Keyword Research

The coffee brand ranking for terms their customers never use

A specialty coffee subscription company had a team of three people doing SEO. They'd spent 18 months writing content and building links. Their traffic was growing.

Then a consultant looked at their keyword rankings and found the problem: they were ranking on page one for "arabica bean sourcing," "single origin processing methods," and "third-wave coffee philosophy."

Their customers — people who wanted great coffee delivered to their door — searched for "coffee subscription box," "best coffee delivery UK," and "fresh roasted coffee online."

The company was ranking for terms their own baristas used. Not terms their customers used.

Their keyword research had been done from the inside of the industry, using the language of coffee professionals. But customers don't speak that language. The SEO effort was real, the traffic was real — and almost none of it was converting because it was the wrong audience.

Keyword research is about understanding how your customers speak, not how you speak.

What makes a keyword valuable

Not all keywords are worth targeting. Four factors determine whether a keyword is worth your time:

Volume without relevance is worthless. Ranking #1 for a query searched 100,000 times per month delivers nothing if the visitor has no interest in what you offer.

Relevance without volume is fine — a keyword searched 200 times per month by exactly the right audience is worth more than 10,000 monthly visitors who bounce.

The sweet spot: a keyword with enough search volume to matter, low enough difficulty that you can realistically rank, and high relevance to your audience and business goals.

The keyword research process

Step 1: Seed keywords

Seed keywords are the broad starting topics that describe what your site is about. They're the input, not the output, of keyword research.

For the coffee subscription company:

• coffee subscription
• fresh roasted coffee
• specialty coffee delivery
• coffee gift

You don't target seed keywords directly (they're too competitive) — you use them to discover the specific, targetable queries within each topic.

Step 2: Expand into long-tail keywords

Long-tail keywords are longer, more specific phrases that branch off from seed keywords. They typically have lower volume but also lower competition — and often higher intent.

"Coffee subscription" → long-tail variations:

• best coffee subscription box UK (commercial intent, specific audience)
• coffee subscription gift for dad (commercial + gifting intent)
• coffee subscription vs buying from a roaster (commercial investigation)
• how to cancel coffee subscription (navigational — low commercial value)
• specialty coffee subscription box with tasting notes (very specific, highly relevant)

Where to find long-tail keywords:

1. Google autocomplete: Start typing your seed keyword and see what Google suggests. These suggestions are based on real searches.
2. "People also ask" box: Google surfaces related questions within search results — each one is a keyword opportunity.
3. "Related searches" at the bottom: Similar to autocomplete — real searches people make after searching your keyword.
4. Competitor content: Read the top-ranking articles for your seed keyword. What subtopics do they cover? What questions do they answer? Each subtopic is a potential keyword.
5. Reddit and Quora: Search your topic. The threads your audience creates reveal the exact language and questions they use.

Using AI for keyword expansion: "I run a [business type] targeting [audience]. My seed keywords are [list]. Generate 30 long-tail keyword variations — phrases my target audience actually types into Google when looking for what I offer. Include a mix of informational, commercial, and transactional intent. Use the language my customers would use, not industry jargon."

Step 3: Research tools

Free tools give you enough data to start:

For most beginners, Google's free tools + Ubersuggest provides sufficient data. Paid tools become valuable when you're managing SEO at scale.

Step 4: Evaluate difficulty

Keyword difficulty is how hard it is to rank on page one for a given query. The fastest way to assess it: Google the keyword and study page one.

Ask yourself:

• Who's ranking? If it's all Wikipedia, Healthline, Forbes, and major brands — high difficulty.
• How long is the content? Long, well-researched guides are harder to displace than thin articles.
• How many backlinks do top results have? (Paid tools like Ahrefs or Moz show this; free tiers give limited backlink data.)
• Are there any smaller, less authoritative sites on page one? If yes, there's an opening.

A rough rule: new sites with low domain authority should target long-tail keywords with 3+ words and low to medium difficulty scores before pursuing competitive head terms.

There Are No Dumb Questions

"How much search volume is 'enough' to be worth targeting?"

It depends entirely on your business. A keyword searched 100 times per month with 80% conversion intent might be worth more than a keyword searched 10,000 times per month with low commercial relevance. For most businesses starting out: any keyword with 50+ monthly searches and clear relevance is worth targeting. For e-commerce or high-value service businesses, even 20 monthly searches for a transactional query can be worth a page.

"What if I can't find keywords with both high volume and low competition?"

That's normal — the sweet spot is rare. The strategic answer is to start with low-volume, low-competition keywords to build domain authority, then target more competitive keywords as your site grows. It's the same as a new restaurant competing for "best restaurant in London" — you don't start there. You start with "best Italian restaurant in Shoreditch" and build from there.

Keyword mapping: one keyword per page

One of the most common SEO mistakes is targeting multiple keywords with the same page, or targeting the same keyword with multiple pages.

The rule: one primary keyword per page.

Every page on your site should have one clear target keyword — the main query it's designed to rank for. This doesn't mean you ignore related terms; modern search engines understand semantic relationships and will rank a page for dozens of related queries. But you write each page with one primary intent in mind.

Keyword cannibalism: When two pages on your site target the same keyword, they compete against each other. Google struggles to decide which one to rank, and often ranks neither well. Solution: identify cannibalism, and either merge the pages into one stronger page or differentiate them to target different intents.

Building a keyword map:

A keyword map is a spreadsheet that assigns a target keyword to every page on your site:

Every page has a purpose and a target. No two pages target the same keyword.

Using AI for keyword mapping: "Here is my site structure: [list pages and URLs]. And here are my target keywords: [list 25 keywords]. Map each keyword to the most appropriate existing page. Flag any keywords that don't have a natural home (they need a new page). Flag any pages that have no keyword assigned (they need targeting or consolidation)."

The content gap: keywords your competitors rank for that you don't

One of the fastest ways to find high-value keyword opportunities is to look at what your competitors rank for — and you don't.

The competitor keyword gap process:

1. Identify 2–3 direct competitors (sites targeting the same audience)
2. Put their domain into Ahrefs or Semrush's "Organic Research" tool
3. Look at their top ranking keywords
4. Filter for keywords where their position is 1–10 and your position is 11+ (or not ranking)
5. These are keywords the market has already validated — searchers are looking for them, competitors are getting traffic — but you don't have content for them yet

Even with free tools, you can do a manual version: Google a few of your target queries, read the top-ranking articles, and list every subtopic they cover that you haven't written about. That's your content gap.

Using AI for gap analysis: "Here are the top 5 ranking articles for the query '[keyword]': [paste titles and brief summaries]. And here is an outline of my existing content on this topic: [paste your content outline]. What topics, angles, and subtopics are covered by competitors that I don't address? What questions does someone searching this query have that none of these articles fully answers?"

Tracking keyword performance over time

After you've done keyword research and published content, you need to track whether it's working.

Google Search Console for keyword tracking:

• Go to Performance → Search results
• Filter by specific pages to see which queries that page ranks for
• Track average position over time — a position trending from 15 to 8 over 3 months means the page is working
• Track impressions: even if you're not getting clicks yet, rising impressions signal that Google is beginning to rank your page for relevant queries

The 90-day tracking rhythm:

• Week 1–4: Page is published. Little to no ranking yet. Normal.
• Month 2–3: Google has crawled and indexed the page. Rankings begin — typically on pages 2–5 initially.
• Month 4–6: If the page is strong, rankings improve. You move toward page one.
• Month 6–12: Compounding. Well-targeted pages often find their stable ranking in this period.

Pages that aren't ranking after 6 months need investigation: is the keyword too competitive? Is the page content genuinely better than what already ranks? Does the page have enough internal and external links?

Using AI to interpret Search Console data: "Here are 3 months of Google Search Console data for my blog post targeting '[keyword]': [paste data — impressions, clicks, position]. The post ranks at position 18 on average. What does this tell me about what I should do to improve the ranking? What would a senior SEO specialist recommend?"

Back to the coffee brand

The coffee company had done real SEO work — 18 months of content, genuine link building, actual traffic growth — and almost none of it was converting because they'd been optimising for the language of people who already worked in specialty coffee, not the language of people who wanted great coffee delivered to their door. "Arabica bean sourcing" is how a barista thinks. "Best coffee subscription box UK" is how a customer searches. The fix wasn't more content or more links — it was switching from industry language to customer language, finding the exact phrases real buyers type into Google. That shift, captured in a keyword research document, reoriented every subsequent piece of content toward an audience that could actually buy. The SEO effort had been real; only the language was wrong.

Key takeaways

• Keyword research is about how your customers speak — not how you speak. The language gap between industry insiders and customers is where most SEO strategies fail.
• Four factors determine keyword value: volume, difficulty, intent, and business relevance. High volume without relevance is worthless.
• Long-tail keywords are where new sites win. Specific, 3+ word phrases have lower competition and often higher purchase intent than broad head terms.
• One primary keyword per page. Keyword cannibalism — multiple pages competing for the same query — undermines rankings on both pages.
• Competitors are a keyword goldmine. The queries they rank for (and you don't) are opportunities the market has already validated.