How to build your own CTR curve from Search Console
Use your own Search Console data to estimate expected click-through rate by ranking position instead of relying on someone else’s CTR study.
4 min read
A CTR curve estimates how click-through rate changes as rankings move up or down.
You can use published CTR studies as a benchmark, but they describe someone else’s queries, brand, audience, and search results.
Search Console gives you enough data to build a curve based on your own performance. It will not predict exactly what every query will do, but it gives you a much better baseline for estimating ranking opportunities on your site.
The basic process is simple: group Search Console data by ranking position, sum clicks and impressions, then calculate CTR for each group.
What is a CTR curve?
A CTR curve shows the typical click-through rate you observe at different organic ranking positions.
The result might look something like:
Position | CTR |
|---|---|
1 | 31.4% |
2 | 17.8% |
3 | 12.1% |
4 | 9.0% |
5 | 6.7% |
Your actual numbers will be different.
The curve becomes useful when you want to estimate the potential value of moving a page or query from one ranking range to another.
It is a benchmark based on your historical data, not a prediction that every query at position 4 will receive the same CTR.
What Search Console data do you need?
In Search Console, open Performance → Search results and choose a date range with enough data to smooth out short-term noise.
Three months can be a reasonable starting point for a site with substantial search volume. Smaller sites may need a longer window.
Export query-level data with:
query
clicks
impressions
CTR
average position
Query-level data is useful because one page can rank at very different positions for different searches. A page-level average can hide those differences.
If you need more data than the Search Console interface export provides, use the Search Console API rather than assuming the exported sample represents the entire site.
How do you calculate a CTR curve?
In a spreadsheet:
1. Create position buckets.
Search Console reports average position, so do not treat 3.4 as if the query always ranked exactly third. For a simple curve, you can round positions into whole-number buckets.
2. Remove very low-impression rows.
A query with a handful of impressions can produce an extreme CTR that adds more noise than information. Choose a minimum impression threshold appropriate for the size of your dataset.
3. Separate branded queries.
Brand searches often have much higher CTR and can distort the top of the curve. Build a non-branded curve for general SEO opportunity analysis, and keep branded performance separate.
4. Group the remaining rows by position bucket.
5. Sum clicks and impressions for each bucket.
6. Calculate CTR using:
CTR = total clicks ÷ total impressions
Do not average the CTR column. A 50% CTR on two impressions should not carry the same weight as a 5% CTR on 20,000 impressions.
Want to check yours?
Run the CTR curve builder →
What does a site-specific CTR curve look like?
After grouping your Search Console queries, you might end up with something like:
Position | Impressions | Clicks | CTR |
|---|---|---|---|
1 | 48,200 | 14,700 | 30.5% |
2 | 39,600 | 7,100 | 17.9% |
3 | 31,800 | 3,900 | 12.3% |
4 | 27,400 | 2,500 | 9.1% |
5 | 24,900 | 1,700 | 6.8% |
These numbers describe this site, during this period, for the queries included in the analysis.
Another site could run exactly the same calculation and get a very different curve.
That is the point.
How should you interpret your CTR curve?
Look at the shape, not just individual percentages.
Where are the largest gains?
You may find a steep difference between positions 1 and 3, while moving from 8 to 7 produces relatively little additional traffic. Or your curve may flatten earlier because AI Overviews, shopping results, local results, or other SERP features capture attention above the organic listings.
Then use those differences to estimate ranking opportunities.
If position 8 typically receives 3% CTR and position 4 receives 9%, a query with 10,000 impressions has roughly:
10,000 × (9% − 3%) = 600 additional clicks of estimated upside
That does not mean the page will reach position 4. It tells you what reaching it could be worth
What can distort a Search Console CTR curve?
Several things can materially change CTR:
Branded queries. Brand searches often have unusually high CTR, especially at position 1.
SERP features. AI Overviews, featured snippets, local packs, shopping results, video, and other features change how much attention reaches organic results.
Query intent. Someone looking for a specific website behaves differently from someone comparing products or researching a broad topic.
Device. Mobile and desktop results can behave differently.
Average position. Search Console position is an average across impressions, not a record that the query appeared in exactly the same position every time.
Small samples. Low-impression position buckets can produce unstable CTR estimates.
A single curve is useful for prioritization. Segmented curves can make it much more useful when you have enough data.
How often should you update your CTR curve?
You do not need to rebuild it constantly.
Quarterly is a reasonable cadence for many sites, but the right frequency depends on how quickly your search landscape changes.
Rebuild it sooner after a major site migration, significant brand shift, large change in search mix, or meaningful change to the SERPs you compete in.
Keep previous versions.
Changes in the curve can be useful information themselves. If CTR at the top positions declines over time, for example, investigate whether the queries, SERP features, devices, or other conditions changed before assuming ranking value simply disappeared.
What are the most common CTR curve mistakes?
Averaging CTR.
Sum clicks and impressions first, then calculate CTR.
Mixing branded and non-branded queries.
Brand traffic can heavily distort the top positions.
Using page-level averages.
One page can rank at many different positions across many queries.
Including extremely low-impression rows.
Tiny samples can create extreme CTR values.
Treating average position as exact rank.
Search Console reports an average across impressions.
Over-segmenting small datasets.
A highly specific curve is not better if every bucket contains too little data to be useful.
What if you do not have enough Search Console data?
Do not force a site-specific curve from a dataset that cannot support one.
Start by increasing the date range and reducing unnecessary segmentation.
If some positions still have very little data, combine them into broader buckets such as:
positions 1–3
positions 4–5
positions 6–10
positions 11–20
If the site is still too small, use a published CTR benchmark as a starting assumption until you have enough first-party data.
A broader curve supported by enough data is more useful than a precise-looking curve built on noise.
Should you build more than one CTR curve?
Yes, when you have enough data for the segments to remain reliable.
The first split I would make is branded vs. non-branded.
After that, useful segments may include:
mobile vs. desktop
page type
query intent
country or market
queries with different SERP features
Do not segment simply because you can. Each split reduces the amount of data behind the curve.
Build the broad curve first. Segment when the difference is meaningful and the sample is large enough to trust.
What should you use your CTR curve for?
Use the curve to make ranking opportunities more concrete.
Instead of saying “this page is in position 8, so we should improve it,” estimate what a realistic move could actually produce.
Combine:
impressions × expected CTR gain = estimated additional clicks
Then compare that potential with the difficulty of improving the page and the business value of the traffic.
The CTR Curve Builder can create the curve from Search Console data. The guide on estimating the traffic value of improving an SEO ranking shows how to use it to prioritize pages.
Rankings tell you where you are. Your CTR curve helps tell you what moving might actually be worth.
Keep reading
