Why two keyword tools give the same keyword two different numbers
Nobody outside Google can see how often a term is searched, so every tool models it — from different sources, with different grouping and different rounding. What that means for the number on your screen.
The short version
Two tools disagree because neither can see Google's numbers, so each one builds its own model from a different mixture of sources. Ahrefs states this outright in its help centre: "Search volume in any SEO tool, including Ahrefs, is always an estimation. Only Google has access to the exact data."
The practical rule that follows is simple and almost nobody applies it: compare keywords against other keywords inside one tool, and never compare one tool's number against another's. A gap between two tools is information about the tools, not about the keyword.
Where each tool's number actually comes from
Google Keyword Planner is the only source with access to Google's own data, and even it does not hand you a raw count. Google defines its figure as the average number of searches for "a keyword and its close variants", based on your chosen month range, location and network, averaged over twelve months by default — and adds that "your search volume statistics are rounded".
Ahrefs describes its own method as "a combination of Google Keyword Planner (GKP), Google trends data and other third party data sources", and says it uses clickstream data to "un-group keyword clusters from Google Keyword Planner and display a search volume for each keyword". Its free Keyword Generator page says its volume estimates are "all powered by clickstream data".
Other vendors publish much less. Semrush advertises a database of "more than 26 billion keywords and growing" without describing how a volume is derived. Mangools claims it has "exact search volumes". SE Ranking describes a "unique database" with regional counts. None of those are descriptions of a method, and nothing on this site treats them as one.
The three mechanisms behind the gap
- Bundling. Keyword Planner reports a term together with its close variants, so plurals and rewordings can share a single number. A tool that tries to split them apart will always differ from one that does not.
- Rounding. Google warns that rounded volumes "might not add up as you'd expect" when you pull several locations. A precise-looking 8,100 is a rounded figure, and two tools rounding at different points diverge.
- Averaging. The default figure covers twelve months. A term that tripled last November and has been flat since carries the same annual average as one that has been steady all year, and tools that weight recent months differently produce different answers.
The study everyone quotes, and who ran it
In November 2021 Ahrefs published a comparison of Google Keyword Planner volumes against Google Search Console impressions for 72,635 keywords in the 1,000 to 10,000 volume range. It reported that "91.45% of search volumes in GKP are overestimates", that 54.28% were drastically overestimated, and that "45.22% are roughly accurate (i.e., deviating from GSC 'impressions' by no more than 50%)".
Ahrefs attributed the gap to Keyword Planner grouping close variants and rounding — the same two mechanisms above. Its help centre elsewhere claims its own estimates were roughly accurate for 60% of studied keywords against the 45% it gives Keyword Planner.
Read that for what it is: a vendor's own study of a competitor's data, published by the vendor, unaudited, and five years old at the time of writing. It is quoted here because it is the most concrete public attempt at the question, and labelled carefully because a vendor marking its own homework is not evidence of superiority. Nobody at Querywell repeated it.
What the tools say about themselves, side by side
How four vendors describe their own volume data, quoted from their own pages, read 24 September 2026
| Source | What it says about its own numbers |
|---|---|
| Google Keyword Planner | An average for "a keyword and its close variants", rounded, over twelve months by default |
| Ahrefs | "Always an estimation. Only Google has access to the exact data." Built from Keyword Planner, Trends and third-party sources |
| Mangools | "We have exact search volumes and the most accurate keyword difficulty" |
| Google Trends | Not a count at all — each point "is divided by the total searches of the geography and time range it represents", scaled 0 to 100 |
Google Keyword Planner
- What it says about its own numbers
- An average for "a keyword and its close variants", rounded, over twelve months by default
Ahrefs
- What it says about its own numbers
- "Always an estimation. Only Google has access to the exact data." Built from Keyword Planner, Trends and third-party sources
Mangools
- What it says about its own numbers
- "We have exact search volumes and the most accurate keyword difficulty"
Google Trends
- What it says about its own numbers
- Not a count at all — each point "is divided by the total searches of the geography and time range it represents", scaled 0 to 100
Two of those four rows flatly contradict each other. That contradiction is the most useful thing on this page, and it is why no tool is crowned most accurate anywhere on this site.
Where the gap is widest
Disagreement is not evenly spread. Two tools will usually land within the same order of magnitude on a common commercial term, because there is plenty of data behind it and every model has seen it. The gaps open up in three places.
- Low-volume phrases, where a model has almost nothing to work from. A term one tool prints as 50 and another prints as 0 is not a contradiction — one model guessed and the other declined to.
- Seasonal terms read at the wrong time of year, because a twelve-month average flattens a spike that has just happened or is about to.
- Close variants, where the whole disagreement is structural: one tool is reporting the family and another is reporting the member.
How to work despite it
- Pick one tool and stay in it for a given project. Relative order inside one model is far more trustworthy than absolute numbers across two.
- Treat a volume as an order of magnitude. The difference between 200 and 2,000 is a decision; the difference between 1,900 and 2,400 is noise.
- Check seasonality separately in Google Trends, which is measuring shape rather than size and is honest about it.
- Use Search Console for anything you already rank for. It reports what Google actually did, not what a model thinks it did.
- If a keyword is going to cost you a week of work, look it up in a second tool — not to average the two, but to see whether they agree on the order of magnitude.
Difficulty scores diverge for exactly the same reason and rather more dramatically, which the note comparing how each vendor builds its difficulty score takes apart. If you are still choosing a tool, the eight keyword research tools with their entry prices is the comparison.
Questions about search volume
Which keyword tool has the most accurate search volume?
There is no way to know from outside Google, and this site did not test it. Ahrefs, which sells one of these tools, states in its own help centre that search volume in any SEO tool including its own "is always an estimation". Treat any site that names a most accurate tool without publishing a method with suspicion.
Should I average the numbers from two tools?
No. Averaging two models built from different sources with different grouping produces a third number that describes nothing. Use one tool for comparisons and treat a second tool only as a sanity check on order of magnitude.
Why is Keyword Planner's number higher than my SEO tool's?
Often because Keyword Planner bundles close variants into one figure while the other tool tries to split them apart, and because Keyword Planner rounds. Ahrefs' 2021 study of 72,635 keywords reported that 91.45% of Keyword Planner volumes overestimated Search Console impressions, and attributed it to exactly that bundling and rounding.
Does search volume even matter, then?
Yes, as a rough size signal. It tells you whether a topic is worth a page at all and how topics rank against each other within one tool's model. What it cannot do is predict your traffic, because that depends on where you rank, what the results page looks like and how many people click anything at all.
Is Google Trends a substitute for search volume?
No, it answers a different question. Google explains that each Trends data point "is divided by the total searches of the geography and time range it represents" and is then "scaled on a range of 0 to 100". That makes it excellent for direction and seasonality and useless for size.
There is no partner link anywhere on this page, so nothing on it earns this site money. Where a tool named here does run a programme we take part in, it is listed — with the ones that pay nothing and never will — on the page that sets out how we get paid.
Picking a tool rather than reading about one? The eight keyword research tools with their entry prices side by side is the page to start from, and what the free tiers actually give you is worth reading first if you have not spent anything yet.