How much a keyword's quality rating moves in a month

You meet it on a Monday. A keyword that read 6 out of 10 last week now reads 4, and nothing in the account changed: no new ad, no landing page edit, no bid change, no match type change.

The columns that show it are Quality Score and its three components, plus the historical variants, which is the pair of readings that starts the argument. Segmenting the table by day makes the drift visible rather than inferred.

Google documents one thing here that explains most of it. Each of the three components is evaluated as Above average, Average or Below average, and that evaluation is a comparison with other advertisers whose ads showed for the exact same search over the last 90 days.

That makes the score relative rather than absolute. Each component is graded against other advertisers who showed for the same search in the last 90 days, not against a fixed bar, so the reading rests on a comparison group the advertiser does not control. Google does not publish how or when the score is recalculated, and this page does not speculate about why any single keyword moved.

Google also lists factors related to ad quality that Quality Score does not capture at all, including the device used in the search, the location of the user, the time of day and assets. So the column is neither a complete picture of ad quality nor a fixed one.

The cost is the ritual. A two point move is the ordinary case here, not the exception, so a weekly review that treats any movement as a finding will generate an action almost every week, each one attached to a cause somebody invented after the fact.

That has a second, worse effect. Once a team believes last week's edit moved the score, it believes the score responds to edits, and the account starts being managed against a number that Google describes as a diagnostic, says is not a key performance indicator, and says should not be optimised or aggregated. The hours go to the column instead of to click through rate, conversion rate and cost per acquisition.

The score is not useless, it is just badly suited to a weekly cadence. A component that moves from Above average to Below average and stays there is worth knowing, because it says your ad or landing page now compares worse against whoever is showing for that search.

The rule that follows is about timing rather than attention. Read it as a periodic diagnostic, look at sustained moves in the components rather than single point changes in the total, and expect drift between any two readings.

It is not a performance metric, and Google says so directly. It is not Ad Rank, and it is not used at auction time to determine Ad Rank.

A score that moved is not proof that anything the advertiser did caused it to move, which is the inference this page exists to block. It is a relative measure, so a move can come from the comparison rather than from the account.

It is also not a claim that Quality Score is random. 64.5% of these keywords moved 2 or more points, but 22.1% did not move at all, and the average range was 2.22 points rather than the full scale.