Click-through rate is the share of people who saw your result in Google and clicked it. Search Console defines it simply: clicks divided by impressions. It is one of the few SEO numbers you can move in an afternoon, because it depends partly on words you control, the title and description your result shows. It is also the number most distorted by AI Overviews, which can answer the question before anyone reaches your result.
This post covers how to find the pages losing clicks, what actually lifts CTR, and how to read CTR now that the answer itself is often the outcome. For the definition alone, see the click-through rate entry in our glossary.
How do you find the pages that are losing clicks?
Open the Performance report in Search Console and switch on all four metrics: clicks, impressions, CTR and average position. Then look at pages, not the site total. A site-wide CTR mixes branded and unbranded queries, top and bottom positions, and tells you nothing you can act on.
Filter for pages that have:
- Plenty of impressions, so a change will be visible in the data within a few weeks.
- A good average position, roughly the top five, so the result is on screen.
- A CTR well below other pages at similar positions on your own site.
Those pages are seen and skipped. Their problem is the result, not the ranking, and that is good news: the result is the cheapest thing on the page to change.
Then open the queries for each page. A page can rank for a query it does not answer, and the low CTR is the searcher telling you so. That is not a snippet problem; it is an intent mismatch, and the fix is a different page or a different target query.
Search Console's average position is the topmost position of your link, averaged across queries, so a page ranking first for one query and twentieth for another can show an average that matches neither. Read the query table before trusting the page's average.
What is a normal CTR by position?
Recent data puts the number lower than many teams expect. Ahrefs' study of 300,000 keywords, published in February 2026 from aggregated Search Console data, found the first result's average desktop CTR on informational keywords was 3.9% in December 2025, down from 7.6% two years earlier, and 1.6% where an AI Overview appeared. Use benchmarks like that for orientation only. Your own pages at similar positions for similar queries are a better yardstick, because CTR varies with query type, brand strength, device and the features on the page.
What actually lifts organic CTR?
Work through these in order. The first two are almost always the cause.
Match the title to the question
The title should tell the searcher this page answers their question. Use the words they used, put the specific thing first, and cut the brand to the end. "How to improve organic CTR" beats "The Ultimate Guide to CTR Success". Numbers and years help only when they are true and relevant.
Write the description as the answer's first line
Google often rewrites descriptions, but it uses yours when it fits the query. Write one or two sentences that answer the question or say exactly what the page gives the reader. A description that restates the title wastes the space.
Earn rich results where they fit
Structured data can make your result more useful: breadcrumbs, product ratings, a video thumbnail, an article date. It does not guarantee a rich result, and the markup must describe what the page visibly shows.
Keep dates honest
For queries where freshness matters, a visible, real update date can earn the click over an older-looking result. Changing a date without changing the page is a short-term trick that readers notice.
Fix the URL and breadcrumb
A readable URL and breadcrumb tell the searcher where they will land. A string of parameters tells them nothing.
Check the page before blaming the snippet
Search the query yourself, in a clean browser, on a phone. Is there an AI Overview above you? A featured snippet? Four ads? A video carousel? If the screen is full before your result appears, no title rewrite will fix CTR.
What does a CTR rewrite look like?
Here is the shape of a typical fix, written out so you can see each decision. The page is a guide that ranks around position three for "how to choose a CRM for a small team" and gets far fewer clicks than the pages around it.
Before. Title: "CRM Buying Guide | Brand". Description: "Everything you need to know about CRMs, from features to pricing and beyond."
What is wrong. The title names a topic, not the question. The description promises everything and says nothing specific. Nothing tells the searcher this page is about small teams, which is the only word in their query that narrows it.
After. Title: "How to Choose a CRM for a Small Team (5 Questions) | Brand". Description: "Five questions that decide which CRM fits a team of 2 to 20: who uses it daily, what it must connect to, and what it costs after year one."
The new result uses the searcher's words, says what the page gives them, and shows the specific angle. The brand moves to the end because the searcher does not know it yet.
Then wait. Compare the next four weeks with the four weeks before, at a similar average position, before deciding whether it worked. If position moved a lot in the same window, the comparison is not clean; wait longer or compare query by query.