CTR by position is the number people reach for when they want to turn a ranking into a traffic forecast. Move from fifth to first, the reasoning goes, and clicks multiply by a known factor. The direction is right. The precision is not. Published curves vary widely, and the variation is not noise: it reflects real differences in what the results page looks like for different queries.
What the CTR by position data shows
The largest public dataset is Backlinko’s Google CTR stats by Brian Dean, updated 16 April 2025. It covers 4 million Google results, drawn from 1,312,881 pages and 12,166,560 queries, using data from Semrush’s Search Console accounts. Search Console data matters here: these are recorded clicks and impressions, not panel estimates.
| Position | Average organic CTR (Backlinko) |
|---|---|
| #1 | 27.6% |
| #2 | 15.0% |
| #3 | 11.0% |
| #10 | 2.8% |
By Backlinko’s measure, the first result is 10 times more likely to receive a click than the tenth. The drop from first to second is the steepest step on the curve: the second result gets a little over half the first result’s rate.
Why CTR studies disagree
Other studies report very different position-one figures, and none of them is simply wrong. CTR depends on query type, device and what else sits on the page. A navigational brand search sends almost everything to one result. A broad informational query shows an AI Overview, a featured snippet, a People Also Ask box and videos before the first blue link, and the click rate for that link falls accordingly.
Our analysis of how AI Overviews affect traffic covers the biggest recent change, and the post on zero-click searches covers queries where the answer is on the results page itself. Both push real CTR below any averaged curve.
So treat a published curve as a shape, not a table to multiply by. The shape is reliable: steep at the top, long and flat after the first few positions. The numbers are an average across queries that may look nothing like yours.
Device adds another layer. On a phone, the first organic result can sit below an entire screen of ads, maps and answer boxes, so “position one” describes a rank, not a place on the screen. A curve built from a mix of devices hides that, which is why two honest studies can publish first-position figures far apart and both be accurate for the queries they sampled.
The curve’s shape is reliable. Its numbers are an average of queries that may look nothing like yours.
What moves CTR at the same position
Position is not the only variable you control. Backlinko’s data includes two findings on titles:
- Title length. Titles of 40–60 characters had the highest CTR, 8.9% better than titles outside that range.
- Question titles. They performed about the same as non-question titles, 15.5% against 16.3%, a difference Backlinko found not statistically significant.
The second finding is the more useful one, because it kills a common habit. Rewriting every title as a question does not, on this evidence, earn more clicks. A title of sensible length that states clearly what the page delivers is the safer bet.
Bear in mind that Google sometimes rewrites the title link it shows, so the title in your CMS is not always the one searchers see. Check the live result before concluding a title test failed. When you do change a title, change only the title, note the date, and compare CTR at a similar average position over the following weeks. Changing the title, the content and the internal links together leaves you unable to say which one moved the number.
Organic CTR by position on your own site
The only curve that matters for your forecasts is yours, and Search Console already holds it. Open the Performance report, enable clicks, impressions, CTR and average position, then export the queries. In a spreadsheet, group queries by rounded position and average the CTR for each group.
Three adjustments make the result usable:
- Remove branded queries. They sit at position one with very high CTR and distort the top of the curve.
- Split by device. Mobile and desktop results pages differ enough to produce different curves.
- Set a minimum impression count. A query with a handful of impressions produces a CTR that means nothing.
Then look for outliers. A query at position three with half your usual position-three CTR is either facing heavy SERP features or carrying a weak title. Check the live results page to see which. Our guide to SERP analysis covers what to look for.
Using CTR to choose what to work on
The practical value of the curve is prioritisation. Because the curve is steep at the top, moving a page from fourth to second is usually worth more than moving another from fifteenth to tenth. Pages sitting just below the top three on queries with real impressions are the cheapest gains on most sites.
To rank the candidates, take each query’s monthly impressions and multiply by the difference between your own CTR at its current position and your CTR at the target position. The result is a rough count of extra clicks, good enough to sort a list of fifty pages into an order of work. It is not a forecast to promise anyone, because the results page for that query may not behave like your average.
The same logic applies when you size a niche. In the database explorer, filter to a niche and sort by estimated traffic; the sites at the top are the ones holding the first few positions on enough queries to matter. Traffic figures there are estimates, and estimates are themselves built on assumed CTR curves, which is one more reason not to treat any single curve as exact.
The limitation: Search Console’s average position is itself an average across every impression, so a query that ranks third on desktop and seventh on mobile reports as fifth, a place it may never have held. Your curve inherits that blur. It is still far better than borrowing someone else’s, because the blur comes from your own queries, devices and results pages rather than from a mix of other people’s.
Recalculate it every few months. Results pages change, AI features expand to new query types, and a curve built last year can overstate what a top position earns now. More data-led posts are in the data archive.
Frequently asked questions
What is the average CTR for position 1 on Google?
Backlinko’s study of 4 million results put it at 27.6%. Other studies report different figures depending on query type, device and SERP features, so measure your own.
How much more do top results get clicked than position 10?
In Backlinko’s data the first result was 10 times more likely to be clicked than the tenth, which averaged 2.8% CTR.
Do question titles get more clicks?
Not on Backlinko’s evidence. Question titles scored 15.5% against 16.3% for other titles, a difference that was not statistically significant.
How do I find my own CTR by position?
Export queries from Search Console’s Performance report, remove branded terms, split by device, and average CTR for each rounded position.
The takeaway Trust the shape of the CTR curve, not anyone’s exact numbers. Build your own from Search Console, and spend effort where a move into the top three is within reach.
