Topical Authority Cluster Map Prompt & Generator

What a topical authority cluster map is

A topical cluster is a group of pages built around one broad subject: a single pillar page that covers the topic comprehensively, surrounded by cluster articles that each answer one narrower question in depth and link back to the pillar. A cluster map is the planning document for that structure — the table that decides which pages exist, what each one is for, and how they link together, before anyone writes a word.

Search engines reward the structure because it demonstrates coverage: a site that answers the main question and every adjacent question, with clear internal links between them, is a more credible result than a site with one good article and nothing around it. But the map is also an editorial discipline tool. It forces three decisions marketers otherwise make by accident: which subtopics deserve their own page, which are sections of an existing page, and which gaps in coverage matter most. A good map prevents the two classic failure modes — a sprawl of thin near-duplicate pages competing with each other, and a single overloaded article trying to answer twelve different intents at once.

Why a prompt, and why the incumbents don't give you one

Plenty of marketing content explains cluster strategy. monday.com describes topical planning inside its prompt framework; SEO guides walk through building a topical map with a chatbot step by step. What almost none of them publish is the actual prompt — the reusable instruction that produces the map. The AirOps prompt library comes closest, with a Claude prompt that returns a pillar plus ten subtopics with key points. It stops at prose lists: no search intent per page, no internal-linking plan, no merge decisions, no use for your existing content.

The generator on this page closes those gaps. It takes your existing content inventory as an input, so the map plans around what you have instead of starting from a blank universe. And it returns the output the way planning actually happens — as a table you can paste into a spreadsheet, sort by priority, and assign.

Who this is for

Content strategists and marketing leads planning a quarter of content; SEO consultants building a coverage case for a client; founders who have published twenty scattered articles and need to know what structure they actually have. If you have no content yet, the map still works — the inventory field is optional — but its real strength shows when it can see your existing pages and plan merges, updates, and gaps against them.

How it works

1. Enter your pillar topic and audience. Be specific about scope: "email deliverability for Shopify stores" produces a usable map; "email marketing" produces a directory.
2. Paste your existing content inventory. Titles and URLs, one per line — blog posts, guides, landing pages, anything topically related. The generator uses it to mark which cluster ideas you already cover.
3. Set the cluster size. Ten to fifteen cluster articles is the range working maps use; fewer under-covers the topic, more produces thin pages.
4. Generate and copy the table. The output is spreadsheet-ready. Columns: page title, search intent, audience task, format, relationship to pillar, internal links in and out, priority, and a merge flag.
5. Review the combine flags before approving anything. Any row the model marked COMBINE is a topic it judges too close to another page to stand alone. Those calls shape your whole structure — check them yourself.
6. Work the map in priority order. High-priority gaps first, merges second, new cluster articles third.

The fields in the map, explained

Search intent (informational, commercial, transactional) decides the page's job: teach, compare, or convert. Mixing intents on one page is the most common reason cluster articles underperform. Audience task states what the reader is trying to get done, in their words — it keeps titles honest. Relationship to pillar records whether the page supports the pillar directly, supports another cluster page, or updates existing content. Internal links in/out is the column incumbents never provide: every cluster page needs a planned link to the pillar and to its two or three nearest siblings, decided at planning time rather than improvised after publication. Priority ranks gaps high, medium, or low based on distance from your product and how completely the topic is already covered.

The rule that makes this generator trustworthy

The prompt below contains a hard instruction most cluster prompts lack: the model must never invent search volume, keyword difficulty, or traffic data. Language models do not have access to keyword databases, but when asked for a content plan they will happily attach plausible numbers to every row — numbers with no source that then drive real budget decisions. This generator forbids it. Prioritization is argued from intent, audience fit, and coverage gaps, all of which the model can legitimately reason about from your inputs. If you need volume data, get it from a keyword tool and paste it in as source material; the map will use it, labeled as yours.

THE PROMPT TEMPLATE

Use this when: planning a new topic area, or auditing a scattered set of existing articles into a structure. Replace: every [bracketed] field. The inventory may be left as "none" for a new site. Works in: ChatGPT (GPT-6.1 / GPT-4o), Claude Sonnet 5.5, Gemini 4 Argon.

ROLE
You are an SEO content strategist building a topical authority cluster map for a content team.

CONTEXT
- Pillar topic: [BROAD TOPIC, SCOPED — E.G. "EMAIL DELIVERABILITY FOR SHOPIFY STORES"]
- Target audience: [WHO THE CONTENT SERVES AND THEIR LEVEL]
- Product or service this content ultimately supports: [PRODUCT + ONE LINE]
- Existing content inventory (titles + URLs, one per line — or "none"): [PASTE INVENTORY]
- Number of cluster articles to plan: [10–15]

INSTRUCTIONS
1. Define the pillar page first: its title, the single intent it serves, and the 5–7 subtopics it must cover itself rather than delegate to cluster pages.
2. Plan the cluster articles as a table. For each: working title, search intent (informational / commercial / transactional), audience task (what the reader is trying to get done), format (guide, comparison, checklist, tutorial, FAQ page), relationship to the pillar, internal links (which pages it links to, which link to it), priority (high / medium / low), and status (NEW, UPDATE of an existing page, or MERGE).
3. Check every proposed cluster article against the existing content inventory. If an inventory page already covers the topic, mark the row UPDATE or MERGE with that page's URL — never plan a duplicate.
4. If two proposed cluster articles would answer substantially the same reader question, combine them into one row and flag it "COMBINE — do not split." One page per distinct reader purpose.
5. Prioritize by closeness to the product and completeness of the coverage gap. State the reason for each high priority in the audience-task column.
6. After the table, list: coverage gaps you deliberately left out and why, and the 3 pages you would build first.

NEGATIVE CONSTRAINTS
- Never invent search volume, keyword difficulty, traffic estimates, or competitor data. If a metric was not supplied in the context, it does not appear in the map.
- No keyword stuffing: do not create separate rows for keyword variations of the same question ("best X", "top X", "X guide" are one page).
- Do not use AI clichés: never write "delve", "tapestry", "game-changer", "unlock", or "elevate".
- Do not invent statistics, quotes, or facts about the topic. The map plans pages; it does not assert claims the pages will make.

OUTPUT FORMAT
A markdown table with these columns, in this order: # | Page title | Intent | Audience task | Format | Relationship to pillar | Links to | Linked from | Priority | Status.
Followed by two short lists: "Deliberate gaps" and "Build first."

Best model: GPT-6.1 — holds a ten-column table schema most consistently across 15+ rows; use Gemini 4 Argon when your existing-content inventory runs to hundreds of pages.

ONE WORKED EXAMPLE

Filled inputs (fictional business, for illustration): Pillar topic — "invoice payment for freelancers." Audience — solo freelancers and micro-studios. Product — Billfold, an invoicing app. Existing inventory — two posts: "How to Write an Invoice (Template)" and "What to Do When a Client Won't Pay." Cluster size — 10.

Filled prompt (excerpt):

CONTEXT
- Pillar topic: invoice payment for freelancers
- Target audience: solo freelancers and micro-studios who invoice clients directly
- Product or service this content ultimately supports: Billfold, an invoicing app for freelance designers
- Existing content inventory: How to Write an Invoice (Template) — /blog/how-to-write-an-invoice; What to Do When a Client Won't Pay — /blog/client-wont-pay
- Number of cluster articles to plan: 10

Output excerpt (first rows of the returned table):

GUARDRAILS

• No fabricated data. The map must never contain a search volume, difficulty score, or traffic figure that you did not supply. If the output includes one anyway, delete the number and distrust the prioritization around it. Mark any metric you add yourself with its source before the map is shared.
• A map is a plan, not research. The model reasons from your inputs; it has not checked what competitors rank for or what your audience actually searches. Validate the top three priorities against a keyword tool or your own Search Console data before committing writing time.
• Human review gate. Merge and priority decisions change site structure and internal linking. A named strategist reviews every COMBINE flag and every UPDATE row before the map becomes assignments. [SOURCE NEEDED] discipline applies to the pages the map produces: briefs generated from this map must carry their own source material.

INTERNAL LINKS

• Content Marketing hub — all three content marketing generators.
• Pillar Article Draft Generator — draft the pages this map prioritizes.
• AEO Answer Block Generator — add quotable answer blocks to each finished page.
• AI Writing & Copywriting Prompts guide — prompt fundamentals behind this cluster.
• Prompt Generator — the general-purpose generator on the homepage.

---