The basic idea
Advertising platforms want people to find ads useful. If searchers click irrelevant ads and quickly leave, they lose trust in the results page, and advertisers waste money. To discourage that, platforms generally estimate how relevant and useful an ad is likely to be, and use the estimate when deciding which ads to show and what to charge.
Some platforms show a simplified version of this estimate to advertisers. Google Ads, for example, has long displayed a quality score for keywords on a scale of 1 to 10. Other platforms use different names and scales. In every case, the exact calculation is not public, it can change, and the number shown is a diagnostic aid rather than the actual auction input. Check the platform’s current documentation for how its version is defined.
What these ratings tend to reflect
While details differ, the ratings usually combine three broad ideas.
Expected click performance
How likely is it that someone searching this term will click your ad, compared with other ads for the same term? The platform estimates this from past performance, which is why new keywords start with little history to go on.
Ad relevance
How closely does the ad match the intent behind the keyword? An ad for “wedding photographer” on a keyword group about “wedding photographer prices” that never mentions prices may be less relevant than one that does. Tightly grouped keywords make it easier to write ads that fit.
Landing page experience
Does the page you send people to deliver what the ad promised, and is it a usable page? This tends to cover things like how well the content matches the search, how clear and trustworthy the page appears, and whether it loads quickly and works on mobile. Page speed and user experience matter here for the same reasons they matter for visitors in general.
Why it matters to you
A higher relevance rating can mean you pay less for a given position, or achieve a better position for the same bid, because the platform sees your ad as more likely to satisfy the searcher. The exact effect varies and is part of the ad auction, which also weighs your bid and other factors.
In practical terms, relevance is a lever that does not depend on spending more. That is attractive for a small advertiser competing against bigger budgets.
What the number does not tell you
- It is not a measure of profit. A keyword with a high rating can still lose money if conversions are rare or your margins are thin.
- It is not the whole auction. The displayed score is a summary, and the real calculation is more detailed and made fresh for every search.
- It can lag. New keywords may show little information, and ratings can take time to move after you make changes.
- It can be noisy on low volumes. A keyword with very few impressions gives the system little to judge.
Judge keywords primarily by cost per acquisition and conversion results. Use the rating as a clue about where a problem might be hiding.
Practical ways to improve relevance
- Tighten ad groups. Keep a small set of closely related keywords together, so every ad can speak directly to the searches that trigger it.
- Echo the search in the copy. Where natural, reflect the wording of the query in the headline. The same principle applies to search ad copy in general.
- Match the page to the ad. The headline and first screen of the landing page should confirm what the ad promised. This is sometimes called message match.
- Improve the page itself. Make it clear, quick and easy to use on a phone, with an obvious next step.
- Add negative keywords. Excluding irrelevant searches can raise the share of clicks that are useful, which helps the click signals the platform reads.
- Prune weak keywords. Terms that attract impressions but rarely draw clicks, or that bring traffic that never converts, can usually be paused.
- Test ad variants. Let alternative wordings run, and keep the ones that perform better on both clicks and conversions.
Avoid chasing the number
It is easy to over-optimize for the rating, for example by writing ads that attract clicks but oversell, or by stuffing a keyword into every line. That can improve a score while worsening real results, and misleading wording can break platform policies and damage trust. The sound order of priority is: be useful to the searcher, deliver what you promise, and then see what the data shows.
Improving relevance tends to help, but it does not guarantee lower costs, better positions or more sales. Results depend on your market, competitors, offer and platform changes.