Keyword mapping: one keyword, one page, written down

You have the keyword list. Six hundred rows, volumes, difficulty scores.

Now comes the step almost nobody writes down: deciding which URL answers which query — and recording that decision somewhere a colleague can read it. That record is the keyword map. Finding the keywords is a separate job, covered in our guide to keyword research; this page is only about what you do with the list afterwards, inside the wider SEO guide.

Skip it and the same thing happens to every site. Two posts chase one query. A writer briefs a page that already exists.

What keyword mapping is — and the rule it enforces

Keyword mapping assigns every keyword cluster exactly one owner URL, and writes that assignment into a document you consult before building anything.

The rule it enforces is the one you have heard as "one keyword, one page". More precisely: one intent, one page. Ten phrasings of the same question are one cluster with one owner, not ten pages.

Be honest about where that rule comes from. It is an operating convention SEOs adopted, not something Google published. Google's Search Advocate John Mueller has pushed back on the panic around overlap directly: "If you have 3 different pages appearing in the same search result, that doesn't seem problematic to me just because it's 'more than 1'," and "pages aren't duplicates just because they happen to appear in the same search results page" (Search Engine Journal, 22 September 2025).

So the map is not defensive magic. It is editorial discipline, and the case for it is in Google's own quality guidance, which asks whether you are "producing lots of content on many different topics in hopes that some of it might perform well in search results" and whether the content "provide[s] substantial value when compared to other pages in search results" (Google Search Central, last updated 5 October 2026). A map is how you answer "no" to the first question and "yes" to the second, on purpose.

One distinction worth keeping clean: a keyword map assigns queries to URLs, while content mapping assigns content formats to funnel stages. Different artefacts, both useful, neither a substitute for the other.

The register decides whether a page exists.

Cluster by intent, not by how the words look

This is where most maps go wrong. Clustering tools group by string similarity, so they put "seo audit" and "seo audit tool" together because they share two words — when one wants an article and the other wants software.

Group by the job the searcher wants done. Four tests, in order:

• Would one page satisfy both searchers? If a single page can genuinely answer both queries without padding, they are one cluster.
• Do the results look the same? Search both terms. If the top results are the same kind of page — both guides, both product listings — the intent matches. If one returns tools and the other returns explainers, split them. Our search intent guide covers the classification properly.
• What page type does the answer need? A definition wants a guide. A comparison wants a comparison page. A calculation wants a tool, not 1,200 words about the formula.
• Is this a variant or a question of its own? Plurals, synonyms and word order are variants — same cluster. A genuinely narrower question is its own row, which is how long-tail keywords earn pages.

Only after the clusters exist do you assign owners. One cluster, one URL, and the secondary terms sit on that same page as sections and subheadings — which is what on-page SEO then works with.

The seven columns a keyword map needs

A three-column map gets abandoned because it cannot answer questions. These seven are the working minimum.

Add difficulty if it helps you sequence work — see keyword difficulty for what those scores mean. Resist adding more columns. Each one is another thing to leave stale.

Keep it in a shared spreadsheet, one row per cluster, filterable by status.

Five signs your site needs one

You rarely decide to build a map. You discover you needed one:

• Two pages trade positions for one query. Filter the Search Console Performance report to a single query, then switch the table to the Pages dimension (Search Console Help). Two URLs under one query, week after week, means neither owns it.
• A page ranks for a term another page targets. The map says URL A owns it; Google shows URL B. One of them is in the wrong place.
• A brief was written without checking anything. If a writer can start a page without consulting a list, you do not have a map — you have a keyword export.
• Nobody can name the owner. Ask which URL is supposed to rank for your most commercial term. A pause is the answer.
• New pages keep sounding familiar. The fifth post that re-explains the same concept is the map's absence showing up in the editorial calendar.

Several of these surface in a routine pass through our SEO audit checklist.

How we run ours: the cannibalisation register

This is Digital Hangover's own practice, on this website, not a template we hand out. Our keyword map's enforcement half is a short document we call the cannibalisation register, and it has three columns.

One rule makes it work: no page gets created until the register has been checked. If the entity already has an owner, the new question becomes a section inside that page, or it does not get made.

The Instagram row is the one that earns its keep. "Instagram marketing" is owned by /blog/instagram-marketing-strategy/, so the obvious-looking slug /blog/instagram-marketing/ is recorded as never-build. Without that row, somebody proposes it roughly once a quarter, because it looks like a gap.

We will not show you a traffic chart for the register — we have not run this site against a version of itself without one. What it changed operationally: the "does this already exist?" conversation takes ten seconds instead of a meeting. How the register sits inside the folder structure is in our write-up of this site's architecture.

When two pages already compete

A map prevents overlap. It does not remove overlap you already shipped. For the diagnosis, go to keyword cannibalisation — this is only the assignment decision at the end of it.

The decision rule: write a one-sentence description of each page's job. If a reader would not see those as two different jobs, they are one page.

Two notes, both from Google. Use a permanent redirect when you consolidate: Google treats it as a signal that the target "should be canonical", and says to use one "when you're sure that the redirect won't be reverted" (Google Search Central, last updated 14 April 2026).

If you cannot redirect — say both pages must stay for users — a canonical tag is the softer option. Google lists "to consolidate signals for similar or duplicate pages" among its reasons for specifying one, and notes that without one, "Google will identify which version of the URL is objectively the best version to show to users in Search" (Google Search Central, last updated 10 July 2026). Deciding yourself is better than letting it be decided.

After any consolidation, update the map the same day. A redirected URL left in the Owner column is how the duplicate comes back.

Keeping the map alive — the part everyone skips

A map written once and never consulted is just a spreadsheet. Four habits keep it real:

1. Make it a gate, not a reference. The brief template starts with "owner URL from the map". No row, no brief. This is the only habit that actually matters.
2. Update it in the same change as the page. New URL goes live, status moves to Live, in the same sitting. Deferred bookkeeping never happens.
3. Review quarterly against reality. Pull your top queries, check the ranking URL against the Owner column, and investigate every mismatch. Those mismatches are the real audit.
4. Log the never-builds. Every rejected page idea gets a row with status "Never build" and one line of reasoning, so the same idea does not come back in six months wearing a new slug.

Two linked jobs sit next to this. Each owner URL needs links pointing at it, or it is invisible however well mapped — see internal linking and orphan pages. And if your map starts generating hundreds of near-identical rows from a template, you are in programmatic SEO territory, which has its own quality rules.

What a working map gives you next quarter

The payoff is not a ranking jump. It is that content decisions stop being arguments.

A keyword arrives. The map says it has an owner, so it becomes a section. Or it says the row is empty, so a brief gets written with the intent and page type already decided. Writers stop duplicating each other. One URL owns one answer, which means there is one version of your answer for anyone to find — a reader, a crawler, or an AI assistant quoting you.

Expect the normal organic timeline, not a switch: meaningful movement in roughly three to six months, compounding after that. The map does not speed that up. It stops you spending the three months competing with yourself.

Build the clusters, then the rows, then the gate. Where clusters sit in a wider plan: SEO strategy and topic clusters.