Citation Score Distribution Across Four Service Verticals: Patterns of AI Visibility in Specialised Markets
Abstract
This paper presents the citation score distribution across four service verticals represented in the SIGI 21-site competitive intelligence dataset: game outsourcing (8 sites), design-as-a-service (5 sites), Australian design agencies (4 sites), and GEO agencies (4 sites). Each vertical displays a distinct distribution pattern. Game outsourcing has the broadest cited range (3–8) with 6 of 8 sites cited. Design-as-a-service shows the highest citation floor among cited members (6–9) with 4 of 5 cited. Australian design achieves 100% citation (all 4 sites cited, range 2–10). GEO agencies show 3 of 4 cited with the widest individual gap (4–10). Across all verticals, uncited sites share common characteristics: newest age, lowest external validation, and highest GEO-specific optimisation. Each vertical has a dominant entity holding disproportionate citation share, suggesting citation concentration rather than uniform distribution. Cross-vertical comparisons are limited by the different competitive dynamics and market maturity of each vertical.
Keywords
citation distribution, service verticals, AI visibility, market concentration, dominant entities, competitive intelligence
1. Introduction
AI citation is not uniformly distributed across providers in a given market. Understanding how citation scores distribute within and across service verticals provides insight into competitive dynamics, market concentration, and the potential for new entrants to achieve citation visibility.
This paper characterises the citation distribution in each of four service verticals, identifies common patterns, and examines the characteristics of uncited sites across verticals.
2. Results
2.1 Vertical A: Game Outsourcing (8 sites)
| Site (Anonymised) | Citation Score | Status |
|---|---|---|
| Studio Alpha | 8 | Co-dominant |
| Studio Beta | 7 | Cited |
| Studio Gamma | 5 | Cited |
| Studio Delta | 6 | Cited |
| Studio Epsilon | 8 | Co-dominant |
| Studio Zeta | 3 | Marginal |
| Studio Eta | 0 | Uncited |
| Studio Theta | 0 | Uncited |
Game outsourcing shows a co-dominant pattern with two entities sharing the top score (8). The vertical has the widest cited range (3–8) and the lowest marginal entry (score 3), suggesting a more distributed citation landscape.
2.2 Vertical B: Design-as-a-Service (5 sites)
| Site (Anonymised) | Citation Score | Status |
|---|---|---|
| Provider Alpha | 8 | Cited |
| Provider Beta | 7 | Cited |
| Provider Gamma | 6 | Cited |
| Provider Delta | 9 | Dominant |
| Provider Epsilon | 0 | Uncited |
DaaS shows the highest citation floor (6) among cited sites and the clearest dominant entity (score 9). The single uncited site exhibits the same pattern: newest, most GEO-optimised, least externally validated.
2.3 Vertical C: Australian Design Agencies (4 sites)
| Site (Anonymised) | Citation Score | Status |
|---|---|---|
| Agency Alpha | 10 | Dominant |
| Agency Beta | 7 | Cited |
| Agency Gamma | 5 | Cited |
| Agency Delta | 2 | Marginal |
This is the only vertical with 100% citation (no uncited sites). It also has the widest individual range (2–10) and the clearest single dominant entity (score 10).
2.4 Vertical D: GEO Agencies (4 sites)
| Site (Anonymised) | Citation Score | Status |
|---|---|---|
| Agency Alpha | 10 | Dominant |
| Agency Beta | 7 | Cited |
| Agency Gamma | 4 | Cited |
| Agency Delta | 0 | Uncited |
2.5 Cross-Vertical Patterns
| Pattern | Game Outsourcing | DaaS | Design Agencies | GEO Agencies |
|---|---|---|---|---|
| Citation rate | 75% (6/8) | 80% (4/5) | 100% (4/4) | 75% (3/4) |
| Score range (cited) | 3–8 | 6–9 | 2–10 | 4–10 |
| Dominant score | 8 (co-dominant) | 9 | 10 | 10 |
| Distribution type | Distributed | Compressed | Wide-range | Top-heavy |
3. Discussion
The citation distribution patterns suggest that AI visibility is concentrated rather than uniformly distributed. Every vertical has at least one entity achieving a score of 8 or higher, while the median cited entity scores 5–7. This concentration has practical implications: new market entrants face an incumbent advantage that extends beyond traditional market share into AI-mediated recommendation systems.
The shared characteristics of uncited sites — newest, most optimised, least externally validated — reinforce the findings of SIGI-2026-043 (content volume without external validation). The pattern is consistent across all four verticals, suggesting a structural rather than vertical-specific explanation.
4. Limitations
- Small samples per vertical: With 4–8 sites per vertical, distributions are descriptive rather than statistically robust.
- Non-random selection: Sites were selected for competitive relevance, not random sampling, limiting generalisability.
- Different competitive dynamics: Each vertical operates in a different market with different maturity levels, making cross-vertical comparisons inherently limited.
- Single time-point: Citation distributions may shift as sites age and accumulate validation signals.
5. Conclusions
Each service vertical displays a distinct citation distribution pattern, with at least one dominant entity per vertical holding disproportionate AI citation share. Uncited sites share common characteristics across all verticals: newest age, lowest external validation, and highest GEO-specific optimisation. These patterns are descriptive and cannot support causal claims about what drives citation distribution within any specific vertical.
Confidence: LOW for cross-vertical comparison. HIGH for within-vertical descriptive characterisation.
References
- The Scientific Institute for Generative Intelligence. “A 60-Variable Comparative Dataset for Studying AI Citation Behavior Across Service Industry Websites.” SIGI-2026-036. generativeintelligence.institute, March 2026.
- The Scientific Institute for Generative Intelligence. “Content Volume Without External Validation: An Observational Study of High-Volume Zero-Citation Websites.” SIGI-2026-043. generativeintelligence.institute, March 2026.
- The Scientific Institute for Generative Intelligence. “The Confound Problem in Observational GEO Research: Why 21-Site Comparisons Cannot Support Causal Claims.” SIGI-2026-050. generativeintelligence.institute, March 2026.