Get Started

Topic Authority Concentration Score: Mapping AI Visibility Breadth vs Depth

LinkedIn @ultrascoutai
UltraScout AI Proprietary Metrics Series · View all guides →
YH
Yuliya Halavachova · Founder & Principal Data Scientist at UltraScout AI

Yuliya developed the Topic Authority Concentration Score to help brands understand the structural shape of their AI visibility — not just how much they appear, but where. She leads UltraScout AI's proprietary metrics research programme from London, UK.

You track how often AI mentions your brand. But do you know which topics those mentions cluster around — and whether that clustering is a strategic asset or a structural liability? The Topic Authority Concentration Score (TACS) answers that question with mathematical precision.

The Core Insight

"Owning one topic deeply beats dabbling in ten. But one topic is also one point of failure. The TACS tells you exactly where your AI authority sits on that spectrum."

— Yuliya Halavachova, UltraScout AI

1. The Problem: Not All Citation Patterns Are Equal

Two brands can have identical overall citation rates — say, 55% across their tracked query set — and yet face completely different strategic realities. One brand appears consistently across twenty topic clusters: pricing, implementation, compliance, integrations, use cases, alternatives. The other appears in one topic cluster with overwhelming frequency, and barely registers in anything else.

Which is stronger? The answer is: it depends — and that "it depends" is precisely what the TACS quantifies.

Brand A: Concentrated

87% of citations in "CRM integrations" topic

Remaining 13% scattered across 8 topics

TACS: 7,700

Concentrated — Fragile

Brand B: Balanced

Citations spread across 12 topic clusters

Largest single topic: 22% of citations

TACS: 1,450

Balanced — Healthy

Both score 55% on raw citation rate. Their strategic positions could not be more different. The TACS captures this distinction — and it is a distinction that matters enormously when a competitor publishes content in your dominant topic, or when an AI platform updates its weighting for a category you overindex on.

2. The TACS Formula: Adapted HHI for AI Visibility

The Topic Authority Concentration Score adapts the Herfindahl-Hirschman Index (HHI) — the economic concentration measure used by antitrust regulators to assess market monopolisation — for the context of AI citation distribution across topics.

The HHI was chosen deliberately. It has a well-understood interpretation, penalises extreme concentration non-linearly (squaring each share magnifies the difference between moderate and extreme concentration), and produces a bounded score that is easy to communicate to non-technical stakeholders.

Topic Authority Concentration Score

TACS = 10,000 × Σ(Topic Share²)

Where Topic Share = Citations on that topic ÷ Total citations across all topics

Score range: 0 – 10,000 (theoretical maximum if 100% of citations on one topic)

Score Ranges and Interpretation

2,500 – 10,000

Concentrated

Dominant in 1–2 topics. Category ownership with single-point-of-failure risk.

1,000 – 2,500

Balanced

Healthy spread across 4–8 topics. Resilient authority with clear focus areas.

Below 1,000

Fragmented

Too thin to own anything. Visible everywhere, authoritative nowhere.

Worked Example: Calculating TACS

Suppose a brand has 200 total AI citations tracked across five topic clusters:

Topic Citations Topic Share Share²
CRM Integration 140 0.70 0.4900
Implementation 30 0.15 0.0225
Pricing 16 0.08 0.0064
Compliance 8 0.04 0.0016
Alternatives 6 0.03 0.0009
Total 200 1.00 Σ = 0.5214

TACS = 10,000 × 0.5214 = 5,214 → Highly Concentrated

This brand is effectively a one-topic brand in AI. If a competitor launches a comprehensive CRM integration hub and earns 20 citations in that space, this brand's dominance erodes rapidly — and there is no other topic cluster providing cushion.

Why Squaring Matters

A brand with citations split 50%/50% across two topics has TACS = 10,000 × (0.25 + 0.25) = 5,000. A brand split 10% across ten equal topics has TACS = 10,000 × (10 × 0.01) = 1,000. The squaring function correctly penalises concentration and rewards genuine breadth — it is not merely proportional.

3. When Concentration Is a Strategic Advantage

Not all high TACS scores are failures. For certain brands and market positions, deep concentration in a single topic is a deliberate strategy — and the right one.

The Case for Concentration: Category Ownership

Category Ownership Signal

When AI platforms consistently cite your brand as the authority on a specific topic, they are — in effect — naming you the category leader. This is the AI equivalent of ranking first for your most important keyword. It builds recognition, trust, and recall.

Example: A cybersecurity firm that owns 80% of citations in "zero-trust network security" queries has built a defensible position. Every buyer asking about zero-trust hears their name. That concentrated authority is extremely valuable.

When concentration is correct:

  • You are a genuine category leader in that topic
  • The topic is high-volume and high-intent
  • The topic aligns with your primary revenue driver
  • You have content depth that makes the concentration defensible

The Case Against Concentration: Single Point of Failure

Single Point of Failure Risk

Concentration becomes dangerous when the dominant topic is vulnerable to external shifts: algorithm updates, competitor content investment, or a change in buyer behaviour that reduces query volume in that topic cluster.

Risk factors for concentrated brands:

  • The dominant topic is a trend, not a permanent buyer concern
  • Multiple well-resourced competitors are investing in that exact topic
  • Your brand's actual product strength spans multiple topics — but AI doesn't know it yet
  • The topic is being commoditised by AI-generated content, eroding individual brand authority

4. When Fragmentation Is the Real Problem

A TACS below 1,000 signals a fragmented authority profile. This is often mistaken for "broad coverage" — it is not. It is dilution.

Fragmented brands appear in many topic clusters but with insufficient citation density in any single cluster to register as an authority. AI platforms are probabilistic: they cite sources that appear consistently associated with a topic. A brand mentioned once across twenty topics does not build the signal required for reliable citation on any of them.

Fragmentation Warning Signs

  • No single topic accounts for more than 15% of citations
  • Citations appear across more than 15 distinct topic clusters
  • Citation rates per topic are below 20% even on core queries
  • Competitors consistently outrank you in topics you believe you own
  • AI descriptions of your brand feel generic rather than specific

The cure for fragmentation is not always to reduce topics. It is to identify the 3–5 topics where you have the strongest underlying content assets and authority signals, then build citation depth in those clusters deliberately — while allowing lower-priority topics to remain at maintenance level.

5. Platform-by-Platform: Breadth vs Depth Preferences

Different AI platforms weight topic breadth and depth differently in their citation behaviour. Based on UltraScout AI's analysis of over 40,000 queries across six platforms in Q1 2026:

Platform Preference Optimal TACS Range Rationale
ChatGPT Moderate Depth 1,500 – 3,000 Rewards clear category associations but also values contextual breadth. Penalises single-topic over-specialisation in conversational contexts.
Perplexity Breadth + Freshness 800 – 1,800 Citation engine pulls from multiple fresh sources. Brands with broader topic coverage appear in more citation chains. Narrow brands are cited less frequently overall.
Claude Deep Expertise 2,000 – 4,000 Prioritises demonstrated expertise over breadth. Concentrated authority on a relevant topic is rewarded. Fragmented profiles are discounted.
Gemini Balanced Breadth 1,000 – 2,500 Integrates Google Search signals. Brands with broader authority across related topics rank better in multi-topic queries. Mirrors Google's E-E-A-T breadth expectations.
Copilot Commercial Intent Depth 1,800 – 3,500 Microsoft's commercial integration means high-intent transactional topics are weighted heavily. Deep authority in commercial topics outperforms broad general coverage.
Grok Current Affairs Breadth 700 – 1,500 Real-time data focus. Brands active across current topics and trending discussions benefit from broader profiles. Narrow evergreen authority is less prominent.

Strategic Implication

If Perplexity is your most important platform (growing fastest in B2B research use cases as of 2026), a TACS above 2,500 is a liability, not an asset. If Claude is critical for your enterprise audience, concentrated expertise is rewarded. Your optimal TACS is platform-weighted — and the UltraScout AI platform calculates this automatically.

6. Case Study: B2B SaaS Brand with Over-Concentrated AI Authority

Case Study: Project Management SaaS — Topic Rebalancing

Profile: Mid-market project management platform, primarily serving engineering teams. Strong SEO presence in "agile sprint planning" content.

Baseline TACS Audit (January 2026):

Topic Cluster Citation Share Strategic Importance
Agile Sprint Planning 73% High — core use case
Kanban Boards 11% High — major feature
Engineering Team Productivity 7% High — ICP topic
Project Timeline Software 5% Medium — expansion topic
Jira Alternatives 4% Very High — acquisition topic
TACS Score 5,479 — Highly Concentrated

The Problem Identified:

The brand's second-highest strategic topic — "Jira Alternatives" — was its most important acquisition driver (buyers actively comparing products are furthest along the purchase journey), yet it captured only 4% of citations. Meanwhile, "Agile Sprint Planning" — largely an informational topic attracting developers doing research rather than buyers evaluating software — consumed 73% of AI authority.

A competitor launched a comprehensive "Jira alternatives" content hub in November 2025, and the brand's TACS climbed from 4,100 to 5,479 within 90 days — not because they grew in their dominant topic, but because the competitor displaced them even in the 4% they held in alternatives queries.

Rebalancing Strategy Implemented:

  • Published 14 new pages targeting "Jira alternatives for engineering teams" query cluster with structured comparison data, user testimonials, and verified third-party reviews
  • Added FAQ schema to existing sprint planning pages linking across to kanban and timeline topics — creating internal topic bridges
  • Created an "engineering team productivity" pillar hub connecting all core topics through a unified narrative
  • Published three original data studies on engineering team productivity metrics (cited by industry publications, increasing third-party authority in that cluster)
  • Implemented llms.txt listing all product use cases explicitly — ensuring AI crawlers understood the full topic scope

Results After 90 Days:

Metric Before After Change
TACS Score 5,479 1,820 -3,659 (healthier)
Jira Alternatives citation share 4% 23% +19%
Overall citation rate 51% 64% +13%
AI-attributed trial signups (est.) Baseline +34% +34%

The reduction in TACS — from 5,479 to 1,820 — corresponds directly to the shift from fragile single-topic dominance to a balanced authority profile. Critically, citation rate grew even as concentration fell, because breadth unlocked new query clusters that previously returned zero coverage.

7. How to Measure Your TACS

Step 1: Define Your Topic Taxonomy

Before calculating TACS, you need a topic taxonomy — a structured list of the query clusters relevant to your brand. This is not your keyword list; it is a higher-level clustering of what buyers ask about in your category.

Topic Taxonomy Guidelines

  • Aim for 8–20 topic clusters. Too few merges meaningfully distinct queries; too many creates noise.
  • Each cluster should contain 5–30 query variants. "Best CRM for small business" and "CRM software for startups" belong to the same cluster.
  • Clusters should map to distinct buyer concerns, not just keywords. "Implementation complexity" and "pricing" are different buyer concerns even if they sometimes co-occur.
  • Include both informational topics (what is X, how does X work) and commercial topics (best X, X alternatives, X pricing) — they serve different intent stages and should be tracked separately.

Step 2: Track Citations by Topic Cluster

Run your query set through each AI platform you track. For each query, record whether your brand was cited. Aggregate citations by topic cluster. Total citations per cluster divided by total citations across all clusters gives you each Topic Share.

Step 3: Calculate and Monitor

Apply the formula: TACS = 10,000 × Σ(Topic Share²). Track this monthly. A rising TACS indicates you are becoming more concentrated — either because you are growing in a dominant topic (potentially good) or because you are losing ground in secondary topics (potentially bad). Context is everything.

8. Strategies for Improving Your TACS Profile

For Over-Concentrated Brands (TACS above 2,500)

Diversification Without Dilution

The goal is not to abandon your dominant topic — it is to build secondary authority clusters that provide resilience. A 70/30 split across two well-chosen topic clusters is far healthier than 90/10 even if it reduces raw citation volume in the dominant topic short-term.

Tactical actions:

  • Identify underserved adjacent topics: What questions do buyers ask before and after your dominant topic? Build content there.
  • Create topic bridges: Internal linking and contextual mentions connect topics in AI's training patterns. A sprint planning guide that naturally references "engineering team productivity metrics" helps AI associate both topics with your brand.
  • Target acquisition-stage queries specifically: These are almost always underrepresented in concentrated brands because they are harder to rank for. Invest disproportionately here.
  • Diversify your external citation sources: If 90% of your third-party citations reference one topic, expand your PR and outreach to adjacent topics.

For Fragmented Brands (TACS below 1,000)

Consolidation Into Core Clusters

Fragmented brands need to consolidate rather than expand. Pick 3–5 topics where you have the strongest existing content assets and the highest strategic alignment with revenue, then build citation depth in those clusters deliberately.

Tactical actions:

  • Audit your existing content: Which topics have the most content? Which have the most backlinks? These are your natural authority clusters — reinforce them first.
  • Consolidate thin content: Multiple short pieces on the same topic dilute authority. Merge and strengthen them into comprehensive resources.
  • Prioritise structured data: FAQ schema, HowTo schema, and entity markup on your core topics gives AI platforms clearer signals about your expertise.
  • Accept that some topics will be de-prioritised: Spreading effort across 20 topics equally keeps you fragmented. Strategic concentration — even temporary — builds the momentum required for citation.

For Balanced Brands (TACS 1,000–2,500)

If your TACS is already in the healthy range, your goal is maintenance and optimisation rather than restructuring. Monitor regularly for drift — concentration can creep upward as competitors displace secondary topic positions, even if you maintain your dominant one. Set alerts for TACS changes greater than 300 points in any 90-day period.

9. TACS and the Broader AI Visibility Framework

The TACS does not operate in isolation. It integrates with UltraScout AI's broader suite of proprietary metrics:

  • Intent Alignment Index (IAI): TACS tells you which topics you are cited on; IAI tells you whether those topics are at the right buyer journey stage. A brand with balanced TACS but all citations in informational topics has an IAI problem.
  • Zero Coverage Risk Score (ZCRS): Your ZCRS identifies which topics competitors dominate where you have no citations. These are the first targets for TACS rebalancing.
  • AI Platform Diversity Index (APDI): TACS is calculated separately per platform — you may be balanced on ChatGPT and concentrated on Perplexity. The APDI tells you whether your overall platform profile is diversified.
  • Citation Freshness Index (CFI): Stale content in your dominant topic is a concentration-plus-freshness compound risk. When TACS is high and CFI is low, the single point of failure is also becoming stale — the highest-risk combination in AI visibility.

Key Takeaway

The Topic Authority Concentration Score transforms "where do we appear?" from a qualitative observation into a quantified, trackable, actionable metric. It tells you not just whether you are visible, but whether the shape of that visibility is structurally sound — or whether a single competitor move could hollow out your AI presence overnight.

Measure your Topic Authority Concentration Score

UltraScout AI calculates your TACS automatically across all tracked platforms and topic clusters, with monthly trend tracking and alerts when concentration shifts beyond safe thresholds.

References

  • Halavachova, Y. (2026). "Topic Authority Concentration Score: Adapting the Herfindahl-Hirschman Index for AI Citation Analysis." UltraScout AI Research Series.
  • U.S. Department of Justice. (2023). "Herfindahl-Hirschman Index." DOJ Antitrust Division. Methodology reference for HHI concentration scoring.
  • Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., & Deshpande, A. (2024). "GEO: Generative Engine Optimization." Proceedings of the 30th ACM SIGKDD Conference. arXiv:2311.09735
  • UltraScout AI. (2026). "AI Platform Citation Behaviour Analysis: 40,000 Query Dataset, Q1 2026." Internal Research Report.