Fanatics has grown from an ecommerce business into one of the world's largest sports platforms, reaching more than 100 million fans across commerce, collectibles, betting and gaming, live events, and entertainment. With roughly $13 billion in expected 2026 revenue, the company is building an ecosystem that connects nearly every part of the fan experience.
Fanatics Betting & Gaming has been expanding rapidly since launching online sports betting in 2023. But that speed created challenges for its data team. As new markets, products, and teams came online, analysts built directly from relatively raw data, business logic spread across hundreds of dashboards, and different teams arrived at meetings with conflicting revenue numbers.
Fanatics needed a way to scale analytics without multiplying inconsistent metrics, engineering overhead, and infrastructure costs.
The company evolved from dbt Core to dbt Platform, established a dedicated analytics engineering team, and standardized critical business metrics. Today, its governed data foundation supports approximately 150 contributors, 15+ dbt projects, and 4,000 models.
That work also prepared Fanatics for something it hadn't initially planned for: AI. The models, definitions, and metadata built for analytics now give agents the business context they need to answer questions reliably.
From metric chaos to a governed data foundation
One of Fanatics' biggest challenges was what the team described as "metric chaos." Gross Gaming Revenue (GGR), for example, could have 3 different definitions depending on whether someone asked trading, finance, or commercial teams. Each had a defensible calculation, but conflicting numbers created confusion and forced stakeholders to reconcile reports rather than act on them.
The problem extended across roughly 200 Tableau dashboards containing custom SQL and independently defined business logic.
Fanatics moved that logic into dbt and established a medallion architecture, with standardized data flowing through bronze and silver layers into governed gold marts where enterprise KPIs are defined. A dedicated analytics engineering team works with business stakeholders to align on critical metrics, giving teams consistent definitions they can reuse rather than recreate.
As adoption grew, Fanatics evolved toward a hub-and-spoke architecture using dbt Mesh. Its central analytics engineering project provides shared, governed data products, while downstream teams such as finance, regulatory, and VIP build on that foundation in their own projects.
That approach now supports roughly 150 contributors across 15+ projects and 4,000 models without requiring every team to maintain its own version of enterprise metrics.
Scaling governed analytics while reducing compute costs
As Fanatics' analytics environment expanded, the team needed to control infrastructure costs without limiting development or data availability.
It strengthened ownership, CI/CD, and cost visibility, using dbt and Snowflake metadata to understand how frequently models run, which workloads consume the most compute, and where processing can be reduced.
Fanatics also adopted dbt State to reuse completed work rather than unnecessarily rebuilding models when code or upstream data hasn't meaningfully changed.
The results were significant. According to Luke Kranz, Senior Staff AI Engineer at Fanatics Betting & Gaming, the team reduced Snowflake usage by an estimated 40,000 credits per month after rolling out dbt State.
Even during periods of heavy activity, including the World Cup, Snowflake spending remained relatively flat despite continued growth in usage.
For Fanatics, that means supporting more analytics workloads without allowing infrastructure costs to grow at the same rate. The team can spend less on unnecessary processing while continuing to expand the data foundation the business relies on.
Giving AI the context to deliver trusted answers
The work Fanatics did to standardize metrics and improve data governance also laid the foundation for natural-language analytics.
Fanatics builds Snowflake semantic views on top of its governed dbt marts, adding sample values, synonyms, verified queries, and business definitions that help agents interpret data correctly. Metadata maintained in dbt flows into Snowflake, where AI tools can use it to understand not just which tables to query, but what the data means.
Kranz emphasized the importance of human-defined metadata. An AI agent might infer what a column called "revenue" means, but it doesn't inherently know how Fanatics defines revenue. By documenting that knowledge in dbt, the team can make its expertise reusable across AI applications rather than relying on agents to guess.
Fanatics also evaluates agents before production, targeting scores above 90% to help ensure reliable answers.
That foundation is already supporting approximately 35,000 AI interactions per month. Employees across analytics, trading, and leadership can ask questions through Snowflake agents and tools such as Slack, accessing governed business information without relying on analysts for every request.
The broader impact is a shift in how the data team serves the business. Instead of repeatedly reconciling metrics or answering routine questions, analysts can encode their knowledge into governed data products and agents that make trusted answers available to more people.
Reducing costs and scaling trusted analytics with AI
With governed dbt models, stronger development practices, and reusable business context in place, Fanatics is seeing results across its data infrastructure, analytics operations, and AI adoption:
- Reduced Snowflake usage by an estimated 40,000 credits per month with dbt State, keeping compute spending relatively flat despite growing workloads and periods of heavy activity.
- Scaled governed analytics across approximately 150 contributors, 15+ dbt projects, and 4,000 models, with shared enterprise metrics maintained through a central analytics engineering foundation.
- Supported approximately 35,000 AI interactions per month, expanding access to natural-language analytics through Snowflake agents and tools such as Slack.
- Expanded self-service access to trusted business data, allowing employees to ask questions independently while reducing reliance on analysts for routine requests.
The impact goes beyond infrastructure savings. Fanatics can support growing demand for analytics without proportionally increasing compute costs or manual support work, while its data team makes trusted metrics and business knowledge available to more people through governed data products and AI agents.
Building business context that AI can trust
Fanatics didn't set out to build an AI-ready data platform. It was solving the challenges created by rapid growth: inconsistent metrics, duplicated logic, rising compute costs, and growing demand for answers.
By standardizing those fundamentals with dbt, Fanatics created a foundation that now supports both analytics and AI. Governed models provide consistent data, while definitions, metadata, and semantic views give agents the context to understand how the business operates.
For data teams looking to scale AI, Fanatics' experience offers a practical lesson: govern your metrics, document what your data means, and make that knowledge reusable. When AI builds on a foundation the business already trusts, it can extend the data team's expertise and help more people get reliable answers without adding more manual work.
