Why is my click-through rate so low in Search Console?

The top and the bottom are counted on different things

Start with the arithmetic, because the definition is not in dispute anywhere.

Microsoft states the standard formula: "The click-through rate (CTR) is the number of times an ad was clicked, divided by the number of times the ad was shown (impressions)."

Nothing controversial there. The trouble is not the division. It is that in some result types the two counts attach to different objects.

A click attaches to the one listing a person chose. An impression attaches to every listing that appeared. When results are presented as a list that expands into a details panel, those are not the same population at all.

Google spells out the effect: "This can cause your impressions to be higher than your clicks."

That sentence is doing more work than it looks. It is not describing a problem. It is describing the designed behaviour of the counting rules.

One visitor action can hand impressions to everybody listed

Here is the mechanism in the form that matters to a reader, and it is worth reading twice.

Google describes what happens on a click: "When a list item is clicked, only a single provider gets the click. However, in the resulting details view, all providers that posted details will get an impression."

Read that from the point of view of a business that was not chosen. Somebody clicked a competitor. The panel opened. You got an impression out of it. Your denominator grew from an action that was never available for you to win.

Impressions can also arrive with no list interaction at all: "If you share a direct link to a details pane with a friend and they open it, there will be one impression for the details view and zero clicks or impressions for the listing view." A shared link produces a view that no search result ever showed.

There is a third case, and it is the most ordinary one. On desktop, expanding a list automatically opens the details panel for the first item. Google notes that this counts as a details impression without a click for the matching list item. Nobody chose anything. The interface opened by itself, and a number moved.

None of these are errors. They are all the same fact seen three ways: an impression records that something was shown, not that somebody considered it.

A second look is still one impression

Impressions count appearances rather than attention, and Google documents that for one result type in a single sentence.

For AMP pages it states: "An impression is counted only once no matter how many times the user sees the page in either the carousel or in the viewer."

Note the limit before borrowing the idea. That sentence is written about one specific result type, not about every result on the page. It is worth quoting because it makes the principle checkable, not because it settles every case.

What it shows is which direction the counting runs. Somebody who scrolled past your listing, came back, and looked again is recorded the same as somebody who glanced once and moved on. The count answers whether something appeared, not how much attention it held.

That matters for the rate. It means the bottom number is not a measure of interest. It does not rise because people are more engaged, and it does not fall because they are less so. It moves when the number of appearances moves.

So when a rate drops there are two candidates worth separating, and only two. Either fewer people chose you, or more appearances were recorded. Those look identical in the rate and have almost nothing in common as problems.

Being seen is measured, not observed

This is the part that explains why no publisher can simply hand you a true number, and the honest statement comes from outside the search industry entirely.

Whether a person actually saw an element on a screen is not a fact sitting in a database. It has to be determined, and the web platform found that genuinely hard. MDN puts it plainly: "Historically, detecting visibility of an element, or the relative visibility of two elements in relation to each other, has been a difficult task for which solutions have been unreliable and prone to causing the browser and the sites the user is accessing to become sluggish."

That difficulty is why browsers eventually shipped a dedicated tool for it, and MDN lists measuring advertising visibility among the reasons it was needed.

The interesting part is what that tool requires you to supply. Visibility is reported as a proportion, described as "a representation of the percentage of the target element which is visible as a value between 0.0 and 1.0." Somebody has to choose how much counts. MDN gives the extreme: "A threshold of 1.0 means that when 100% of the target is visible within the element specified by the root option, the callback is invoked."

So "was it seen" is a question with a dial on it. Move the dial and the same afternoon of real human behaviour produces a different number of impressions. Nothing about the visitors changed.

That is the honest frame for any impression count anywhere. It is a measurement made under a chosen rule, not a headcount.

What to check, in order

Five checks, and all of them are free.

• Find out which result types your pages appear in before judging the rate. List and panel results follow different counting rules from ordinary blue links.
• Compare like with like. Filter to one search appearance type rather than reading a single site-wide number that blends several counting rules together.
• Look at clicks on their own first. Clicks are the count that always attaches to a deliberate choice, so a rising click count with a falling rate is usually good news.
• Check whether impressions grew before concluding that clicks fell. A rate can drop while nothing about your performance changed.
• Only then treat the listing itself as the problem, and test the title and description.

The arithmetic makes the trap obvious. Suppose your listing earns 40 clicks from 400 impressions in a month, which is a rate of 10 percent. Next month a competitor runs a campaign, more people open the panel your listing sits in, and impressions reach 800 while your clicks rise to 48.

Your rate is now 6 percent, and it looks like a serious decline. But clicks went up by 8, which is a 20 percent increase in actual visitors. The figures are an illustration. The lesson is that the rate fell because the denominator was filled by other people's activity, and a business that reacts by rewriting a title that was working would be fixing the wrong thing.

Working out which counting rules apply to your pages before anybody rewrites them is part of what we do on search.

Terms used on this page

• Click-through rate

Sources

• What are impressions, position, and clicks? (Google Search Console Help)
• Report attributes and performance statistics (Microsoft Advertising, Microsoft Learn)
• Intersection Observer API (MDN Web Docs, Mozilla)

Last reviewed 2026-09-12.