Quality Score and What It Really Costs You

Quality Score is the most misunderstood number in the interface. It is treated as a grade to be maximised, blamed for costs it does not directly set, and quoted in agency reports as though it were a result. It is none of those things. It is a diagnostic, and once you read it as one it becomes genuinely useful.

What the number is

Quality Score is a one-to-ten estimate that Google reports against each keyword in a Search campaign. It does not exist for display, video, Demand Gen or Performance Max.

It is built from three components, and each is reported not as a number but as one of three statuses — below average, average or above average:

• Expected click-through rate. How likely your ad is to be clicked when it shows for that keyword, with position and format effects stripped out.
• Ad relevance. How closely your ad’s message matches the intent behind the search.
• Landing page experience. How relevant and useful the page you send people to is once they arrive.

Three details change how you should read those statuses. They are relative — you are being compared with other advertisers who competed for the same keyword, so “below average” describes your position in a field, not an absolute failure. They are based on roughly the previous ninety days of history. And the underlying history is drawn from exact matches of that keyword rather than every query it ever served, which is why a keyword’s score can look disconnected from the traffic it is actually bringing you.

A dash rather than a number means the keyword has not had enough impressions yet. Nothing is wrong.

Getting the columns on screen

The score is hidden by default. On the Keywords page, open the columns editor, find the quality score group, and add: Quality Score, Exp. CTR, Ad relevance and Landing page exp. There are historical versions of each as well, which show the score as it stood in a past period rather than today’s value. Those are the ones to use if you want to see whether a change you made moved anything.

Then add Cost to the same view and sort by it descending. This is the whole trick, and the next section explains why.

The mistake almost everyone makes

Averaging Quality Score across an account and reporting the average is meaningless. Every keyword counts equally in that average, including the four hundred that have never spent a pound.

What matters is the quality of the keywords that take your money. Sort by cost, look at the top twenty rows, and read the three components on those. A 4/10 on a keyword spending a third of your budget is a live problem. A 4/10 on a keyword that gets nine impressions a month is a rounding error, and hours spent improving it are hours wasted.

The second mistake is treating a low score as automatically bad. Some keywords are structurally difficult:

• Competitor terms score poorly on ad relevance and landing page experience almost by definition. Your page is not about their brand. That is not fixable and it is not a reason to stop, if the maths works.
• Very broad head terms attract clicks from people with a dozen different intentions, so expected click-through rate suffers.
• Brand terms score highly for reasons that have nothing to do with your skill. People searching your company name click your ad. Feeling good about a 10/10 on your own brand keyword is feeling good about the weather.

A keyword that converts profitably at 5/10 is worth more than one that converts nothing at 9/10. Profit is the target. Quality Score is a hint about the route.

How it relates to what you pay

Here is the nuance that gets flattened in most explanations. The one-to-ten number is not itself the input to the auction. Ad Rank is calculated fresh at auction time from your bid, the quality of your ad and page in that specific context, the ad rank thresholds, the competitiveness of the auction, the search context — device, location, time, the nature of the query — and the expected impact of your assets.

So the number in your keyword table does not directly set your cost per click. What it does is summarise the same underlying signals that the auction is judging. Improving genuine relevance tends to mean you can win the same visibility for a lower bid, or better visibility for the same bid. Google does not publish an exchange rate, and anybody who quotes you one for how much a point of Quality Score saves has made it up.

What actually moves each component

Expected click-through rate. This one responds to ad copy that matches the search and to ad group structure. If a single ad group holds twenty loosely related keywords, no ad in it can be specific to all of them, and this component is the first casualty. Split by intent, not by volume. Note that pausing a bad keyword does not repair the component — the history stays attached.

Ad relevance. The cheapest to fix. Get the keyword’s language into a headline. Write ads for the theme of the ad group rather than for the business generally. Use the search terms your customers actually type, which you will find in the search terms report, not the ones your marketing deck prefers.

Landing page experience. The slowest to fix and usually the most valuable. Google describes this in terms of relevant and original content, transparency about who you are and what you do, easy navigation, and pages that load quickly on mobile. In practice the biggest single improvement in most accounts is sending traffic to a page about the thing that was searched for, instead of the homepage. That is a website job as much as an ads job.

One thing worth stating plainly: none of these three components can be improved by fiddling with bids, budgets or bid strategies. They respond to relevance and nothing else.

When to look at it at all

Quality Score earns attention in two situations. When your costs per click have risen and you want to know whether the cause is competitive pressure or your own deteriorating relevance. And when you are restructuring an account and want a before-and-after read on whether tighter ad groups delivered anything.

Outside those cases, checking it monthly is a habit rather than a decision. It is a rear-view mirror. The forward-looking numbers are cost per enquiry and the quality of the enquiries you get, and those are the two an honest report leads with.

If your report from an agency leads with Quality Score and impression share and never quite gets to what a customer cost, that is worth asking about — and it is the sort of thing an independent paid search review is meant to surface.