CTR

CTR is a measure of agreement between two things: the offer in the creative, and the person it was shown to. That is why a low figure on its own diagnoses nothing. A compelling ad shown to the wrong audience and a weak ad shown to exactly the right one produce the same disappointing number for opposite reasons, and the fix for one makes the other worse.

Separating them takes one more piece of information. If the audience is tightly defined and known to be right, the creative is the variable. If the creative has performed elsewhere, the targeting is. When neither is known, testing creative first is usually the cheaper experiment, because creative can be changed in an afternoon and an audience takes a fortnight to learn.

On search the diagnosis is narrower and more useful, because the audience declared itself by typing the query. A low CTR on a search term means the ad did not look like the answer to that search, which is nearly always a relevance problem rather than a persuasion one.

Expected click-through rate is one of the three components of Google Quality Score, alongside ad relevance and landing page experience. Quality Score feeds Ad Rank, and Ad Rank decides both whether your ad appears and what the click costs.

The consequence is that CTR is not only a traffic metric on search. An ad with a better expected CTR can hold a position a competitor is bidding more for, and hold it more cheaply. Improving relevance is therefore the only lever that increases volume and decreases cost at the same time, which is why the unglamorous work of tightening ad groups so the ad matches the keyword keeps paying.

Note the word expected. Google is predicting the rate for that auction using historical performance for exact matches of the keyword, adjusted for context. A brand-new keyword carries an estimate rather than a measurement, so a poor component score on something with no impressions is not yet evidence of anything.

Social platforms decide what to serve by estimating the revenue each impression will produce, which is roughly the bid multiplied by predicted engagement. An advertiser with a modest bid and a creative people actually click can outcompete one with a high bid and a dull creative, because the platform earns more per impression from the first.

This is the mechanism behind the common observation that creative is the biggest lever on social performance. It is not only that a good ad persuades more people. It is that a good ad is cheaper to deliver, so the same budget reaches more of them.

It also explains why the performance of a creative degrades before the audience has technically seen too much of it. Engagement falls slightly, the delivery system responds by placing it less favourably, costs rise, and the decline compounds.

A branded search term routinely clears 20% because nobody is more relevant to the client name than the client. A cold display placement on a news site lives under 0.1% because almost nobody wanted an ad at all. Both can be performing exactly as they should, and an industry average across them describes nothing that exists.

The comparisons that carry information are narrower. A campaign against its own recent history tells you whether something changed. Two creatives in the same ad set against each other tells you which one works better for that audience. The same keyword against its own figure before and after an ad rewrite tells you whether the rewrite helped.

When a client arrives with a benchmark from an industry report, the useful response is to ask what campaign types it covered and over what period. Usually the answer is that nobody knows, which makes the point without an argument.

CTR reacts faster than any cost metric, because it sits earlier in the chain. Impressions have to turn into clicks before clicks can turn into conversions, so a change in click-through shows up a week or two before cost per result moves.

The pattern worth watching for is CTR falling while spend, targeting and bids all hold steady. On social that is creative fatigue, usually with frequency rising alongside it. On search it is more often a competitive shift: somebody new is showing above you, or a rival rewrote their ad copy.

Both are far cheaper to address at the point CTR moves than at the point cost per conversion does. Which is an argument for looking at this number weekly rather than folding it into a monthly report, and one of the clearest cases where the reporting cadence itself decides whether a problem gets caught.