule, and extensions can perform well on revenue and ROAS while still showing only Average component ratings in the Quality Score column.

Temper And Forge’s client work adds another structural example. HIPAA-compliant and PHI-related advertisers often cannot run traditional conversion tracking, which limits the feedback signals that can be used to evaluate and improve quality. In TF’s experience, these accounts can carry depressed Quality Scores even when they are tightly managed and profitable, because the diagnostic does not see the full picture of value.

Because Quality Score is computed from historical impressions on exact searches, it also lags behind current changes. A recently improved landing page or new messaging may not be reflected in the 1–10 number for some time, even if auction-time quality has already improved and CPCs are falling. This is one reason TF does not treat a low Quality Score as proof of poor management.

When Temper And Forge ignores Quality Score and what it watches instead

Google’s help pages state that Quality Score is not a key performance indicator and should not be optimized or aggregated. Temper And Forge’s practice aligns with that guidance. TF does not watch the Quality Score number closely and does not treat it as a guiding principle for optimization.

TF’s focus is on ROAS, CPC, deliverability, and search impression share as the spend modulator. If costs creep up or impression share is lost to rank, TF works through the auction-time quality levers that feed Ad Rank, such as ad relevance and landing page experience, and through bidding decisions. Chasing the visible 1–10 score as if it were a performance KPI would invert the priority, since Google explicitly warns against optimizing for it.

This approach also avoids the misconception that an average Quality Score across the account is a useful health metric. Google’s documentation warns against aggregating Quality Score, and there is no account-level score in the product. TF evaluates health through performance metrics and context, not through a rolled-up Quality Score.

When Quality Score is worth reading

Even though TF treats Quality Score as secondary, the diagnostic can still surface useful hints. Because it breaks into expected CTR, ad relevance, and landing page experience, it can point to which part of the user journey is weakest relative to other advertisers on the same queries. A Below average label on landing page experience, for example, is a clear signal that your page is underperforming peers on that traffic.

Quality Score is also helpful when you see delivery or cost problems and want a quick directional check. If search impression share lost to rank is high, and Quality Score components show Below average expected CTR and ad relevance, that supports a focus on improving ad quality and relevance alongside bids. Used this way, the diagnostic is a secondary pair of eyes that can confirm or challenge your working hypothesis about why Ad Rank is low.

What it does not provide is a numeric roadmap from score to CPC. Google publishes no formula, component weighting, or QS-to-CPC mapping, and explicitly warns against treating the score as a KPI. The misconception that optimizing the Quality Score number is equivalent to optimizing the account ignores this documented limitation.

What a “–” Quality Score means

Sometimes the Quality Score column shows a dash instead of a 1–10 value. Per Google’s help content, this means there are not enough searches exactly matching that keyword to determine a score. The system does not have sufficient historical impression data on exact searches to assign a diagnostic rating.

A dash does not mean the keyword is broken or that it is excluded from the auction. The auction still uses auction-time quality signals for any eligible impression. It simply means the offline diagnostic cannot be computed yet, so there is no Quality Score to interpret.

Temper And Forge’s practice is to treat Quality Score as a secondary pair of eyes: mostly Google noise with occasional useful insight. TF watches ROAS, CPC, deliverability, and search impression share, and when costs move, it works through auction-time quality levers and bidding rather than chasing the visible 1–10 Quality Score number.