Long tail keyword statistics at a glance
| Category | Statistic | Source |
|---|---|---|
| Distribution | Keywords with fewer than 10 monthly searches make up almost 93% of Ahrefs' US keyword database, 2.3 billion of them | Ahrefs, 2026 |
| Distribution | Just under 18,000 US keywords get more than 100,000 searches a month | Ahrefs, 2026 |
| Distribution | More than 95% of conversational long tail keywords have no measurable search volume | Ahrefs, 2026 |
| Impressions | 81.1% of informational organic impressions in Seer's 53-account cohort come from one and two word queries; 4.9% come from queries of four words or more | Seer Interactive, 2026 (computed) |
| Conversion | Paid search conversion climbs from 0.17% at one word to 1.94% at six words, then dips | NP Digital, May 2025 |
| Conversion | Queries of three words or more convert 6.1x better than one and two word queries | NP Digital, 2025 (computed) |
| Query length | The average Google AI Mode query runs 7.22 words against 4.0 words for a traditional search | Semrush, 2025 |
| Query length | Search-enabled ChatGPT prompts grew from 4.7 words to 8.7 words in twelve months | Semrush, 2026 |
| Query length | 15% of the queries Google sees every day have never been searched before, published by Google in 2017 and restated in 2025 | Google, April 2017, Search Engine Journal, March 2025 |
| AI search | 8% of one or two word searches produce an AI summary, against 53% of searches of ten words or more | Pew Research Center, 2025 |
| AI search | Comparison queries trigger an AI Overview 95.4% of the time; question queries 85.9% | Seer Interactive, 2026 |
| Zero volume | Zero and low volume keywords ranking top five drove 11.3 impressions each per month in Ahrefs' own test | Ahrefs, 2022 |
| Zero volume | It takes roughly 8,850 zero-volume keywords at #1 to match one 100,000-volume keyword at #1 | Ahrefs and First Page Sage (computed) |
How is search demand distributed across keywords?
Search demand follows a cliff, not a curve: 93% of the keywords in Ahrefs' US database get fewer than 10 searches a month, while just under 18,000 keywords get more than 100,000.
| Statistic | Source |
|---|---|
| Keywords with fewer than 10 searches per month account for almost 93% of Ahrefs' US keyword database, or 2.3 billion keywords. | Ahrefs, 2026 |
| Just under 18,000 keywords in the US have search volumes over 100,000 searches per month. | Ahrefs, 2026 |
| Ahrefs has discovered 110 billion keywords in total and tracks the 28.7 billion most popular ones. | Ahrefs, 2026 |
| 94.74% of keywords get 10 monthly searches or fewer, and 0.0008% get more than 100,000. | Ahrefs, 2026 |
| More than 95% of conversational long tail keywords have no measurable search volume at all. | Ahrefs, 2026 |
| 96.55% of the 14 billion pages in Ahrefs' Content Explorer index get zero traffic from Google, and another 1.94% get between one and ten visits a month. | Ahrefs, 2023 |
| The average page ranking #1 also ranks in the top 10 for nearly 1,000 other relevant keywords, from a sample of 3 million queries. | Ahrefs, 2017 |
| About 95% of all pages have zero backlinks, from an analysis of 11.8 million search results. | Backlinko, April 2025 |
Two of those are dated: the 96.55% zero-traffic figure is from 2023 and the thousand-keywords-per-page study is from 2017. The 2017 one still matters most, because it says long tail traffic is mostly a by-product of ranking for something else.
Every keyword tool sells you the top of this distribution and hides the bottom, because the bottom is unsellable: no volume estimate, no difficulty score, no forecast. If your keyword plan only contains terms your tool could price, you planned around 7% of the keyword universe.
Do zero-volume keywords get searched at all?
Zero-volume keywords do get searched, but barely. Ahrefs ranked its own blog in the top five for low and zero volume terms in 2022 and measured 11.3 impressions per keyword each month. Fewer than 1% of those keywords cleared 100 impressions.
| Statistic | Source |
|---|---|
| Low and zero volume keywords where the Ahrefs blog ranked top five drove 11.3 impressions each per month; fewer than 1% of them drove more than 100 impressions. | Ahrefs, 2022 |
| It takes roughly 8,850 zero-volume keywords ranked at #1 to match the monthly clicks of one 100,000-volume keyword ranked at #1. | Ahrefs 2022 and First Page Sage 2025 (computed) |
| 15% of the queries Google handles every day are new and have never been searched before. Google published the figure in April 2017 and John Mueller restated it in March 2025, saying it has held steady despite large language models. | Google, April 2017, Search Engine Journal, March 2025 |
| The #1 organic result earns a 39.8% click-through rate; the #10 result earns 1.6%. | First Page Sage, May 2025 |
| On mobile, Sistrix measured 28.5% CTR at position 1 and 2.5% at position 10 across more than 80 million keywords. | Sistrix, 2020 |
The Ahrefs test is small and its author says so. I use it because it is the only public experiment that measures what a zero-volume keyword pays out.
Google's 15% figure is the most misused number in long tail marketing. It does not mean 15% of search traffic is unclaimed. It means Google fields roughly a billion novel strings a day, most of which nobody will type again. You cannot target a query that happens once.
How long is the average search query in 2026?
The average traditional Google query is 4.0 words and the average Google AI Mode query is 7.22 words, from Semrush's 2025 clickstream study. Search-enabled ChatGPT prompts reached 8.7 words in early 2026.
| Interface | Average query length | Period |
|---|---|---|
| Google traditional search | 4.0 words | May to July 2025 |
| Google AI Mode | 7.22 words | May to July 2025 |
| ChatGPT with search enabled | 4.7 words | Jan to Feb 2025 |
| ChatGPT with search enabled | 8.7 words | Jan to Feb 2026 |
| ChatGPT without search | 24.9 words | Jan to Feb 2025 |
| ChatGPT without search | 13.5 words | Jan to Feb 2026 |
Source: Semrush AI Mode study (about 69 million US desktop sessions) and Semrush ChatGPT clickstream analysis (US clickstream, October 2024 to February 2026).
| Statistic | Source |
|---|---|
| The average AI Mode query is almost twice the length of a traditional Google query, 7.22 words against 4.0. | Semrush, 2025 |
| Between 92% and 94% of AI Mode searches ended without a click to an external site. | Semrush, 2025 |
| Search-enabled ChatGPT prompts grew 85% longer in twelve months, from 4.7 words to 8.7 words. | Semrush, 2026 (computed) |
| Between 65% and 85% of ChatGPT prompts could not be matched to any traditional search term in Semrush's database of more than 27 billion keywords. | Semrush, 2026 |
| The share of ChatGPT prompts written in traditional search language rose from 18.9% to 34.9% between October 2025 and February 2026. | Semrush, 2026 |
| Queries of three to four words were the most common length bracket in Semrush's clickstream study of 609,809 US searches, with one and two word queries second. | Semrush, 2022 |
| The typical Google voice search answer is only 29 words long, from 10,000 Google Home results. | Backlinko, 2018 |
Query length is moving in two directions at once. Chat prompts are getting shorter as people learn a tool with a search box behind it does not need a paragraph, and search queries are getting longer as people learn Google will accept one. The two interfaces are converging around seven or eight words, which is long tail territory by any definition used since 2005. The 2022 Semrush brackets and the 2018 voice study are dated, and they are the baseline that shift is measured against.
This convergence is the real story on the page. For twenty years the advice was to target long tail keywords because they are easier to rank for. The new reason is simpler: the average query is becoming a long tail query. You no longer have to go looking for the tail. It arrives in your Search Console report on its own.
Do long tail keywords convert better than head terms?
Long tail keywords convert far better than head terms. NP Digital measured 0.17% conversion on one-word paid search queries and 1.94% on six-word queries in May 2025. Queries of three words or more convert 6.1 times better than one and two word queries.
| Query length | Conversion rate |
|---|---|
| 1 word | 0.17% |
| 2 words | 0.35% |
| 3 words | 1.02% |
| 4 words | 1.61% |
| 5 words | 1.58% |
| 6 words | 1.94% |
| 7 or more words | 1.83% |
Source: NP Digital, May 2025, a study of 40 companies running paid search, branded terms excluded. This is vendor data from an agency's own client accounts, so read the shape, not the absolute numbers.
| Statistic | Source |
|---|---|
| Queries of three words or more convert 6.1 times better than one and two word queries. | NP Digital, 2025 (computed) |
| Organic search traffic converts at 4.9% and paid search at 5.4%, from 110 million sessions and 5 million conversions. | Ruler Analytics, May 2026 |
| Average Google Ads cost per click is $5.42 and average cost per lead is $66.69. | WordStream and LocaliQ, 2026 |
| Google holds 91.1% of worldwide search engine market share, so Google-based query-length findings cover most of the market. | StatCounter, August 2026 |
The dip at seven words is the detail most summaries drop. Conversion rises until a query names a product or a problem, then flattens and slightly reverses, because queries that long usually belong to researchers rather than buyers. Six words is the commercial peak here.
I would not use these percentages as targets. A 1.94% conversion rate inside an agency's own paid accounts says nothing about your funnel. The ratio is what travels: the words a head-term campaign is proud of convert at roughly one sixth the rate of the words it ignores. If you have never segmented a paid conversion report by query word count, that is an afternoon with a high hit rate.
How do AI Overviews treat long queries differently?
AI Overviews target long queries hardest. Pew Research Center found AI summaries on 8% of one and two word Google searches in March 2025, against 53% of ten-word-plus searches.
| Statistic | Source |
|---|---|
| 8% of one or two word searches produced an AI summary, against 53% of searches of ten words or more, from 68,879 searches by 900 US adults in March 2025. | Pew Research Center, July 2025 |
| 18% of all Google searches in the Pew sample produced an AI summary. | Pew Research Center, 2025 |
| 60% of question-format searches triggered an AI summary, against 36% of searches written as a full sentence with a noun and a verb. | Pew Research Center, 2025 |
| Users clicked a result on 8% of visits to pages carrying an AI summary, against 15% of visits without one. | Pew Research Center, 2025 |
| Among informational queries, AI Overviews appeared on 27.3% of one-word queries, 34.8% of two-word, 31.1% of three-word and 35.5% of four-word-plus queries. | Seer Interactive, 2026 |
| 81.1% of informational organic impressions in Seer's 53-account cohort came from one and two word queries; only 4.9% came from queries of four words or more. | Seer Interactive, 2026 (computed) |
| AI Overview prevalence by intent: 36% of informational queries, 8% of commercial and 5% of transactional. | Seer Interactive, 2026 |
| Comparison queries trigger an AI Overview 95.4% of the time and question-format queries 85.9%, from a small base of 280 comparison and 1,413 question queries. | Seer Interactive, 2026 |
| Brands cited inside an AI Overview earn 120% more organic clicks per impression than brands that are not cited. | Seer Interactive, 2026 |
| Organic CTR on AI Overview queries recovered from a 1.3% floor in December 2025 to 2.4% in February 2026. | Seer Interactive, 2026 |
Pew and Seer disagree about how strongly query length drives AI Overview triggering. Pew sees a steep gradient, 8% to 53%, across all searches; Seer sees a flat one, 27.3% to 35.5%, but only inside informational queries, where an overview is likely however the query is phrased. Both hold: length predicts AI Overviews mostly because long queries are usually informational. For how the two engines split traffic, see the ChatGPT vs Google search statistics page.
The uncomfortable part is that long tail keywords and AI Overviews want the same queries. The phrases that convert best in paid search are the ones most likely to be answered above the organic results. Being cited still pays 120% more clicks per impression, but the old promise of cheap traffic from specific questions has expired.
What does long tail traffic cost compared with head terms?
Long tail traffic is cheap per keyword and expensive per click. A zero-volume keyword ranked first returns about 4.5 clicks a month: Ahrefs' 11.3 impressions times First Page Sage's 39.8% CTR.
| Statistic | Source |
|---|---|
| A zero-volume keyword ranked #1 returns about 4.5 clicks a month (11.3 impressions x 39.8% CTR). | Ahrefs 2022 and First Page Sage 2025 (computed) |
| For every 1,000 US Google searches, 360 clicks reach the open web; in the EU the figure is 374. | SparkToro and Datos, July 2024 |
| 58.5% of US Google searches ended in zero clicks, and 59.7% of EU searches. | SparkToro and Datos, 2024 |
| Buying the equivalent traffic is priced at $5.42 per click and $66.69 per lead in Google Ads. | WordStream and LocaliQ, 2026 |
The SparkToro clickstream figures are from July 2024 and marked dated, since the click rate has fallen further since. They set the ceiling for this arithmetic: any long tail strategy competes for a shrinking pool of open-web clicks. For the full channel cost comparison, the organic vs paid search statistics page has CPC and retainer data side by side.
Here is the trade I would make. Long tail keywords are not a traffic strategy, because 8,850 of them at position 1 equals one good head term. They are a margin strategy: you publish a page for a term you can win, it picks up adjacent phrases as Ahrefs' 2017 study predicted, and those phrases convert six times better. The volume never impresses a board. The pipeline does.
How we calculated the original numbers
Four figures on this page do not appear anywhere else. Here is the arithmetic, so you can check it or swap the inputs.
- 6.1x, the conversion premium of long tail queries. NP Digital's May 2025 data gives 1.02%, 1.61%, 1.58%, 1.94% and 1.83% for queries of three to seven-plus words, a mean of 1.596%. One and two word queries convert at 0.17% and 0.35%, a mean of 0.26%. 1.596 divided by 0.26 is 6.14. NP Digital does not publish the query volume behind each bucket, so this is an unweighted mean of the seven buckets, not a volume-weighted average. Branded terms are excluded from the study, so it is not a brand-search artefact.
- 4.9% of informational impressions come from four-word-plus queries. Seer Interactive publishes organic impression counts next to its AI Overview rates for the 53 accounts it tracks: 25,042,440 one-word, 32,598,962 two-word, 9,935,297 three-word and 3,513,670 four-or-more, totalling 71,090,369. The four-word-plus bucket is 4.94% of that, and one and two word queries together are 81.08%. Most keywords are long tail; most impressions are not.
- 8,850 zero-volume keywords to match one head term. Ahrefs' 2022 test gives 11.3 impressions a month per low or zero volume keyword ranking top five. At First Page Sage's 39.8% CTR for position 1 that is 4.4974 clicks. A 100,000-volume keyword at position 1 returns 39,800 clicks. 39,800 divided by 4.4974 is 8,850. Both inputs flatter the long tail, since 39.8% assumes a page with no AI Overview.
- 85% growth in search-enabled ChatGPT prompt length. Semrush puts search-enabled ChatGPT prompts at 4.7 words in January and February 2025 and 8.7 words in the same months of 2026. 8.7 divided by 4.7 is 1.851, an increase of 85.1% in twelve months.
What the 2026 numbers say
The distribution is a cliff and the tools show you the top of it. With 93% of Ahrefs' US keyword database under 10 searches a month, a plan built from filtered keyword exports rests on a rounding error of total demand. Treat a keyword list as a topic map, not a traffic forecast.
Long tail converts and does not scale. Six times the conversion rate on queries of three words or more is replicable. Needing 8,850 zero-volume keywords to match one head term is why it cannot carry a business alone. The pattern that works is the one Ahrefs measured in 2017: rank for something findable, collect the adjacent phrases as a by-product.
Query length is rising and Google is answering the long ones itself. AI Mode queries run 7.22 words and Pew found AI summaries on 53% of ten-word-plus searches. The tail is getting bigger and less clickable at once. Being cited inside the answer, worth 120% more clicks per impression in Seer's data, is the version of long tail SEO that still pays in 2027.
FAQ
What percentage of keywords are long tail?
Almost 93% of the keywords in Ahrefs' US database get fewer than 10 searches a month, or 2.3 billion keywords (Ahrefs, 2026). Ahrefs also reports that 94.74% of keywords globally get 10 monthly searches or fewer, and 0.0008% get more than 100,000. Long tail dominates keyword count, not search volume.
Do long tail keywords convert better than short ones?
Yes, by about six to one. NP Digital's May 2025 study of 40 companies running paid search measured 0.17% conversion on one-word queries and 1.94% on six-word queries, so queries of three words or more convert 6.1 times better than one and two word queries. Conversion peaks at six words and dips slightly beyond that.
How many words is the average Google search in 2026?
The average traditional Google query is 4.0 words and the average Google AI Mode query is 7.22 words, from Semrush's analysis of about 69 million US desktop sessions between May and July 2025. Search-enabled ChatGPT prompts averaged 8.7 words in early 2026.
Are zero search volume keywords worth targeting?
Only as a by-product. Ahrefs' 2022 test found low and zero volume keywords where its blog ranked in the top five delivered 11.3 impressions each per month, and fewer than 1% cleared 100 impressions. It takes roughly 8,850 of them at position 1 to match one 100,000-volume keyword at position 1.
Do AI Overviews appear more often on long queries?
Yes. Pew Research Center found AI summaries on 8% of one and two word Google searches and 53% of searches of ten words or more, from 68,879 searches in March 2025. Seer Interactive's 2026 data shows a flatter gradient inside informational queries, 27.3% at one word to 35.5% at four words or more.
How many Google searches are brand new each day?
15% of the queries Google handles every day have never been searched before. Google published that figure on its blog in April 2017, and John Mueller restated it at Search Central Live NYC in March 2025, saying the share has stayed near 15% despite large language models changing how people phrase things.
Sources
- Ahrefs, Long-Tail Keywords: What They Are and How to Get Search Traffic From Them (2026)
- Ahrefs, SEO Statistics (2026)
- Ahrefs, 96.55% of Content Gets No Traffic From Google (2023)
- Ahrefs, Should You Target Zero-Volume Keywords? It Depends (2022)
- Ahrefs, How Many Keywords Can You Rank For With One Page? (May 2017)
- Semrush, Google AI Mode's Early Adoption and SEO Impact (2025)
- Semrush, ChatGPT Traffic Analysis: Insights From 17 Months of Clickstream Data (2026)
- Semrush, Zero-Clicks Study (2023, data from May 2022)
- Pew Research Center, Google users are less likely to click on links when an AI summary appears (July 2025)
- Search Engine Journal, Google Revisits 15% Unseen Queries Statistic In Context Of AI Search (March 2025)
- Seer Interactive, AIO Impact on Google CTR: 2026 Update (April 2026)
- NP Digital via Neil Patel, Conversion Rate by Keyword Length (May 2025)
- First Page Sage, Google Click-Through Rates by Ranking Position (May 2025)
- Sistrix, Why Almost Everything You Knew About Google CTR Is No Longer Valid (July 2020, updated July 2025)
- Backlinko, We Analyzed 11.8 Million Google Search Results (April 2025)
- Backlinko, Voice Search SEO Study (February 2018)
- SparkToro and Datos, 2024 Zero-Click Search Study (July 2024)
- Ruler Analytics, Conversion Rate Benchmarks (May 2026)
- WordStream and LocaliQ, Search Advertising Benchmarks (2026)
- StatCounter, Search Engine Market Share Worldwide (August 2026)
- Google, Our latest quality improvements for Search (April 2017)
Methodology. Numbers were collected in September 2026 and checked on the original publisher's page, not in statistics roundups. Vendor claims without a stated sample or method were excluded, and vendor data that is included is labelled as vendor data in the sentence. Figures older than 24 months are flagged as dated in the text: the Ahrefs studies from 2017, 2022 and 2023, the Semrush query-length brackets (2022), the Backlinko voice search study (2018), the Sistrix CTR study (2020) and the SparkToro clickstream study (July 2024). Google's "15% of queries are new" line first appeared on Google's blog in April 2017 and was restated by John Mueller in March 2025, so both years are given. The NP Digital page returns a 403 to automated requests; its figures were verified through a text mirror of the same URL on 20 September 2026. Last updated 20 September 2026.