Topical Authority Audit: 7-Step Checklist (+ AI Citation Readiness)

A topical authority audit is a systematic evaluation of how thoroughly your website covers a specific subject area — what you've published, what's missing, how deep it goes, how it's linked, who links to it, and whether search engines and AI assistants actually treat you as the source.

That last part is new. In 2026, an audit that stops at Google rankings is checking half the scoreboard. ChatGPT, Perplexity, and AI Overviews answer a growing share of your buyers' questions, and they cite whoever is easiest to quote — not necessarily whoever ranks.

Why audit at all? Ahrefs' well-known search traffic study found that roughly 96% of pages get zero organic traffic from Google. Most of those pages aren't badly written. They're structurally isolated: covering a topic without the supporting subtopics, links, and depth that signal expertise.

Full honesty: we ran this exact audit on our own site after Google crawled 10 of our articles and indexed none of them. That specific failure has its own playbook: what to do when Google crawls a page and declines to index it. Crawled. Evaluated. Skipped. The checklist below is the one we used to work out why, and it's the same diagnostic we'd run before any plan to build topical authority. Audit first. Then treat.

This page has promised a free checklist since day one, so here it is — inline, ungated, copy-paste ready. Drop it into your notes app or project board.

• [ ] 1. Inventory: Export every indexed URL. Tag each with topic cluster, subtopic, content type, word count, last update, and 90-day traffic.
• [ ] 2. Gap analysis: Map your top 3 search competitors' subtopics against yours. Score each gap on demand, competitor coverage, and topical necessity.
• [ ] 3. Depth check: Per page: does it beat half the current SERP, answer the follow-up questions, and carry a real update from the last 12 months? Flag failures.
• [ ] 4. Internal links: Find orphan pages (zero inlinks). Confirm every cluster article links to its pillar and 2–3 siblings.
• [ ] 5. Backlink relevance: Categorize referring domains by topic. Record the on-topic share.
• [ ] 6. Prioritize: Score every issue 1–5 for impact and effort. Multiply. Work top-down.
• [ ] 7. AI citation readiness: Check answer-first structure, FAQs and schema, and quotable passages. Ask ChatGPT and Perplexity your cluster's core questions and record who gets cited.

The rest of this guide tells you exactly how to run each step, and what done looks like.

Before you can find gaps, you need to know what you have. Obvious. Almost nobody does it properly.

Pull your full URL list from Google Search Console or a crawler like Screaming Frog, then build a spreadsheet with these columns:

• URL and page title
• Primary topic cluster
• Specific subtopic (the narrow angle within the cluster)
• Content type (pillar, cluster article, FAQ, glossary, case study)
• Word count and last updated date
• Organic traffic, last 90 days

The column that trips people up is subtopic. Don't label everything "SEO" or "pricing." A post on pricing psychology and a post on freemium conversion are different subtopics under the same cluster. Get granular now — every later step depends on it.

Sort by cluster when you're done. You'll see immediately which clusters have fifteen articles and which have two. That imbalance is your first clue about where your authority is real and where it's paper-thin.

A coverage gap analysis compares what you've published against what the sites already ranking for your topics have published. Pick your top three search competitors — not business competitors, the sites that keep showing up for your target queries — and map every subtopic they cover that you don't.

You can do this manually: crawl competitor sitemaps and bucket their URLs into the same structure you built in Step 1. Tedious, but free. (Finding these gaps automatically is exactly what Inbounder's audit and topical-map features do, if you'd rather skip the spreadsheet.)

Then score each gap on three dimensions:

1. Search demand: does the subtopic have meaningful query volume?
2. Competitive coverage: do all three competitors cover it while you don't? That's a glaring hole.
3. Topical necessity: would a genuine expert have to cover this angle? Low-volume subtopics still count toward authority.

High on all three: top of the creation queue. Covered by one competitor with negligible demand: still on the list, near the bottom.

One warning. A gap analysis is only as good as the map you compare against. If you haven't defined your clusters yet, start with how to create a topical map and come back.

Having a page on a subtopic isn't the same as covering it. Thin content can drag a whole cluster down, because search engines evaluate cluster quality, not just cluster size.

For each page in your inventory, ask three blunt questions:

• Does it answer the primary query better than at least half of what currently ranks?
• Does it cover the follow-up questions a reader naturally has next?
• Has it been meaningfully updated in the last 12 months?

Any "no" gets a flag. Watch for the classic thin-content signals too: pages indexed for months that rank for nothing, sub-500-word posts on topics that demand depth, and multiple pages targeting nearly identical subtopics and cannibalizing each other.

Google's Search Quality Rater Guidelines describe the bar as experience, expertise, authoritativeness, and trust. Thin pages fail all four. So does bumping the "updated" date without changing anything substantive — freshness you didn't earn is a flag, not a fix.

Step 4: internal links. Internal linking is how search engines understand your topic relationships, and it's where most audits get lazy. Fixing broken links is 20% of the job. The real question: does your link architecture mirror your clusters? Every cluster article should link up to its pillar and across to two or three siblings. Run a crawl, filter for pages with zero inlinks, and you'll almost certainly find orphan pages you forgot existed — invisible to crawlers, contributing nothing. We break down the full pattern in internal linking for topical authority.

Step 5: backlink relevance. A hundred links from random directories move you less than fifteen from respected sites in your niche. Export your referring domains, categorize each by primary topic — a rough manual pass works — and record what share is on-topic. If your profile skews heavily off-topic, your authority signal is diluted no matter how large the raw count looks. Note the ratio in your audit sheet; it becomes a target, not just a diagnosis.

You now have gaps, thin pages, orphans, and off-topic links. Do not try to fix everything at once.

Score each item 1–5 on likely impact and 1–5 on effort, then work the quadrants. High-impact, low-effort items go first:

• Adding internal links to orphan pages (15 minutes a page, immediate crawl benefit)
• Refreshing thin pages that already rank on page two
• Writing the missing cluster articles all three competitors cover

Low-impact, high-effort items go to the backlog, not the trash. A fifty-searches-a-month subtopic that needs original research still belongs on the list. Just not this month.

Work the list in order, ship weekly, re-score monthly. This roadmap is the audit's entire payoff — a sequence, not a pile.

AI citation readiness is whether an AI assistant retrieving a few passages from your site would find a complete, quotable answer — and whether assistants cite you today. This step didn't exist in the original version of this audit. It's now the one we'd run first.

Four checks:

1. Answer-first structure. Under every question-shaped heading, is the first sentence a complete answer? LLMs lift passages, not pages. A heading followed by three paragraphs of wind-up is unquotable.
2. FAQs and schema. Does each key page carry a real FAQ section and structured data (Article, FAQPage)? These map cleanly to how assistants chunk and retrieve content.
3. The three-passage test. Pick a page. Imagine an LLM retrieved any three passages from it. Would at least one contain a self-contained answer — claim and context in the same breath? If every passage depends on the one before it, rewrite the openers.
4. Are you cited today? Ask ChatGPT and Perplexity the questions your cluster targets — ten prompts is enough for a spot-check — and record who gets cited each time. Write it down as N-of-M ("cited in 2 of 10 answers"), not a vibe. The competitor domains in those citations are your Step 2 gap list, updated for AI search.

That N-of-M habit matters. It's exactly how Inbounder's visibility module reports AI citation rates — counts with receipts, linked to the actual answers — because a citation rate you can't verify is a citation rate you'll fool yourself with.

Rankings and citations belong on the same scorecard; see measuring topical authority for how to track both. Then repeat the whole audit quarterly. Competitors publish, content ages, and AI answers reshuffle constantly. The first audit is a diagnosis. The habit is the authority.