Long-Tail Keywords: Where a New Site Should Actually Start

Every new site owner wants to rank for the two-word term. "Email marketing." "Project management." "Running shoes."

And every new site owner is competing there against domains with fifteen years of links, full-time content teams, and a brand people already search by name. You will not win that this year. You may not win it ever, and that's fine, because it was never the route.

The route is the specific question nobody has bothered to answer properly.

The short version

A long-tail keyword is a longer, more specific query with lower search volume and lower competition. "Why do my emails go to spam after changing DNS providers" instead of "email deliverability."

Three reasons they're the right starting point:

1. You can actually rank, which means traffic now
2. Intent is sharper, so they convert better
3. Coverage compounds — twenty specific pages build the authority that eventually wins the head term

The head term isn't a target. It's a reward for having covered its subject thoroughly.

What counts as long-tail

Less about word count than about specificity, though they correlate.

Head term: "email deliverability" — broad, high volume, brutal competition, mixed intent.

Mid-tail: "improve email deliverability" — narrower, moderate volume, still competitive.

Long-tail: "why do my emails go to spam after changing DNS providers" — highly specific, low volume, often barely contested, and unmistakable intent.

The practical marker: can you tell exactly what problem the searcher has? If yes, it's long-tail. "Email deliverability" could be someone with a crisis, a student, or a job hunter. The DNS question could only be someone with a specific problem you can solve today.

Why they work better on a young site

You can rank. This is the whole argument. A page ranking fifteenth for a 20,000-volume term gets nothing. A page ranking first for a 40-volume term gets 40 chances. Small numbers you actually receive beat big numbers you don't.

Intent is sharper. Someone typing a fifteen-word question has thought about their problem and is looking for a solution now. Someone typing two words might be doing anything. Long-tail visitors convert better — often dramatically — because the query itself filters for readiness.

Volume aggregates. No single long-tail query is impressive. Fifty of them, each pulling 30 to 80 visits a month, is meaningful traffic — and it's traffic that doesn't vanish when one head-term ranking shifts.

It compounds into authority. Twenty focused pages on email deliverability, properly interlinked, demonstrate command of the subject. That's what eventually makes the head term reachable. You don't attack it — you earn it. The cluster model.

How much volume is too little?

The honest answer is that this is the wrong question, and it trips people up.

Zero reported volume doesn't mean zero searches. Keyword tools report what they have data for, and a large share of Google queries are unique — never searched before in that exact form. Tools can't see those.

Some of the most valuable pages target queries every tool called zero, because they answer a real question that gets asked in a hundred slightly different phrasings, none individually big enough to register.

Better questions than "what's the volume":

• Do real people actually ask this? (Check Reddit, forums, your own inbox.)
• Is the intent clear enough that a visitor would be worth having?
• Does it connect to something I sell or want to be known for?

A query with 20 estimated searches that matches your service exactly is worth more than one with 2,000 that doesn't.

Where to find them

Autocomplete, forked. Type your topic and add a letter — "email deliverability a," then b, then c. Each fork produces a different set of real queries.

People Also Ask, expanded. Click a question and more appear. One seed query can produce twenty specific questions, each a potential page.

Reddit and forums. The single best source, because people describe problems in their own words before they've learned the jargon. "My emails are going to junk" not "inbox placement optimisation." That vocabulary gap is exactly where uncompeted queries live, since everyone else is writing for the professional phrasing.

Search Console. If you have any history at all, you're already getting impressions for long-tail queries you never targeted. Validated demand with confirmed relevance. How to mine it.

Your own client questions. The questions people pay you to answer are, by definition, questions with commercial value.

Question modifiers. Take your topic and prepend: why, how, what, when, which, does, can, should, is. Then add situational modifiers: after, without, for small business, on WordPress, when using Gmail.

One page or many?

The main way people waste long-tail work.

Close variants belong on one page. "How to set up DMARC," "DMARC setup guide," and "how do I configure DMARC" are the same question. One page, three phrasings.

Genuinely different questions get different pages. "What is DMARC" and "DMARC not working after migration" are different intents.

The test: would the ideal answer to both queries be the same page? Same page, one post. Different page, two posts.

Split close variants and you get cannibalisation — your own pages competing, signals divided, both underperforming. Spotting and fixing it.

When do you graduate to head terms?

Not on a schedule. On evidence.

You're ready when:

You have coverage. Fifteen or twenty solid pages on the subject, interlinked. The pillar for a head term should sit at the centre of a cluster, not float alone.

You're already ranking for mid-tail terms in the same territory. That's the signal Google has begun to associate you with the subject.

You've earned some links, usually to your best long-tail pages rather than to anything you targeted.

You can genuinely produce something better than what currently ranks. If you can't articulate what your version does that the top three don't, you'll publish and place eleventh.

The pattern that works: publish long-tail supporting posts first, build the cluster, then write the pillar targeting the head term — with fifteen relevant internal links already pointing at it from day one. That's a completely different starting position from a lone post launched into an established field.

The catch worth knowing

Long-tail queries are more exposed to AI answers than head terms are.

A specific question with a specific answer is exactly what an AI Overview handles well, and someone who gets their answer in the results may never visit. That's real and it's changed the arithmetic.

Two adjustments rather than a change of strategy:

Favour long-tail queries where the answer isn't a single fact. "Why did my emails go to spam after a migration" needs a diagnostic process, not a sentence. Those hold up better than "what does SMTP 550 mean."

Add something a summary can't replace — your own diagnostic order, a screenshot from real work, a specific case. Which formats survive.

Frequently asked questions

What counts as a long-tail keyword?
A longer, more specific query with lower search volume and less competition. Length correlates but specificity is the real marker — the test is whether you can tell exactly what problem the searcher has. "Email deliverability" could be anyone; "why do my emails go to spam after changing DNS providers" is one person with one problem you can solve.

How much search volume is too little?
This is usually the wrong question. Keyword tools report zero volume for queries they simply have no data on, and a large share of Google searches are unique phrasings never seen before. Better tests: do real people ask this on forums or in your inbox, is the intent clear, and does it connect to what you sell? A 20-search query matching your service beats a 2,000-search query that doesn't.

When should I target head terms?
When you have fifteen to twenty interlinked pages covering the subject, you're already ranking for mid-tail terms in that territory, you've earned some links, and you can articulate what your version would do better than the current top three. The working pattern is to publish the supporting posts first, then write the head-term pillar with internal links already pointing at it.

What to do next

Pick your subject and open Reddit. Search it, and read how people phrase their problems — not the answers, the questions.

Write down ten of them in the exact words they used. That's your next ten posts, and most of them will have no competition worth the name, because everyone else is writing for the professional phrasing nobody actually types.

Free: The SEO audit checklist.

---

Related guides

• Keyword research without expensive tools — the wider process
• How to read a SERP — confirming a long-tail query is winnable
• Internal linking and topic clusters — turning coverage into authority
• 6 content formats that still get clicks — protecting against AI answers