.webp){kind=logo}
CTR by Position in 2026: How to Read Google's Click Curve
This guide shows you how to read CTR by position the way searchers see the page: by results layout, by query type and by the noise in your own Search Console data. Every example uses the calculator above, so you can check each number yourself.
What CTR by position measures
The classic curves came from results pages with ten blue links. Today the same position can sit below an AI Overview, ads, a video block or a local pack. A single curve then overstates clicks on busy pages and understates them on clean ones, and your targets drift away from what real people do.
Figure 1. Planning scenarios at position 4.3. The factors do not measure causal click losses.
How to use the calculator
Averages can look compatible even when you mix reporting periods. For example, last month’s clicks divided by a year of impressions produce a low rate that no title change can explain.
Why the same position earns different clicks
A fourth result can appear near the top on one search and below an AI summary, shopping block and ads on another. The ranking number alone does not describe what a person sees.
You can check representative queries in the same country and on the same device as your report, then tick only the features you see. The factors create a scenario you can test; they do not provide a measured country-specific curve.
How to interpret the AI Overview factor
Pew reported clicks on traditional results in 8% of visits with an AI summary and 15% without one. The study compared whole search visits, not the CTR of each individual position, so 8/15 cannot be treated as a universal ranking penalty.
You can treat 0.53 as a starting sensitivity assumption and compare it with 1.00 and your own segmented observations. This shows how much your conclusion depends on the assumption before you prioritize work.
How to read Search Console average position
Search Console position is an average across impressions, and the topmost result from your property can determine its reported position. A value of 4.3 does not mean your page always occupied one fixed slot.
You can narrow the report to a query, page, country and device before you calculate CTR. Linear interpolation keeps decimal inputs usable, but a lookup at an average is still an approximation of a changing distribution.
Why brand queries need a separate curve
People searching your name already expect to find your site. If you combine those queries with broad product searches, high branded CTR can hide weak non-brand performance.
You can enter your own brand names and common variants in CSV mode. The audit uses literal matches and labels those rows separately, so you can review ambiguous names instead of trusting automatic intent detection.
What the Wilson interval does and does not show
A sample of 10 clicks has more sampling uncertainty than 10,000 clicks. A Wilson interval describes that uncertainty under a binomial model, but repeated users, query mixtures and layout changes can violate its simple assumptions.
You can show the observed interval next to the sensitivity band, then gather another comparable period before you act. A non-overlapping interval is a reason to investigate, not evidence that the benchmark is correct or a snippet caused the gap.
How to improve CTR at the same position
A low modeled CTR can reflect weak wording, mismatched intent, a competing brand or an unrealistic curve. For example, a page about maintenance may attract fewer clicks for a query that asks for a repair service.
You can test one relevant change at a time, such as a clearer title or a description that states what the page answers. A matching promise gives real people a reason to choose the result. Accurate claims matter especially for health, finance and other trust-sensitive topics.
When to compare a ranking move with a snippet test
The target-position forecast combines a new position with unchanged SERP assumptions. Comparing it with actual clicks also includes any existing benchmark gap, so the two uplifts answer different questions.
You can use the isolated ranking uplift to compare ranking opportunities and the modeled current-position gap to select snippet tests. You can then check outcomes against the same filters and period rather than treating either forecast as guaranteed traffic.
CTR by position checklist
A detailed calculator can still give an unreliable answer if the inputs describe different audiences or if you accept every default without checking it.
You can confirm the date range, traffic basis, query type, SERP context and sensitivity assumptions before you save the result. Keeping those choices in the exported report makes the comparison repeatable and transparent.
