The click curve, and why it keeps flattening
The blended organic click-through rates used here — roughly 27 % at position one, 16 % at two, 11 % at three, falling below 3 % by position eight — come from the large-scale SERP studies published by Advanced Web Ranking and corroborated by independent analyses of Search Console exports. They are averages over millions of queries and hide enormous variance: a branded navigational query at position one can exceed 60 %, while a query answered directly in the snippet can drop the first result below 15 %.
What has changed since 2023 is the amount of the result page that resolves the query before anyone scrolls. Featured snippets, People Also Ask, product carousels, map packs and now AI Overviews all absorb intent. Google has not published the click loss, and the independent measurements disagree with each other, which is why this simulator exposes the SERP context as a slider you set rather than a constant hidden in the code. The discounts applied — from a 18 % reduction for ads to roughly half for an AI Overview stacked with ads — sit in the conservative middle of published ranges.
The practical consequence is that position targets should be set per intent, not per site. On informational queries where an AI Overview now answers the question, the honest forecast for position three may be lower than the forecast for position eight on a commercial query with no rich results at all. This is the mechanism behind the shift towards commercial and transactional content that most serious programmes made over the last two years, and it is covered in depth in our guide to SEO and GEO.
Where keyword volumes go wrong
Every third-party volume is a model, not a measurement. Tools infer monthly searches from clickstream panels and from Google's own bucketed ranges, then smooth them over twelve months. The error bars are wide, especially in smaller languages and on long-tail terms, and seasonal products can be off by an order of magnitude in either direction depending on the month you look.
The most reliable correction available to you is free: your own Search Console impressions. For any query you already appear on, impressions are a direct measurement of how often the query ran and you were eligible, which anchors the model far better than a third-party estimate. Forecast the queries you already touch from your own data, and reserve tool volumes for the terms you have never ranked for.
Coverage is the other quiet exaggeration. A keyword set of 250 terms will not all land on one page at the same position; a realistic page wins the head term and a long tail of variants, and misses the ones with a different underlying intent. Setting coverage to 100 % is how forecasts end up three times too high, which is why the default here is deliberately lower.
- Anchor forecasts on Search Console impressions wherever you already appear
- Treat tool volumes as ranges, not numbers, especially outside English
- Check seasonality before annualising a single month
- Never assume full coverage of a keyword set on one page
- Re-check the SERP layout every quarter — rich results change
- Forecast clicks, then convert to revenue; never forecast revenue directly
Turning a forecast into a plan
The most useful output of this simulator is not the click number, it is the shape of the curve. Look at the gap between position ten and position five, then between five and one. On most keyword sets, the first move is worth more than the second and costs a fraction as much, which tells you where to spend: pages stuck on the second page usually need one honest content improvement and three internal links, not a link-building campaign.
Use the paid equivalent as a sanity check, not as a valuation. If buying the same clicks would cost 40,000 a month, and the SEO programme to earn them costs 4,000, someone should ask why the paid team is not already buying them — often the answer is that they are, and the incremental value of the organic click is lower than the model suggests. That is a healthy conversation to trigger before a budget is approved rather than after.
Finally, tie the forecast to the people who will deliver it. A forecast is a hypothesis about capacity as much as about demand: it assumes someone will write the pages, fix the templates and earn the links. If you do not have that capacity, the useful next step is comparing SEO agencies and independent consultants against the workload this forecast implies, or sizing it with the budget estimator.