Long-Tail Keyword Research: A 7-Step Process With Exact Filters and Real Data

Long-tail keyword research is the process of finding specific, low-volume searches your site can realistically rank for, checking what the searcher wants, and grouping those searches into pages. In practice it takes seven steps: list seed topics, expand them with Google and a keyword tool, mine your own Search Console, filter to roughly 10–1,000 monthly searches and low difficulty, check the results page, cluster close variants, and prioritize.

The numbers to aim for come from real targeting data. Of the 189,369 keywords Outrank researched and wrote articles for in the last 90 days, the median had 170 searches a month and a keyword difficulty of 14. Below you get those benchmarks as a chart, the exact filters to set in each tool, a regex for Search Console, a worked example and a checklist you can tick off.

#What volume and difficulty should a long-tail keyword have?

There is no official cutoff, so the most useful benchmark is what sites that grow actually target. Outrank picks keywords for over 4,300 websites, using live search volume and keyword difficulty. This is the distribution of those targets over the last 90 days:

What this means for your filters:

• Volume: 10 to 1,000 searches a month. Only 3.3% of targets were above 1,000. Keep a few zero-volume keywords if they are clearly what your buyers ask; tools often under-report brand-new or very niche queries.
• Difficulty: under 20 for new sites, under 30 once you rank. The scale differs by tool. Semrush, for example, labels 0–14 "very easy" and 15–29 "easy". See what keyword difficulty measures before you trust any single score.
• Word count: a hint, not a rule. Ahrefs defines long-tail keywords by low search volume, not length. If you want the full distinction, read short-tail vs long-tail keywords.

#Step 1: List 5–10 seed topics

Seeds are the broad topics your business is about: products, services, problems you solve and the audiences you serve. Pull them from your navigation, your best-selling products, sales call notes and support tickets.

For the worked example in this guide, imagine an online office furniture store. Its seeds are "standing desk", "office chair", "monitor arm", "desk converter" and "home office setup".

#Step 2: Expand each seed with Google

Google shows you real long-tail phrasing for free, in three places:

• Autocomplete. Type the seed followed by a modifier and read the suggestions: "for", "vs", "how", "why", "best", "near me", or each letter of the alphabet.
• People Also Ask. Every question box is a long-tail query, and clicking one loads more.
• Related searches at the bottom of the results page.

Here is what Autocomplete returned for two modifiers on October 19, 2026 (U.S., English):

In two minutes you have a dozen phrases with clear intent, and every one of them is something people actually type.

#Step 3: Expand with a keyword tool and set the right filters

A keyword tool adds the numbers (volume, difficulty) and finds thousands of variants Google won't show you. The filters are what turn a huge list into a shortlist.

• Ahrefs Keywords Explorer: enter the seed, open Matching terms, then set a Volume filter (for example 10–1,000) and a KD filter (max 20 for a new site). Switch to the Questions tab for question-shaped long-tail queries. Parent Topic tells you if a keyword is just a variant of a bigger one.
• Semrush Keyword Magic Tool: enter the seed, set Volume and KD % filters, and use Advanced filters → Word count to drop one- and two-word head terms. The Questions option isolates "how", "what" and "why" queries.
• Google Keyword Planner: free with a Google Ads account, but Google states that keywords with very low search volumes aren't discoverable in it. That is most of the long tail, so treat it as a sanity check, not a source.
• Outrank's free blog keyword generator: enter your site's URL and get blog keyword ideas with search volume and difficulty.

Export the filtered list to a spreadsheet. For a seed like "standing desk", expect a few hundred rows after filtering.

#Step 4: Mine your own Search Console

Your Search Console Performance report lists the long-tail queries Google already shows your pages for. It is the only source that combines real demand with your real ranking position, and it costs nothing.

Go to Search Console → Performance → Search results → Add filter → Query → Custom (regex). Search Console uses RE2 syntax, matches partially by default, and is case-sensitive unless you start the pattern with (?i). Two patterns worth saving:

• Questions: (?i)^(how|what|why|which|when|where|who|can|does|is|are|should)\b
• Queries of five words or more: ^(\S+\s+){4,}\S+$

Then sort by impressions and look at the Position column.

Worked example, using Outrank's own data. These are real question queries from outrank.so's Search Console over the 90 days to October 2026, and what each one tells us to do:

The rule of thumb: positions 5–20 are "striking distance", where one well-targeted section can move a page up. Queries in position 40+ with plenty of impressions often need their own page. We used this list to plan the refresh of our own keyword research guides, including the short-tail vs long-tail and keyword research tutorial pages.

#Step 5: Check the results page for intent and weak spots

Google every keyword on your shortlist before you commit. You are checking three things:

1. Intent and format. If the top results are product pages, a blog post won't rank, and vice versa. Match the format that already wins (guide, list, comparison, tool).
2. Weak results. A Reddit thread, a forum post, a thin page or an outdated article in the top 10 means Google is short of good answers. That is your opening.
3. What sits above the links. An AI Overview, a video carousel or a local pack pushes organic results down and lowers the clicks you can expect.

For "standing desk for tall person", for example, you would check whether shoppers see product listings (build or improve a category page) or buying guides (write an article).

#Step 6: Group close variants into one page

Long-tail research produces many near-duplicates. "Standing desk for tall person", "best standing desk for tall people" and "adjustable desk for tall people" (all real Autocomplete suggestions) are one intent, so they belong on one page. Writing three pages would split your signals and make them compete (keyword cannibalization).

The reliable test is SERP overlap: if two keywords share several of the same top-10 URLs, Google treats them as the same intent. Our guide to keyword clustering explains the method. In the worked example, the "standing desk" list collapses into page ideas like these (variants are real Autocomplete suggestions; check each pair's SERP overlap before you merge):

#Step 7: Prioritize and map keywords to pages

Score each page idea on two things: how close the searcher is to needing what you sell, and how winnable the keyword is (difficulty plus what you saw on the results page). Publish the high-relevance, easy wins first. Add each keyword to a sheet with these columns: keyword, volume, KD, intent, target URL (existing or new), status and publish date.

Then track it. After four to eight weeks, check the page in Search Console: which queries is it earning impressions for, and at what position? Feed the new queries back into Step 4. Long-tail research is a loop, not a one-off project.

If you want this loop to run without the spreadsheet, it is what Outrank automates: it researches keywords with live volume and difficulty, favours the low-difficulty long tail shown in the chart above, then writes and publishes one article a day to your site.

#Long-tail keyword research checklist

Once you have the list, turn it into content: our guide to creating long-tail keywords and writing for them covers modifier formulas and page structure, and the keyword research example walks through a complete plan from seed to final list.

#FAQ

#How do I find long-tail keywords for free?

Use Google Autocomplete with modifiers like "for", "vs" and "how", the People Also Ask box, and Related searches. Then filter your own Search Console queries with a regex for questions and five-plus-word queries. Free tools such as Ahrefs' keyword generator and Outrank's blog keyword generator add volume and difficulty.

#What is a good search volume for a long-tail keyword?

Most long-tail targets fall between 10 and 1,000 searches a month. Across 189,369 keywords Outrank targeted in 90 days, the median was 170 searches a month and 96.7% were under 1,000. Relevance matters more than volume: a 50-search keyword that matches what you sell can beat a 900-search one that doesn't.

#Should I target keywords with zero search volume?

Sometimes. Keyword tools estimate volume from samples and often show zero for niche or new queries that real buyers type, and Google says 15% of daily searches are brand new. Target a zero-volume keyword when customers clearly ask it, ideally as a section of a page that also targets a measurable keyword.

#How many long-tail keywords should one page target?

One primary long-tail keyword per page, plus its close variants and related questions as sections. If two keywords share most of the same top-10 results, they belong on the same page. If the results differ, give each its own page.

#What keyword difficulty is good for long-tail keywords?

For a new site, stay under 20 on a 0–100 scale; established sites can go up to about 30. Outrank's median target has a difficulty of 14, and 81% of its targets are under 30. Difficulty scores differ between tools, so always check the actual results page too.

#What is the difference between long-tail keyword research and regular keyword research?

Regular keyword research looks at every keyword in your market, including competitive head terms. Long-tail keyword research filters for specific, lower-volume, lower-difficulty searches with one clear intent, which are the keywords a new or growing site can rank for in months rather than years.