Long tail keywords are specific, lower-volume searches that sit far from the popular head terms. Each one is searched less often, but together they make up a large share of demand, and the people typing them usually know exactly what they want.
By Kristijan GjorgjioskiIndependent SEO consultantPublished 18 October 2026 · 6 min read
Long tail keywords are the specific searches most sites ignore because each one looks too small to matter.
That's the mistake.
A single long-tail query might bring you a handful of visits a month.
But those visitors typed exactly what they needed. And there are thousands of queries like it.
In this guide you'll learn what long-tail keywords are, why they tend to convert better, where to find them, and how to turn them into pages without creating a mess.
Long-tail keywords are specific search queries with relatively low individual search volume, usually describing a narrow need, audience or problem, that together account for a large share of all searches.
The name comes from a demand curve.
Plot every search in a market by how often it's typed.
A few broad terms sit at the tall end, the "head". After them comes a long, flat tail of specific searches, each typed rarely.
Here's an example from home services:
The long-tail version tells you almost everything: the problem, the constraint and what the person hopes to avoid.
They're not defined by length.
Long queries are usually long tail, because extra words add detail.
But a short query for a rare spare part can be long tail too. And a five-word phrase that huge numbers of people type every month is not.
They're not automatically easy either.
Some long-tail searches still return big sites with broad pages.
Lower volume often means less competition, but you still have to check.
And they're not a separate strategy from your main keywords.
Long-tail queries usually belong to the same topics as your head terms. They're the detailed questions that sit underneath them.
Here's how the two ends of the curve compare:
| Head terms | Long-tail keywords | |
|---|---|---|
| Example | "accountant" | "accountant for UK sole traders selling on Etsy" |
| Searches per query | High | Low, often reported as zero |
| What the searcher wants | Unclear, many possible goals | Specific and easy to read |
| Who ranks | Directories, big brands, national sites | Often a mix, with weaker pages |
| Page that wins | Hub, category or broad service page | Focused guide, niche service page, FAQ section |
| Typical stage | Early, browsing | Closer to a decision or a fix |
The middle of the curve, sometimes called "chunky middle" keywords, sits between the two. "Accountant for sole traders" is a good example: more specific than the head, broader than the tail.
Nobody can promise a set conversion rate for long-tail traffic, and I'd be wary of anyone who quotes one.
But the pattern I see across client accounts is consistent, and there are clear reasons for it.
Someone typing "accountant" might be a student, a job seeker or a business owner.
Someone typing "accountant for UK sole traders selling on Etsy" has a business, knows roughly what they need, and is checking who can help.
Specific words signal a specific person.
That makes it far easier for you to build a page that speaks to them.
A broad page has to serve every reader a little.
A long-tail page can serve one reader completely: the exact problem, the exact constraint, the next step that fits.
When your page matches the query that closely, the visitor doesn't need to keep searching.
That's what turns a visit into an enquiry.
Most competitors chase the head terms.
That leaves the long tail served by forum threads, thin pages and loosely related articles from big sites. A focused page often beats them, which is why my guide on how to find low competition keywords starts with specific queries.
Keyword tools estimate volume from samples.
Specific queries are where those samples are thinnest, so many long-tail searches show zero.
Competitors who sort by volume never see them.
You can, if you look in the right places.
The best long-tail sources are the ones closest to real customers.
Your Performance report lists searches your site already appears for.
Filter for queries of four words or more, or use a regex filter for question words such as how, what, can and does.
You'll usually find specific searches no tool suggested.
Read contact form messages, quote requests and call notes.
Copy out the exact phrases people use to describe their problem.
Each one is a search someone probably typed before they got in touch with you.
Autocomplete, the People Also Ask box and related searches all surface specific phrasings.
Type your topic plus "for", "without", "vs" or "can I" and see where Google takes you.
Reddit threads, trade forums and product reviews show how people describe problems in their own words, including the details they care about.
Tools help you expand a seed list fast.
Just don't drop the zero-volume rows. In niche markets they're often the valuable ones.
If you're collecting long-tail queries across many services, automating the keyword expansion saves hours without removing the judgement about which ones matter.
Finding long-tail queries is easy.
Using them well, without creating a mess, is where most sites go wrong.
"How to fix a leaking shower tray" and "leaking shower tray repair" almost certainly return the same pages.
They need one page, not two.
My rule: if two queries share most of their top results, they belong together.
The comparison of keyword clustering methods covers how to do this at scale.
A long-tail query earns its own page when it has its own intent and its own results.
Otherwise it becomes a section, an H2 or an FAQ answer on the page that owns the broader topic.
| Signal | Own page | Section on an existing page |
|---|---|---|
| Results page | Different from the broader topic | Same as the broader topic |
| Intent | Distinct (cost, comparison, audience) | A sub-question of the main intent |
| Depth needed | Hundreds of words to answer well | A paragraph or two |
Long-tail guides work best as feeders.
Each one should link to the service, product or category page it supports, and that page should link back.
This is the logic behind topic clusters.
Don't build dozens of near-identical pages that swap one word each.
That splits your signals and creates cannibalisation. The result can also look like doorway pages to Google.
One genuinely useful page beats twenty thin variations.
Two projects show the pattern well, without needing any numbers.
A bone health screening provider had traffic resting on a pair of searches. The content plan I built for it went after the long-tail questions patients actually ask.
A pharma manufacturing software firm had a set of head terms already covered. The growth came from building a long tail of buyer pain points beneath them.
In both cases the head terms stayed.
The long tail filled in the questions buyers asked before and after them.
You don't need a tool to start.
Pull your last twenty enquiries, copy out the phrases customers used, and search each one.
Wherever the results are weak, you've found a long-tail page worth building.
When you want those queries mapped against your head terms, scored and assigned to pages, that's the core of my keyword research work. For the full background on reading each query's goal, start with my explainer on what search intent is, then see how I approach keyword research for businesses like yours.
There is no fixed number. Long tail keywords are usually three words or more because extra words make a search specific, but the defining feature is low individual demand and a narrow intent. A short query for an obscure product part can be long tail, while a long phrase that thousands of people type every month is not.
Often, but not always. Fewer sites build pages for very specific searches, so the results are frequently weaker. Some long-tail queries are still dominated by large sites ranking a broad page, though. Check the results page for each one rather than assuming a long phrase means low competition.
Yes. Specific questions are exactly the kind of query where searchers want a precise, trustworthy answer, and pages that answer them clearly are good candidates to be cited. Long-tail queries that involve a decision, a local detail or an action also still send clicks, because a summary rarely completes the job.
No. Create a separate page only when a long-tail query has its own intent and its own set of results. Close variations of one question belong on the same page. Building many near-identical pages splits your signals and can look like doorway content.
About the author
Kristijan Gjorgjioski
Independent SEO consultant · 6+ years in SEO · 60+ client websites worked on
I'm an independent SEO consultant with 6+ years in SEO. I run audits, migrations and content programmes for businesses in Europe, the UK and the USA, and I build the automation behind my own research.
Book a free 30-minute call. I look at your site first and tell you plainly what I would fix, and in what order.