Do zero-volume keywords actually drive traffic? Ahrefs' own data says mostly not.

A tip that has survived a decade of repeating

Search "zero volume keywords" and the advice barely varies: your keyword tool shows 0 or "not enough data" because search engines round small numbers down, long-tail and voice-search phrasing slip past the sampling, and a keyword showing zero volume can secretly deliver steady traffic once you rank for it. It gets presented as a loophole, the tool is wrong and you get to exploit the gap. It is repeated often enough across blogs and tool vendors that it reads as settled fact rather than a claim anyone actually tested against real data.

What happened when Ahrefs tested its own advice

Ahrefs' Joshua Hardwick ran the actual test in October 2022, on Ahrefs' own blog. He pulled the top-five ranking positions for keywords Ahrefs' own tool reported at 10 or fewer monthly U.S. searches, then checked what those pages actually earned in Google Search Console over a 28-day window. The result: the average keyword produced about 11.3 real impressions a month, and fewer than 1% of them broke 100 impressions. The zero-or-near-zero estimate matched what actually happened. There was no hidden reservoir of traffic the tool had failed to count.

Hardwick's own conclusion narrowed the advice rather than repeating it: target a reported-zero keyword if it is an early sign of a trend nobody has search history for yet, or if it is a highly specific phrase you already know converts, from sales calls, support tickets, or a niche you understand from experience, regardless of what a tool says about it. That is a much colder pitch than "free undiscovered traffic," and it is the version backed by an actual measurement rather than a hook.

Flag this plainly: it is one experiment, on one site, from 2022, and Hardwick said so himself, calling it "a (very) small experiment" to be read "with a grain of salt." Nobody has published a larger, independent version of it since.

Why the number was never a precise count to begin with

Ahrefs' CMO Tim Soulo went further in March 2024, arguing in "There's No Such Thing as Accurate Search Volume" that the entire premise of a single precise number was shaky before anyone started arguing about zero. Google Keyword Planner numbers are rounded annual averages assigned to volume buckets shared across groups of similar keywords, not a count kept for one exact phrase. What shows on your screen as "10" or "0" next to a specific string was frequently never measuring that string alone.

Soulo also pointed back at Ahrefs' own older research, by his account roughly seven years old and worth reading as dated rather than current, which found that a page ranking for one target keyword typically ranks for around 1,000 related keywords at the same time. If a single page earns traffic across a thousand different queries, the volume number attached to any one of them was never a clean prediction of what that page would earn. Treating a "0" as meaningfully different from a "10" assumes a precision the underlying data never had.

The gap no tool will ever close

There is a separate, permanent version of the same problem, and it comes from Google directly. At Search Central Live NYC, reported by Search Engine Journal on 28 March 2025, Google's John Mueller confirmed that 15% of all queries Google sees are new every day, a figure he said Google has recalculated repeatedly and found essentially unchanged over time. Mueller said he expected large language models and AI systems to push that share higher, and it hasn't: "it's still hovering around that number."

That 15% is structurally invisible to any keyword tool, not because the tool is bad but because a query has to exist before anything can measure it. It sits underneath a second, newer blind spot we covered in keyword research for AI search: AI engines split a single prompt into sub-queries through what Google calls query fan-out, and nobody, including Google, publishes that list either. One invisible zone predates AI by decades. A second one arrived with it. Neither is a zero-volume keyword waiting to be found. Both sit permanently outside what any tool reports, which is a structural fact about how search works, not a temporary gap the next keyword tool update fixes.

What this changes about the research

None of this means ignore small numbers. It means stop treating a 0 or a 10 as a signal to chase and start treating coverage as the strategy. The buyer question behind a reported-zero phrase is usually already covered, in different wording, by a broader page built to answer the whole topic, which is the same reason publishing more separate pages tends to underperform one page that actually covers the ground. Multiplying thin pages to catch every zero-volume variant is also how keyword cannibalization starts, competing against your own coverage instead of a competitor's.

Ahrefs' own two exceptions are worth keeping: an early trend with no search history yet, and a specific, high-intent phrase you already know converts from a channel other than search. Both are judgment calls made with outside information, not something a keyword tool surfaces on its own. Everything else in the reported-zero bucket is better spent building the page broad enough to catch it by accident, including phrasing nobody has typed yet and the fan-out sub-queries no dashboard will ever show you. It is the same principle behind clustering by intent instead of listing strings, which is how our own Keyword Strategy is built.

What we could not verify

No study, ours or anyone else's, has measured whether zero-volume keyword targeting affects AI citation rate specifically. The coverage argument above is a reasoned extension of the fan-out and fresh-query problem, not a tested result, and it is stated as inference rather than fact for that reason. Hardwick's 11.3-impression figure comes from one small, self-reported experiment on Ahrefs' own blog in 2022, has not been independently replicated at any scale, and Ahrefs itself called the sample size out. No comprehensive, cross-site dataset exists showing what share of reported-zero keywords outperform their estimate in general; treat the Ahrefs number as directional for one site, not a universal rate. Whether Google Keyword Planner's bucket-and-rounding methodology has changed since Soulo described it in March 2024 could not be independently confirmed beyond Ahrefs' own account of it.

Sources

Ahrefs, Joshua Hardwick, "Should You Target Zero-Volume Keywords? (It Depends)", 21 October 2022. Ahrefs, Tim Soulo, "There's No Such Thing as 'Accurate' Search Volume", 18 March 2024. Search Engine Journal's report on Google's John Mueller at Search Central Live NYC, "Google Revisits 15% Unseen Queries Statistic In Context Of AI Search", 28 March 2025. Google's own description of query fan-out in AI features and your website.

Common questions