How does AHREFS and SEMRush and Moz, etc, know the volume of searches for a given keyword/phrase?
- July 27, 2024
- General
AI Summary
SEO tools do not read live data from Google. They estimate monthly search volume by combining Google Keyword Planner ranges with anonymized clickstream data from browser panels, then extrapolating and calibrating those signals with statistical models. Because each tool uses a different panel and model, their numbers differ, so search volume is best read as a directional range rather than an exact count.
- Only Google knows true query counts, so every third party volume figure is a model output.
- The two core inputs are Keyword Planner ranges and clickstream panel data from real browsers.
- Tools extrapolate a small panel to the whole population and correct for bias, geography, and seasonality.
- Different panels and models are why Ahrefs, Semrush, and Moz report different numbers for the same keyword.
A recurring question among practitioners is how tools like Ahrefs, Semrush, and Moz can publish a monthly search volume for a keyword when Google never shares exact query counts. The short answer is that they do not know the true number. They estimate it. This page explains the data sources and modeling behind those estimates, and how to use the figures responsibly.
Google is the only source of the true number
Google is the only party that knows exactly how many times a query was searched. The closest official signal is Google Keyword Planner, and even that reports broad ranges, for example 1K to 10K, rather than a precise value, and those ranges are aimed at advertisers. So the exact monthly volume you see in an SEO tool is never pulled from Google. It is reconstructed from other evidence.
The two core data inputs
The first input is Keyword Planner data, which gives a coarse bucket and a directional sense of scale. The second, and the one that lets tools produce a specific number, is clickstream data. Providers buy or license anonymized browsing data from panels of real users, often gathered through browser extensions, apps, and internet service partnerships. Within that panel they can count how often a keyword was searched and which results were clicked.
A panel is only a sample of the internet, so the raw panel count is not the real volume. This is where modeling comes in.
How the model turns a sample into a number
The tool extrapolates the panel count up to the full search population, using the known size and skew of the panel. It then calibrates that estimate against Keyword Planner ranges so the output stays inside a plausible band. On top of that it corrects for panel bias, since panel users are not perfectly representative, and adjusts for geography and seasonality. For rare or new keywords with too little panel data, machine learning fills the gaps by inferring volume from similar terms and SERP characteristics.
What each data source contributes
| Data source | What it provides | Limitation |
|---|---|---|
| Google Keyword Planner | Official range and scale | Broad buckets, advertiser focused |
| Clickstream panels | Actual query and click counts | Only a sample, panel bias |
| SERP and CTR data | Click distribution by position | Estimated, feature dependent |
| Machine learning model | Fills gaps for rare keywords | Inference, not measurement |
How to use volume estimates responsibly
Because the number is modeled, treat it as a range, not a promise. Do not switch tools mid project and expect the figures to match, since each vendor uses a different panel and calibration. Use volume for relative comparison and prioritization: keyword A is roughly ten times bigger than keyword B is a safe read, while keyword A gets exactly 8,100 searches is not. Pair volume with intent and difficulty before you commit resources.
For related guidance, see our references on search volume, keyword difficulty, a practical guide to judging keyword difficulty, and the keyword research guide.
Source: https://old.reddit.com/r/bigseo/comments/17ebu28/how_does_ahrefs_and_semrush_and_moz_etc_know_the/
Frequently Asked Questions
Do Ahrefs and Semrush get search volume directly from Google?
No. Google never shares exact query counts. Tools estimate volume by combining Google Keyword Planner ranges with anonymized clickstream data from browser panels, then calibrating that with statistical models. The number you see is a model output, not a reading from Google.
What is clickstream data?
Clickstream data is anonymized browsing activity licensed from panels of real users, typically collected through browser extensions, apps, and internet provider partnerships. Within that panel a tool can count how often a keyword was searched, which it then extrapolates to the full population.
Why do different tools show different search volumes?
Each vendor uses a different clickstream panel, a different extrapolation method, and different calibration against Keyword Planner. Those differences compound, so the same keyword can show meaningfully different numbers across Ahrefs, Semrush, and Moz.
How accurate are search volume estimates?
They are directionally useful but not exact. For high volume, well sampled keywords they tend to be reasonably close, while for low volume, new, or niche keywords they can be far off because there is little panel data to model from.
Should I trust the exact monthly number?
Treat it as a range and a relative signal rather than a precise forecast. It is reliable for comparing the rough scale of keywords and prioritizing work, but not for promising a specific amount of traffic.
How do tools estimate volume for rare keywords?
When a keyword has too little panel data, machine learning fills the gap by inferring volume from similar keywords, related terms, and SERP characteristics. This is why very obscure keywords often show round or clearly modeled numbers.
Claude Vincent is a technical SEO consultant focused on crawlability, rendering, and AI-search visibility. He writes the field guides and case studies at SEO ProCheck, with a bias toward the durable, unglamorous work that decides whether search engines and AI answer engines can actually read and cite a site.
About SEO ProCheck
Technical SEO consulting and GEO strategy with 20 years of enterprise experience. Case studies, resources, and tools for search and AI visibility.
Work With Me
Technical SEO audits, GEO strategy, site migrations, and international SEO. Hourly consulting for teams who need hands-on support, not just reports.
Subscribe to our newsletter!
