Bradley-Terry Aggregation: How AI Ranks Contractors in Competitive Markets

What Is the Bradley-Terry Model?

The Bradley-Terry model is a statistical framework for estimating the probability that one item will be preferred over another in a pairwise comparison. Originally developed for sports ranking and voting systems, it has become one of the most widely adopted methods for aggregating AI model performance data — including the pairwise preference data collected in AI citation research. In the context of GEO, Bradley-Terry aggregation converts millions of individual AI responses — each of which implicitly or explicitly ranks businesses against each other — into stable, probabilistic ranking scores that predict which contractor an AI will cite most often.

Why Pairwise Comparison Matters for Contractors

AI recommendation systems do not produce rankings in a vacuum. They are implicitly making pairwise comparisons every time they generate a response. When ChatGPT recommends 'Smith Roofing' instead of 'Jones Roofing' in response to 'best roofer in Atlanta,' it is making a pairwise judgment — even if it never explicitly compares the two. At scale, across millions of queries, these implicit pairwise judgments accumulate into statistical patterns. Bradley-Terry aggregation makes these patterns visible and quantifiable.

The Pairwise Matrix for Local Markets

In a local market with 10 competing roofing contractors, there are 45 possible pairwise matchups (10 choose 2). Bradley-Terry aggregation collects data on each matchup — how often does Contractor A appear instead of Contractor B, across all relevant queries? — and uses those win/loss rates to estimate each contractor's underlying 'strength' parameter. The strength parameter is then used to calculate the probability that any given contractor will be cited in a new query, even one where direct pairwise data is sparse.

What Drives Your Bradley-Terry Score?

Your Bradley-Terry strength parameter in a local AI market is determined by the consistency and breadth of your citation signals across multiple dimensions. It is not a single metric — it is a composite measure that rewards businesses that appear across a wide range of query types, locations, and AI platforms simultaneously.

Content Authority Signals

AI systems weight content that directly answers specific questions about specific services in specific locations. Contractors with comprehensive content — service pages that answer 'what does it cost,' 'how long does it take,' 'what areas do you serve,' 'what licenses do you hold,' and 'what can I expect from the process' — provide richer citation material than contractors whose websites are primarily marketing brochures. More citable content means more wins in the pairwise comparison matrix.

Entity Strength

Your business entity — the structured representation of your name, address, phone, service area, and business category in knowledge graphs and structured data — directly influences your Bradley-Terry position. Strong entity signals produce consistent citations across query variations. Weak entity signals (inconsistent NAP data, missing schema markup, thin or absent knowledge graph presence) produce inconsistent citations, which translates directly to a lower strength parameter in the aggregation model.

Third-Party Citation Depth

AI systems use third-party mentions — review platforms, news articles, industry directories, local publications — as validation signals when deciding whether to cite a business. A contractor mentioned across 50 authoritative third-party sources has stronger validation than one mentioned in only 5. In Bradley-Terry terms, this validation depth increases your win probability in pairwise matchups against competitors with thinner third-party footprints.

How Aggregation Affects Your Citation Probability

Once Bradley-Terry aggregation has produced strength parameters for all competitors in a market, those parameters translate directly into citation probabilities for future queries. If your strength parameter is twice that of your nearest competitor, you have roughly twice the probability of being cited in any given query in your market cluster. If your strength parameter is half that of the market leader, you have roughly half the citation probability.

How to Improve Your Bradley-Terry Position

Improving your Bradley-Terry position requires winning more pairwise matchups — appearing more consistently across more query types, locations, and platforms than your competitors. The strategy is systematic: identify which matchups you are currently losing (where do competitors appear instead of you?), then target those specific coverage gaps with content and schema improvements. This approach is more efficient than trying to improve across all dimensions simultaneously.

Identifying Your Pairwise Losses

Run query variation audits that directly compare your citation rate against specific competitors on matched query sets. For each query where a competitor appears and you do not, that is a pairwise loss. Map those losses by service type, location, and query formulation. You will find patterns — you might win on emergency service queries but lose consistently on project planning queries, or win in your core city but lose in adjacent markets. Those patterns define exactly where to invest your GEO resources.

Closing Coverage Gaps Systematically

Each coverage gap you close converts a pairwise loss into a win or draw. Over time, accumulating wins across multiple gap categories improves your aggregate strength parameter. This is why GEO is a compounding strategy: early wins create momentum, early wins reduce competitive threats, and the cumulative strength parameter improvement accelerates as you cover more of the query variation landscape.

The Market Leader Advantage

Contractors who achieve a dominant Bradley-Terry position in their market enjoy a self-reinforcing advantage. High citation rates generate more customer interactions, which generate more reviews, which improve third-party validation signals, which further increase citation probability. Conversely, contractors who are rarely cited lose the opportunity to build these reinforcing signals. The gap between market leaders and laggards in AI-driven local markets widens faster than in traditional search — making the window for GEO investment especially valuable right now.

Know Your Position in the Pairwise Matrix

Market Disruptors runs competitive pairwise citation analysis as part of every GEO engagement. We map your win/loss rates against your top competitors across your core query clusters, estimate your effective Bradley-Terry position, and build a strategy that targets the highest-value pairwise wins first. Book a free strategy call to see exactly where you rank against your competitors in the AI citation hierarchy — and what it would take to move to the top.