How does AHREFS and SEMRush and Moz, etc, know the volume of searches for a given keyword/phrase?

Every SEO tool prints a tidy monthly search volume next to a keyword, yet none of them can see Google's real query counts. This article explains where that number actually comes from, why Ahrefs, Semrush and Moz rarely agree on it, and how to read and blend the figures without over trusting any single one.

The short answer is that they do not know it, they estimate it. Google is the only party that can count how many times a query is typed, and it does not publish that number. Everything you see in a keyword tool is a model built from partial signals, calibrated to be useful for comparison. Understanding how that model works tells you exactly how much to trust the figure, and where it breaks down.

Where the raw signals come from

Three inputs do most of the work. The first is Google Keyword Planner, the free tool inside Google Ads. It reports volume in broad buckets, for example 1K to 10K, and it is aimed at advertisers rather than at giving precise organic numbers. It anchors the model but is far too coarse on its own. The second input is clickstream data: anonymized records of what real people search and click, gathered from panels of users through browser extensions, apps, and data partnerships. The third is each tool's own crawl of the search results, which captures rankings, SERP features, and how many keywords a page ranks for. The search volume definition in the lexicon covers the term itself in more detail.

How the model turns signals into a number

A clickstream panel is only a sample, perhaps a few million users, so the tool must scale it up to represent everyone who searches. That extrapolation is where the modeling happens. The engine blends clickstream counts with Keyword Planner ranges, corrects for known bias in the panel, for example an over representation of certain countries or demographics, and adjusts for seasonality so a spike in December does not distort the annual figure. For rare and long tail keywords, where the panel may show only a handful of searches or none at all, machine learning fills the gaps by inferring volume from related terms and from the behavior of similar queries. The output is a single tidy number like 8,100, which hides a genuine range underneath.

Why the tools disagree

Because each vendor runs a different panel, a different scaling model, and a different refresh cycle, the same keyword can show meaningfully different volumes across tools. This is not a defect: it is the natural result of estimating an unknowable quantity from different samples. Head terms with heavy search activity tend to converge, because there is plenty of clickstream signal to anchor them. Long tail terms diverge more, because the model is doing more of the guessing. When a number matters to a decision, check it in two or three tools and treat the spread as your confidence interval. The related keyword difficulty guide explains why a companion metric like difficulty is also modeled and should be read the same way.

How to read volume like a practitioner

Use volume for what it is good at: comparing keywords against each other and sizing the rough opportunity of a topic. Do not use it to promise an exact traffic number to a client or a stakeholder. A term listed at 8,100 might deliver a fraction of that in clicks once you account for your ranking position, the click through rate at that position, and how many SERP features push organic results down the page. Pair tool volume with your own Google Search Console impressions once you rank, since that shows how often your pages actually appeared for a query. For the full workflow around choosing and validating terms, see the keyword research fundamentals guide.

Volume data sources compared

How the major tools differ in practice

The vendors describe their methods a little differently, and those differences explain most of the gaps you see. Ahrefs leans heavily on a large clickstream dataset and reports a global volume as well as a country level one, which is why its figure can look higher than rivals for international terms. Semrush blends clickstream with its own keyword database and a machine learning model, and shows an intent label beside the volume. Moz calibrates its estimate against Google Keyword Planner ranges plus a clickstream signal and presents volume as a bucketed range rather than a single hard number, which is arguably the most honest way to display an estimate. None of these is the truth: each is a different lens on the same figure that only Google can actually count.

Blending figures into one working number

When a decision rides on volume, do not average the tools blindly. Pull the estimate from two or three sources, drop any obvious outlier that sits far from the rest, and take the median of what remains as your working figure. Treat the full spread as a confidence interval: a term that reads 6,000 in one tool and 12,000 in another is really "roughly ten thousand, give or take," so plan content and forecasts against the low end. Round aggressively too. A keyword shown as 8,100 is no more precise than one shown as 8,000, and pretending otherwise builds false confidence into any traffic model you hang off it.

Search Console impressions as your ground truth

Once a page ranks, you no longer have to guess. Google Search Console reports the exact number of times your pages appeared for a query under Performance, Search results, filtered by query, and that is real Google data rather than a modeled estimate. Compare the impression count against the tool volume for the same term: if a tool said 8,000, you rank on page one, yet you see only 2,000 impressions a month, the tool overstated demand or the query is more seasonal than its flat annual number implied. Over time this builds a private calibration for your niche, letting you trust or discount specific tools for the keywords that matter to you. Read impressions next to average position and click through rate in the same report to turn interest into a realistic traffic expectation.

What has changed since this discussion

Clickstream supply has become less stable in recent years, as browsers tightened privacy and several data brokers changed hands or shut down, which is one reason volume figures can shift between tool updates. At the same time, zero click searches and AI overviews mean a high volume keyword no longer guarantees clicks even at the top of the results. The practical takeaway is unchanged but sharper: volume tells you interest exists, not how much traffic you will capture, so always pair it with position and click through data.

Frequently Asked Questions

Source: https://old.reddit.com/r/bigseo/comments/17ebu28/how_does_ahrefs_and_semrush_and_moz_etc_know_the/