Long tail keywords are specific search queries with low search volume, usually three or more words, such as "trail running shoes for wide feet" instead of "running shoes". Each one brings few searches, but together they make up most searches, and their precise wording tells you what the searcher wants, so they are easier to rank for.
Key takeaways
- A keyword is long tail because its volume is low compared with its topic's head terms, not because of its word count.
- Keywords with fewer than 10 monthly searches make up almost 93% of Ahrefs' US keyword database.
- AI search widens the tail: Google says early AI Mode users ask questions two to three times longer than classic searches, and fan-out splits each question into sub-queries.
- Find them in Search Console, autocomplete, People Also Ask, community questions and the headings of pages that already rank.
- Target one page per intent, not one page per phrase.
What are long tail keywords?
Long tail keywords are the many specific, low-volume queries at the far end of a topic's search demand curve. Sort every query about running shoes by monthly searches and a few head terms ("running shoes") take huge volume, a middle band takes moderate volume, and thousands of specific queries ("how often should you replace running shoes") take a handful each. That long, flat end of the curve is the long tail, a name borrowed from Chris Anderson's 2004 Wired essay The Long Tail, about niche products that each sell little but together add up to a large market. Search demand works the same way: keywords with fewer than 10 searches a month make up almost 93% of Ahrefs' US keyword database (Ahrefs).
Long tail research is one part of keyword research, and two points trip people up:
- Length is a symptom, not the definition. Most long-tail queries have three or more words because each word narrows the need, but as Ahrefs puts it, "it is not the length of a keyword that makes it a long-tail".
- Long tail is relative. There is no fixed cut-off. Where the head term gets 1,000 searches a month, a 40-search query is long tail; in a broad consumer topic, a 900-search query can be too.
Variant long tail vs standalone long tail
Ahrefs separates "supporting" long-tail keywords from "topical" ones (and adds a "conversational" tail of AI prompts); we use plainer labels for the first two:
- Variant long tail is a rarer way of asking for the same thing as a bigger query, such as "wide toe box trail running shoes" next to "best trail running shoes for wide feet". The page that ranks for the main query usually ranks for its variants too.
- Standalone long tail is a sub-topic the main page doesn't answer, such as "how to clean trail running shoes". It needs its own section or page.
Long tail vs short tail keywords
The volume bands below are illustrative; they vary by niche.
Short tail (head) | Middle | Long tail | |
|---|---|---|---|
Example | running shoes | trail running shoes | trail running shoes for wide feet |
Typical monthly searches | tens of thousands+ | hundreds to low thousands | zero to a few hundred |
Intent | vague or mixed | partly clear | clear |
Competition | very high | medium | usually lower |
Page that tends to rank | homepage, category, major guide | category or comparison | specific article, product page or section |
Why do long tail keywords matter more in AI search?
Long tail keywords matter more in AI search because AI answers are assembled from many specific sub-queries, and the questions people ask AI features are longer.
People ask longer questions. Google reported that early users of AI Mode were asking questions "two or three times the length of traditional search queries" (Google, July 2025). And the tail never runs out: in 2019 Google said that 15 percent of the queries it sees every day are ones it hasn't seen before (Google, October 2019).
One question becomes many searches. Google says AI Overviews and AI Mode may use a "query fan-out" technique, issuing multiple related searches across subtopics and data sources (Google Search Central). Its AI optimization guide shows "how to fix a lawn that's full of weeds" fanning out into "best herbicides for lawns", "remove weeds without chemicals" and "how to prevent weeds in lawn". Each sub-query is a long-tail keyword.
Citations come from beyond the top 10. In Ahrefs' March 2026 study of 863,000 keyword results pages, 37.9% of URLs cited in AI Overviews also appeared in the top 10 results for that query (Ahrefs). Ahrefs suggests fan-out may explain the gap: AI Overviews seem to lean on pages that rank for the fan-out sub-queries. For a newer site, ranking for specific sub-questions is a realistic route into AI answers before it can rank for the head term.
What AI search does not change
The same Google guide warns against reading this as "add more keyword variations": its AI systems understand synonyms and meaning, so "you don't have to worry that you don't have enough 'long-tail' keywords". Treat long tail keywords as a map of the questions your page must answer, each in its own headed section, not as strings to repeat.
Why long tail keywords are easier to rank for
Long tail keywords are easier to rank for because few pages are built for the exact need, and the pages that rank often answer it only in passing.
- Weaker competition. For "running shoes" you face brands and big retailers. For "how often should you replace running shoes if you run 20 miles a week" you may face a forum thread and one paragraph in a general guide.
- Clearer intent. The longer query says whether the searcher wants to buy, compare or learn, so you can match the format; our search intent guide shows how to read it.
- They add up. One good page usually ranks for dozens of related variants.
- They build topical depth. Pages that answer one topic's long-tail questions, linked to a pillar page, form topic clusters.
- Low reported volume is not zero demand. A question your customers keep asking can deserve a section even when a tool shows no volume.
How to find long tail keywords
The fastest way to find long tail keywords is to start with the queries you already get impressions for, then widen the net. Seven sources, roughly in order of reliability:
1. Your own Search Console queries
In Google Search Console, open the Performance report, select Queries and add a query filter set to Custom (regex):
what|how|when|whyshows question queries; Google's Search Central team suggests this filter for spotting queries your content should answer (Google Search Central blog).^(\S+\s+){4,}\S+$shows queries of five words or more. This one is our own pattern, not Google's: check it in Search Console on a few queries you know before relying on it.
As a rule of thumb, queries with impressions where your average position sits between about 8 and 30 are the cheapest wins: Google already connects your page to them, but the page doesn't answer them fully.
2. Autocomplete, People Also Ask and related searches
Type the head term followed by question words ("running shoes why", "running shoes how") or single letters and note the suggestions. Expand People Also Ask boxes, which load more questions as you open them, and check the related searches.
3. Community and customer questions
Reddit, Quora, forums, reviews, support tickets and sales calls give you customers' exact words, including problems no keyword tool lists yet.
4. Keyword tools for volume
Size the ideas with a keyword tool. Google Keyword Planner is free with a Google Ads account; our guide to Google Keyword Planner for SEO shows how to filter it for longer phrases. Treat low numbers as estimates, not proof that nobody searches.
5. The headings of pages that already rank
Read the H2s and H3s of the top five results for your head term. Sub-questions most of them cover are expected; ones only a single page covers, or none do, are your openings.
6. AI answers
Ask ChatGPT, Gemini or Google's AI Mode the broad question and note how the answer is organized and which sources it cites. It won't show the sub-queries it ran, but its sections are a reasonable proxy for the sub-topics, and the cited pages show who you compete with.
7. Modifiers
Add qualifiers to any head term: audience (for beginners), constraint (under $300, for small spaces), comparison (vs, alternative to), problem (why is my, not working), place or language.
How to do this in RankTera
In RankTera's KW Research, type a keyword and pick a country. RankTera reads the top Google pages (you choose up to 5 in the Select SERP Results picker, with the first 5 pre-selected, and can add your own URL, file or text), then lists the phrases those pages use under their H2–H6 headings with each one's average count (AVG COUNT), plus the pages' average word count. Track the keywords you choose in the Rank Tracker, one country per keyword: it checks your Google position and, weekly, whether ChatGPT and Gemini cite your domain.
What we found: the terms top pages use for "long tail keywords"
We ran "long tail keywords" through RankTera's KW Research on October 15, 2026, using US Google results. The five top-ranking pages it analysed (semrush.com, ahrefs.com, brightedge.com, backlinko.com and blog.hubspot.com) averaged 1,845 words, and the run found 85 keywords. The most-used terms:
Term | Average count per page |
|---|---|
long-tail keywords | 9 |
long tail keywords | 6 |
search volume | 4 |
long-tail keyword | 3 |
keyword research, keyword difficulty | 2 each |
search queries, AI systems | 2 each |
The ranking pages explain the concept through search volume and keyword difficulty, and "AI systems" appears about twice per page even on this definitional topic. In our separate analysis of six other top Google results that day, the median page had 2,190 words, and only two of the six used a table.
Long tail keyword examples
These long tail keyword examples show the intent each query signals and the page that usually fits.
Head term | Long tail keyword | Intent | Content that fits |
|---|---|---|---|
running shoes | trail running shoes for wide feet | Commercial | Buyer's guide or filtered category |
payroll software | payroll software for a 10-person company | Commercial | Use-case or comparison page |
payroll software | how to run payroll for contractors in the UK | Informational | Step-by-step guide |
rank tracker | how to track keyword rankings in another country | Informational | Tutorial |
banana bread | why is my banana bread dense in the middle | Informational | Troubleshooting article |
dentist | emergency dentist open on Sunday in Leeds | Transactional (local) | Location page with opening hours |
hotels in Lisbon | quiet hotel in Alfama with parking | Commercial | Shortlist or comparison |
standing desk | standing desk for a small apartment under $300 | Commercial | Shortlist with a comparison table |
How to target long tail keywords without cannibalization
To target long tail keywords without cannibalization, map each keyword to one page by intent: variants join the page that already serves that intent, and only a different need gets a new page. Google's AI optimization guide says that creating "separate content for every possible variation of how people might search", fan-out queries included, primarily to manipulate rankings or AI responses violates its scaled content abuse policy, and that more pages don't make a site higher quality. Splitting variants across pages also makes your own URLs compete for one query, which is keyword cannibalization.
The one-page-per-intent test
- Search both keywords in a private window, in your target country.
- Compare the top 10. If several of the same URLs rank for both (our rule of thumb is four or more), they share an intent and belong on one page.
- If the results barely overlap, give the second keyword its own page, or its own section if the page stays focused.
- If the results show a different page type (videos, product pages, a calculator), build that format, not another article.
Where each long-tail keyword goes
Situation | What to do |
|---|---|
Variant of a keyword a page already targets | Use its wording in a heading, sentence or FAQ on that page |
Sub-question the page should answer | Add an H2 or H3 that asks it and answer in the first two sentences |
Distinct need with its own results | New page, linked to and from the pillar page |
Two of your pages rank for the same query | Merge into the stronger page and redirect the other |
Customer question with almost no volume | FAQ entry or help article |
Whatever the placement, use the searcher's wording where it reads naturally and never repeat a phrase to hit a count: Google matches meaning, not exact strings.
Worked example: turning 10 long-tail keywords into 4 pages
This example is illustrative; the volumes are not real data. A new dog-training site can't rank for "puppy training" yet. It collects ten long-tail keywords from Search Console, autocomplete and client emails, checks which pairs share first-page results, and maps them:
Keyword | Monthly searches (example) | Intent | Page |
|---|---|---|---|
how to stop a puppy biting | 880 | Informational | A: puppy biting guide |
how to stop puppy biting when excited | 170 | Informational | A (variant) |
puppy biting ankles when walking | 140 | Informational | A (H2 section) |
how long does the puppy biting phase last | 390 | Informational | A (H2 section) |
crate training a puppy at night | 720 | Informational | B: night crate training |
puppy crying in crate at night first week | 260 | Informational | B (variant) |
how long can a puppy stay in a crate | 590 | Informational | B (H2 section) |
best crate size for a labrador puppy | 210 | Commercial | C: crate buying guide |
wire vs plastic dog crate | 170 | Commercial | C (H2 section) |
puppy training classes vs online course | 90 | Commercial | D: comparison page |
Four pages now cover 3,620 example searches a month (A: 1,580, B: 1,570, C: 380, D: 90) instead of ten thin pages competing with each other. The site tracks each page's main keyword plus a variant, and if "how long can a puppy stay in a crate" later returns mostly crate-schedule pages, it moves to its own page.
Frequently asked questions
What is a long-tail keyword example?
"How to stop puppy biting when excited" is a long-tail keyword: it is specific, gets far fewer searches than "puppy training", and states exactly what the searcher wants. "Payroll software for a 10-person company" and "why is my banana bread dense in the middle" are two more.
How many words is a long tail keyword?
A long tail keyword usually has three or more words, but word count is not the rule. A keyword is long tail when its search volume is low compared with the head terms of its topic. Short queries can be long tail in small niches, and some long questions get high volume, so judge by volume and specificity.
Are long tail keywords still worth targeting with AI Overviews and AI Mode?
Yes, more than before. Google says AI Overviews and AI Mode may fan one question out into several related searches, and many cited pages rank outside the top 10 for the original query. Pages that answer the specific sub-questions well get more chances to be retrieved and cited. Write for the questions, not for exact phrase counts.
Should I create a separate page for each long tail keyword?
No. Create one page per search intent. If two long-tail keywords return mostly the same Google results, one page should target both. A new page is justified only when the results show a different need or page type. Separate pages for every variation split your signals and, at scale, can breach Google's scaled content abuse policy.
Can a beginner rank for long tail keywords?
Yes. Long tail keywords are the usual starting point for new sites because competition is lighter and the intent is clear. Pick questions your audience really asks, answer each one better than the current results, link the pages under a pillar page, and track their positions to see which need work.
Find and track your long tail keywords in one place
Long tail keywords pay off when each page answers one intent completely and you can see which pages rank for which questions. RankTera's KW Research shows the terms the top pages use under their headings, and the Rank Tracker follows your Google positions and ChatGPT/Gemini citations for every keyword. Start your 14-day free tier — no card needed; it includes 3 keyword researches, 3 AI articles and weekly tracking for 300 keywords. Plans are on the pricing page.