Topical Authority vs Domain Authority for AI Citations: What the Data Says

For about a decade, Domain Authority has been the metric the SEO industry uses to predict competitive outcomes. Higher DA means better rankings. Better rankings mean more traffic. Clean logic. Measurable signal.

Then AI search showed up and mostly ignored it.

Wellows’ research found topical authority (measured as the breadth of keywords a domain ranks for within a topic) correlates with AI citation at r=0.41, the strongest individual predictor they measured. Domain Authority? It explains less than 4% of citation variance (r²=0.032).

I’m Gabriel Bertolo. I run Radiant Elephant, a boutique SEO and GEO agency out of Northampton, Massachusetts. I’ve been building content architectures for clients for over 13 years, and the topical authority finding lines up with everything I’ve watched happen across our client portfolio since AI Overviews started rolling out. The clients with deep, structured content clusters are getting cited. The clients with high DA but thin coverage on any given topic are not.

I ranked topical authority alongside 14 other proven tactics in the complete evidence review covering every major GEO study. This article goes deeper on the mechanism, the fan-out data, and how to actually build the clusters.

Why DA fails and topical depth wins

The reason maps directly to how retrieval-augmented generation works under the hood.

When someone asks ChatGPT or Perplexity a question, the system doesn’t just run one search. It breaks the query into multiple sub-queries (fan-out) and searches for the best-matching content across each sub-question.

Kevin Indig’s research quantified this: 89.6% of ChatGPT queries generate two or more follow-up searches, and 32.9% of cited pages appeared only in fan-out query SERPs. A third of all cited content would never have been found through traditional single-keyword targeting. The AI found it because it was searching for answers to sub-questions the user never explicitly asked.

Sites with comprehensive topic coverage give the AI more citation opportunities across those fan-out sub-queries. A domain with one great page about a topic has one shot at getting cited. A domain with a pillar page plus fifteen supporting articles has sixteen shots, each matching a different sub-query.

Domain Authority doesn’t capture this. A site with DA 80 and one thin page on a topic loses to a site with DA 30 and deep, structured coverage. Growth Memo’s March 2026 analysis showed the compounding effect: the top 10 domains in a given topic cluster take 46% of all ChatGPT citations for that topic. The top 30 take 67%. Once you establish deep coverage, the advantage compounds. Late entrants face a steep climb.

We proved this in practice. In our national GEO case study, the client had a DR of 21 at kickoff. They were competing against Fortune 500 companies with DR 70-80+. We built comprehensive pillar-and-spoke content clusters targeting five distinct stakeholder roles across their industry. The client outranked FDA.gov (DR 92) in AI search. That’s not supposed to happen in a DA-driven world. But DA isn’t driving AI citation. Topical depth is.

The fan-out data that proves clusters work

Surfer SEO’s study of 10,000 keywords gave us the most direct evidence.

Pages ranking for fan-out sub-queries are 161% more likely to be cited in AI Overviews. Fifty-one percent of all AI Overview citations go to pages ranking for both the main query and at least one fan-out query. Under 20% go to pages ranking only for the main query.

Marie Haynes was among the first to document Google’s query fan-out mechanism, noting in March 2025 that queries have “ultimately turned into conversations.” BrightEdge’s data reinforces this: 82.5% of AI citations go to deep, nested topic pages. Only 0.5% cite homepages.

The pattern is consistent. AI systems are not looking for the most authoritative domain in a general sense. They’re looking for the best answer to each specific sub-question. And the best way to have the best answer to many sub-questions is to have content that actually covers many sub-questions.

That’s what topical authority is. Comprehensive coverage of a subject, organized so each piece stands alone as an answer to a specific question while connecting to the broader topic through internal linking.

This connects directly to why answer-first content structure matters at the page level. Topical authority creates coverage across sub-queries. Answer-first structure ensures each page is extractable when the AI finds it. And it’s why AI Overviews now pull 62% of citations from outside the top 10. The fan-out mechanism reaches deeper into the index, rewarding depth over ranking position.

How we build topic clusters for clients at Radiant Elephant

This is the process I follow for every content cluster we build, whether the client is a professional services firm here in the Pioneer Valley or a national B2B manufacturer.

Start with a pillar page. Comprehensive piece, 2,500-5,000 words, covering your core topic. Not a surface-level overview. A genuine deep dive that addresses multiple user intents, includes specific data, and references the sub-topics your spoke pages cover in detail. The pillar is the central node in the cluster.

Build 15-20 spoke pages around that pillar. Each addresses a specific subtopic, question, or angle. Each targets a different query the AI might generate during fan-out. “What is [topic]?” is one spoke. “How does [topic] compare to [alternative]?” is another. “[Topic] for [specific industry]” is a third.

Use bidirectional internal linking. Every spoke links to the pillar, and the pillar links to every spoke. Spoke pages link to 2-3 related spokes. No orphan pages. Every page in the cluster connects to at least 3 others. This creates a content graph AI systems can traverse, building confidence that your domain covers the full topic.

Target the fan-out queries specifically. Don’t just optimize for your primary keyword. Figure out what second, third, and fourth questions a user would ask after the first one, and build content for those. Tools like iPullRank’s Qforia (Gemini-powered) and Wellows’ Query Fan-Out Generator can simulate the sub-queries AI systems generate.

Prioritize depth over breadth across topics. A domain that covers three topics with deep cluster architecture outperforms one that covers thirty topics with one article each. AI rewards concentration. This is what I tell every prospective client: stop publishing thin content about everything. Go deep on the topics that drive revenue, and own those topics completely.

If you’re building authority from scratch, start with content clusters and topical authority.

Five implementation examples

Example 1: Marketing agency building a “Local SEO” topic cluster

Pillar page: “The Complete Guide to Local SEO in 2026” (3,500 words)

Spoke pages (18 total): “How to Optimize Your Google Business Profile (Step-by-Step).” “Local SEO Citations: What They Are, Where to Build Them, and Which Ones Matter.” “How to Get More Google Reviews (Without Violating Google’s Guidelines).” “Local SEO for Dentists: The Complete Strategy.” “Local SEO for Lawyers: Rankings, Reviews, and Reputation.” “Local SEO for Restaurants: Google Maps, Yelp, and Beyond.” “NAP Consistency: Why Your Name, Address, and Phone Number Need to Match Everywhere.” “Local Link Building: 12 Strategies That Work in 2026.” “How to Rank in Google’s Local Pack.” “Google Business Profile Categories: How to Choose the Right Ones.” “Local SEO vs National SEO: What’s Different and Why It Matters.” “How to Respond to Negative Google Reviews (With Templates).” “Local Schema Markup: LocalBusiness, GeoCoordinates, and sameAs.” “Multi-Location SEO: Managing 5+ Locations Without Duplicate Content.” “Local SEO Audit Checklist: 47 Things to Check.” “Hyperlocal Content Strategy: Neighborhood and City Pages.” “Local SEO Pricing: What Agencies Actually Charge in 2026.” “Local SEO Case Study: From Page 3 to Local Pack in 90 Days.”

Internal linking: Every spoke links to the pillar in the first paragraph and a “Related Reading” section at bottom. Pillar links to every spoke from the relevant section. Spokes cross-link to 2-3 related spokes (Google Reviews spoke links to the Negative Review Response spoke and GBP Optimization spoke). No orphan pages.

Fan-out query coverage: A query like “local SEO” generates sub-queries about citations, reviews, Google Business Profile, and pricing. Each maps directly to a spoke page. The agency can capture citations across 18 different fan-out sub-queries from a single topic cluster.

Example 2: E-commerce brand building a “Running Shoes” cluster

Pillar page: “How to Choose Running Shoes: The Complete Buyer’s Guide” (4,000 words)

Spoke pages (15 total): “Best Running Shoes for Flat Feet in 2026 (Tested by a Biomechanics Lab).” “Best Running Shoes for Wide Feet.” “Best Running Shoes for Plantar Fasciitis.” “Trail Running Shoes vs Road Running Shoes: How to Choose.” “How Much Should Running Shoes Cost? A Price-to-Performance Analysis.” “When to Replace Running Shoes (It’s Not Just About Mileage).” “Running Shoe Drop Explained: Zero Drop vs 12mm.” “Carbon Plate Running Shoes: Do They Actually Make You Faster?” “How to Break In New Running Shoes.” “Running Shoe Size Guide: Why Your Running Shoes Should Be Half a Size Up.” “[Brand A] vs [Brand B]: 2026 Side-by-Side Comparison.” “Best Running Shoes Under $100.” “Stability vs Neutral Running Shoes: Which Do You Need?” “Running Shoe Rotation: Why You Need More Than One Pair.” “How We Test Running Shoes: Our 200-Mile Testing Protocol.”

Why this cluster wins AI citations: “What running shoes should I get for flat feet” generates fan-out sub-queries like “best stability shoes for flat feet,” “arch support running shoes,” “flat feet shoe recommendations.” The dedicated spoke matches the primary query. The stability vs neutral spoke matches sub-queries. The testing methodology page provides the original data and E-E-A-T signal strengthening the entire cluster.

Example 3: Financial advisory firm building a “Retirement Planning” cluster

Pillar page: “Retirement Planning Guide: Everything You Need to Know About Retiring in 2026” (5,000 words)

Spoke pages (20 total): “How Much Money Do You Need to Retire? (2026 Calculator and Benchmarks).” “Retirement Savings by Age: How Much Should You Have at 30, 40, 50, 60?” “401(k) Contribution Limits 2026.” “Traditional IRA vs Roth IRA: Which Is Better for You?” “Roth Conversion Strategies for High-Income Earners.” “The 4% Rule in 2026: Does It Still Work?” “Required Minimum Distributions (RMDs): 2026 Rules.” “Social Security Benefits: When to Claim and How Much You’ll Get.” “Medicare Enrollment: Timelines, Costs, and Common Mistakes.” “How to Create a Retirement Income Plan.” “Retirement Tax Planning: Minimize Your Tax Burden in Retirement.” “Should You Pay Off Your Mortgage Before Retiring?” “How to Choose a Financial Advisor for Retirement Planning.” “Early Retirement (FIRE): How to Retire at 45, 50, or 55.” “Retirement Planning for Self-Employed and Small Business Owners.” “Retirement Planning for Women: Addressing the Retirement Gap.” “Estate Planning Basics: Wills, Trusts, and Beneficiary Designations.” “Long-Term Care Insurance: Do You Need It?” “Retirement Investment Strategy: Asset Allocation by Age.” “How Inflation Affects Your Retirement Savings.”

Fan-out coverage: “How much do I need to retire” generates sub-queries about Social Security timing, Medicare costs, RMD rules, inflation impact. Each has a dedicated spoke. The firm can appear in the AI response multiple times, with different spokes answering different sub-questions.

Example 4: SaaS company building a “Data Analytics” cluster

Pillar page: “What Is Data Analytics? The Complete Guide for Business Teams” (3,500 words)

Spoke pages (16 total): “Descriptive vs Predictive vs Prescriptive Analytics.” “[Product] vs Tableau: Feature, Pricing, and Performance Comparison.” “[Product] vs Power BI: Which Is Better for [Use Case]?” “Data Visualization Best Practices: 11 Rules for Clear Charts.” “KPI Dashboards: How to Build One That People Actually Use.” “SQL for Business Analysts: The Essential Queries.” “Data Cleaning: How to Fix Messy Data Before Analysis.” “What Is a Data Warehouse? (And When Do You Need One?)” “Data Analytics ROI: How to Measure the Business Impact.” “Data Analytics for Marketing Teams.” “Data Analytics for Sales Teams: Pipeline Analysis and Revenue Forecasting.” “Data Analytics for HR: Workforce Planning and Retention Metrics.” “How to Hire a Data Analyst: Skills, Titles, and Salary Benchmarks.” “Data Governance: What It Is and Why Your Analytics Are Useless Without It.” “Customer Case Study: How [Client] Reduced Churn by 23% Using [Product].”

Why the vertical spokes matter: The role-specific spokes (Marketing, Sales, HR) target the fan-out queries AI generates when different professionals ask about analytics. An engineering manager’s sub-queries differ from a marketing director’s. Vertical spokes match role-specific sub-queries directly. The comparison pages match the 32.5% of AI citations that go to comparative content. And each multi-platform AI search strategy benefits from deeper topic coverage because every platform’s fan-out mechanism draws from the same cluster.

Example 5: Real estate agency building a “Buying a Home in [City]” cluster

Pillar page: “How to Buy a House in [City]: The Complete 2026 Guide” (4,500 words)

Spoke pages (17 total): “[City] Housing Market Report: Prices, Trends, and Forecast 2026.” “Best Neighborhoods in [City] for Families.” “Best Neighborhoods in [City] for Young Professionals.” “How Much House Can You Afford in [City]? (2026 Calculator).” “First-Time Home Buyer Programs in [State]: Down Payment Assistance and Tax Credits.” “[City] Property Tax Rates by Neighborhood.” “Closing Costs in [State]: What to Expect and How to Negotiate.” “Home Inspection Checklist: What to Look For.” “How to Get Pre-Approved for a Mortgage in 2026.” “Fixed vs Adjustable Rate Mortgages at 6% Interest Rates.” “[City] School District Rankings and Home Values.” “Condo vs Single Family in [City].” “New Construction vs Existing Homes in [City].” “How Long Does It Take to Buy a House in [City]?” “Working With a Real Estate Agent in [City].” “[City] vs [Nearby City]: Where to Buy in 2026.” “Investment Properties in [City]: ROI by Neighborhood.”

Freshness integration: The market report updates monthly. Property tax rates update annually. First-time buyer programs update when state programs change. The mortgage spoke updates when rates shift. This cadence keeps the cluster inside the 13-week citation window year-round. Local queries generate heavy fan-out. “Should I buy a house in [city]” triggers sub-queries about neighborhoods, prices, schools, taxes, and market trends. A 17-spoke cluster dominates the entire fan-out landscape.

Depth wins. Breadth doesn’t.

The brands winning AI citations aren’t the ones with the highest DA. They’re the ones that own their topic completely. DA explains less than 4% of the thing we’re now optimizing for. Topical authority explains 10x more.

Build deep. Build connected. Build around the fan-out queries AI systems actually generate. That’s how I approach content architecture at Radiant Elephant, and it’s the pattern the data supports across every major study. I laid out the full evidence in the GEO research review covering all 15 proven tactics.