What expected CTR measures
Expected CTR is a forecast, not a record of what happened. For each keyword, Google estimates how likely your ad is to be clicked if it shows for that keyword, and reports the estimate as below average, average or above average against other advertisers on the same term. It is one of the three components reported in Quality Score, alongside ad relevance and landing page experience.
The important detail is that the forecast is normalised. Google strips out the advantage of sitting high on the page and the lift that ad assets give, so an ad is not credited simply for occupying a good slot. What is left is a judgement about the wording: does this ad look, to someone who typed that query, like the answer they wanted? A keyword with little history borrows an estimate from similar terms until it has data of its own.
Why expected CTR matters
It feeds Ad Rank, so it changes both whether your ad shows and what each click costs. Two advertisers bidding the same amount on the same keyword will not pay the same price if one of them writes ads that answer the search and the other does not.
It is also the clearest read you have on whether your ad text is speaking the customer’s language. Conversion data tells you what happened after the click; this rating tells you what people thought of the ad before they clicked, which is a signal you cannot get anywhere else in the account.
Common mistakes with expected CTR
Chasing clicks for their own sake is the big one. A headline that overpromises will lift the rating and quietly ruin the cost per lead, because the extra clicks come from people who were never going to buy. A rating is only worth improving when the traffic it brings is traffic you want.
Comparing across keyword types is the second. Branded searches will nearly always rate better than generic ones, because the person already knows who you are. Judging a generic keyword against a branded one tells you nothing. The third is impatience: the rating is recalculated as impressions accumulate, so a low-traffic keyword takes time to reflect a rewritten ad.
How to act on it
Start with the ad group rather than the ad. A keyword rated below average is often sitting among keywords with a different intent, so the ads have to hedge and end up specific to nothing. Split the group so each set of ads can name the thing the searcher asked for.
Then use the search terms report to see the wording people actually typed, and borrow it. Add negatives for the queries you never want. Give a responsive search ad genuinely different headlines rather than paraphrases of one idea, so the system has something to test. Re-read the diagnostics after the new ads have gathered impressions, and judge the trend rather than a single day.
