A long-tail keyword is a specific search query that relatively few people type, such as “emergency locksmith Walthamstow Sunday” rather than “locksmith”. Each one brings modest traffic on its own, but together they make up a large share of searches, and the people typing them usually know exactly what they want.
How long-tail keywords work
The name comes from a chart, not from the number of words. If you plot every query in a market by how often it is searched, a few popular terms sit at the top and a vast number of rarely searched ones trail away in a long, flat tail. A head keyword such as “accountant” sits at the top; “accountant for Airbnb hosts in London” sits in the tail.
Long-tail queries tend to be longer because being specific takes more words, but not always. A rarely searched two-word product code is long-tail too. What defines the tail is low demand for each individual query combined with a precise need.
They come in a few recognisable kinds:
- Refinements of a head term by place, type, price or audience: “vegan wedding cake Brighton”.
- Questions Covered in more detail under question keywords: “do I need planning permission for a garden office”.
- Problem descriptions “boiler losing pressure overnight”.
- Comparisons and specifics “Xero or FreeAgent for a sole trader”.
Why it matters
For a small UK business, the head terms in its market are often dominated by national brands, directories and comparison sites. The long tail is where a smaller firm can compete, because a well-written page that answers a specific question precisely can outrank a large site’s generic page. The search intent behind long-tail queries is usually clearer too: someone who searched “commercial EPC for a shop in Kent” is far closer to an enquiry than someone who searched “EPC”.
Long-tail demand also matters more as search habits change. Conversational searches, voice queries and the questions people type into AI-assisted search tend to be long and specific, and pages that answer specific questions clearly suit that style of search well.
Common mistakes
- Dismissing terms because a tool shows zero or tiny volume. Tools under-report rare queries, and Search Console often reveals real searches the tools miss.
- Creating a separate thin page for every variation (“plumber Hackney”, “plumber in Hackney”, “Hackney plumbers”), which leads to keyword cannibalisation and doorway pages.
- Forcing the exact phrase into awkward sentences, when Google understands close variants perfectly well.
- Chasing long-tail terms that have no connection to what you sell.
How to act on it
Start with your own data. In Google Search Console’s Performance report, filter queries by page and look at the long, specific searches already bringing impressions. They show the language customers use and the questions your pages only half answer. Add the questions customers ask on the phone and by email, and the People Also Ask boxes for your main terms.
Then group the queries by intent. Variations that want the same answer belong on one page, which should answer them all thoroughly, often with a section or FAQ for each angle. Genuinely different needs get their own page. My free keyword clustering tool helps sort a list into those groups, and mapping each group to a single page is the core of the keyword research service I run. Revisit the Search Console data every few months, because new phrasings appear as customers’ needs and vocabulary change.