Long-tail keywords are winnable for a small site because the answer is specific, not because the volume is small. That distinction is the whole thing, and most coverage gets it backwards.
"How do I fix a WordPress database connection error after migrating hosts" has a real answer, and a page that gives it exactly beats a general WordPress troubleshooting guide that mentions it in passing — regardless of which site is bigger.
"WordPress" has no answer at all. It has a Wikipedia page and an official site, and nothing you write will change that.
So the strategy is not "target low-volume keywords". It is: find the questions with precise answers, and answer them precisely. Low volume is a side effect, and occasionally it is not even true.
What long-tail actually means
Two definitions circulate and they are not the same.
By length: queries of four or more words. Convenient, and wrong often enough to mislead — "best CRM" is three words and brutally competitive; "wordpress 6.8 speculative loading" is four and specific.
By specificity: queries where the searcher has already narrowed what they want. This is the useful definition, because it maps to whether a precise answer exists.
The practical test: can this question be answered definitively, in a way that would satisfy the person asking?
- "Email marketing" — no. It is a subject, not a question
- "Why are my emails going to spam" — nearly. Several causes, and a good page can cover them
- "Why do my WordPress contact form emails go to spam but not my newsletter" — yes. That has a specific answer and very few pages give it
Winnability tracks that specificity, and length only correlates with it.
Why specific queries are winnable
Four reasons, and only one is about competition volume.
The answer is verifiable. A precise question has a right answer. You can be demonstrably correct, which is much harder to compete with than being generally good.
Big sites cover them badly. A large publisher's page on the broad topic mentions your specific case in one sentence, because it is covering forty cases. A page devoted to that one case beats a sentence inside a larger page, and this holds regardless of the sites' relative authority.
Intent is unambiguous. A specific query tells you exactly what the person wants, so the format decision is made for you — the reading method is in search intent.
And the searcher is further along. Someone asking a precise question has usually already tried the obvious thing. They are closer to needing something, which is why specific traffic converts better than broad traffic at any volume.
The volume trap
The reason people avoid these queries is that tools report low or zero volume, and there are three problems with that.
Zero often means "below the reporting threshold", not zero. Tools round and suppress small numbers. A query reported as zero can still be searched steadily.
Google's own volumes are rounded averages. Its documentation states plainly: "Your search volume statistics are rounded. This means that when you get keyword ideas for multiple locations, the search volumes might not add up as you'd expect." These are estimates presented with false precision, and treating a reported 10 as meaningfully different from a reported 30 is over-reading the data.
And the variations are invisible individually. One page answering a specific question ranks for dozens of phrasings of it, most of which no tool reports separately. The traffic arrives in aggregate from queries you never targeted — which is why Search Console consistently shows pages ranking for things nobody researched.
The honest framing: you cannot forecast long-tail traffic accurately, in either direction. Which is an argument for judging these pages by whether the question is real, not by a number.
The four sources that are not tools
Because tools are worst exactly where long-tail lives.
1. Your own Search Console
The best source, and free. Queries you already appear for, in your own account.
Two searches to run:
- Filter queries containing "how", "why", "what", "can", "does" — questions you are being shown for
- Sort by impressions with low clicks — you are being shown and not chosen, which usually means your page nearly answers the question
Rare queries are omitted — Google states that "Tables in the performance reports omit rare queries to protect user privacy" — so what you can see is a sample of the tail rather than all of it. Even the sample is better than any external estimate.
2. The questions you are actually asked
Support emails, sales calls, comments, the message that starts "quick question".
These are long-tail queries in their original form, before anyone rewrote them into keyword phrasing. They are also proof of demand that no tool can give you: someone needed this badly enough to ask a human.
Keep a running list. It is the single most valuable content input a small business has and it costs nothing to collect.
3. Results-page features
"People also ask" boxes, related searches at the bottom, autocomplete suggestions.
These are Google showing you related demand, sourced from real queries. Expanding a "people also ask" box generates more, and the tree goes deep.
Read them for phrasing as much as for topics. How people actually ask is rarely how you would write it.
4. Where people ask each other
Forums, communities, Q&A sites, the comments under competitors' articles.
Look for questions that got no good answer, or an answer that is now out of date. That combination — real demand, poor supply — is the clearest opportunity signal available, and it needs no subscription.
Finding them
Six steps.
Step 1 — Start with Search Console, filtered to questions
Performance → Queries, filter for question words, last three months. Read the list rather than sorting it.
Step 2 — Add the questions you have actually been asked
From email, support, calls, comments. Write them in the asker's words, not cleaned up.
Step 3 — Expand each with the results page
Search it, read "people also ask" and the related searches, and note the phrasings you would not have chosen.
Step 4 — Test each for a definitive answer
The gate. Can this be answered in a way that satisfies the person asking? If the honest answer is "it depends on many things", it is a broad topic wearing a specific-looking phrase, and it belongs in a pillar rather than its own page.
Step 5 — Group ruthlessly
Most specific questions are variants of each other. Five phrasings of the same problem are one page — the same test as in keyword research: would one page satisfy all five searchers?
Splitting them is how cannibalisation starts — see keyword cannibalisation.
Step 6 — Answer the question in the first paragraph
Then support it.
This is the entire craft of long-tail content. Someone with a specific question has no patience for context, and a page that makes them scroll has already lost to one that did not.
It is also what makes the passage extractable when the answer is being selected rather than clicked — the argument in content formats that survive AI search.
What goes wrong
Writing a thin page per variation. Twenty near-identical short pages is closer to Google's description of scaled content abuse — "when many pages are generated for the primary purpose of manipulating search rankings and not helping users" — than to a long-tail strategy. The specificity has to be in the answer, not only in the title.
Padding a specific answer to look substantial. If the answer is four hundred words, four hundred words is the page. Google's guidance asks directly whether you are "writing to a particular word count because you've heard or read that Google has a preferred word count? (No, we don't.)"
Targeting specificity that nobody searches. Long-tail is not "invent narrow phrases". Every source above is evidence of a real question — that is the point of using them.
Ignoring the pillar. Specific pages need somewhere to belong, or they sit isolated with no internal authority reaching them. See internal linking and topic clusters.
Assuming low volume means low value. A query with three searches a month from someone about to buy is worth more than three hundred from people who are not.
What to expect
Slower and steadier than broad keywords.
- Individual pages get modest traffic. The return is the aggregate across many
- Rankings arrive sooner than for competitive terms, because there is less established competition
- They hold longer, because the answer stays correct if you keep it correct
- Traffic arrives from queries you never targeted, which is normal and is the tail working
- Conversion is better than broad traffic, because the searcher is further along
The strategic argument: thirty specific pages that each rank for a cluster of related questions is a more reliable position than one page competing for a head term against sites with resources you do not have. And it compounds, because a body of specific correct answers is exactly what makes the broad term winnable later.
Frequently asked questions
What are long-tail keywords?
Queries where the searcher has already narrowed what they want, so a specific answer exists. Length correlates with this but does not define it — "best CRM" is short and competitive, while a four-word technical query can be highly specific.
Are long-tail keywords still worth targeting?
Yes, and for a small site they are usually the only winnable ground. The advantage is that a page devoted to one specific question beats a sentence about it inside a larger page, regardless of which site is bigger.
How do I find long-tail keywords?
Search Console filtered to question words, the questions you are actually asked by customers, "people also ask" and related searches on the results page, and forums where the question got no good answer. Tools are weakest exactly here.
Why do keyword tools show zero volume for long-tail keywords?
Because tools round and suppress small numbers, so zero often means "below the reporting threshold". Google's own volumes are rounded averages by its own documentation, so precision at the low end is not there to read.
How long should a long-tail article be?
As long as the answer needs. Google states it has no preferred word count, and padding a specific answer to look substantial makes it worse for the person who came for the answer.
Should I create a separate page for every long-tail variation?
No. Most specific questions are variants of each other, and one page answers all of them. Splitting them causes your own pages to compete and, at scale, resembles what Google calls scaled content abuse.
Do long-tail keywords convert better?
Generally yes, because someone asking a precise question has usually already tried the obvious thing and is further along. Volume is a poor proxy for value at this end of the curve.
How long do long-tail pages take to rank?
Usually faster than competitive terms, because there is less established competition — and they tend to hold their position longer, since a correct specific answer stays correct if you maintain it.
What to do next
Open your sent email and find the last five times someone asked you a question about your subject. Write those questions down in their words.
At least two of them are pages nobody has written properly, and you already know the answers — which is the entire advantage a specialist has over a large publisher.
Related guides
- Keyword research that starts with your own data — where long-tail fits in the method
- Search intent: the four types and what they change — why specific queries are easier to classify
- Keyword cannibalisation — what over-splitting produces
- Keyword research without paid tools — the sources in full
- Internal linking and topic clusters — where specific pages belong
- Content formats that survive AI search — answering in the first paragraph
- Refreshing old blog posts — keeping specific answers correct
Free: The 60-Minute Email Authentication Fix
A no-fluff checklist to set up SPF, DKIM & DMARC correctly and pass Gmail & Yahoo's sender requirements.
Muhammad Basim has worked in digital marketing since 2013, focused on email deliverability and AI-assisted content production. He is the author of The Email Deliverability Playbook and The Email Copywriting Playbook.
Related Articles
Email Deliverability: Why Authenticated Emails Still Land in Spam
Email deliverability is whether your message reaches the inbox rather than the spam folder, and it is decided by three factors: technical infrastructure, list quality, and sending behaviour. Authentication is one item inside the first factor. Getting it right is necessary, and it is nowhere near sufficient. That gap explains the most common complaint in […]
Linkable Assets: What Actually Earns Links
A link is a citation, and a citation requires that someone writing about your subject needed you to make their point. That is the whole mechanism, and it is the reason most "linkable content" earns nothing. The test, before you build anything: Could a writer covering this topic finish their sentence without referencing you? If […]
Link Outreach Emails That Get Replies
Outreach fails for exactly two reasons, and only one of them gets written about. The first is that the email was not worth replying to. That is the reason every guide addresses, and the advice — personalise, be brief, offer value — is correct and insufficient. The second is that the email never arrived. Nobody […]