Making live college-football APIs ready for customer adoption.

Prioritize the complete customer requirement

Major operators needed pricing and results during college-football games. Getting them live required clear availability states and the market coverage each customer needed. I made closing that gap the top priority when I took dedicated NBA and CFB roadmap ownership in early 2025.

Swish had begun development in 2024. I defined product and API requirements and led release acceptance with Data Science, Data Engineering, Software Engineering, Trading and Customer Success. Engineering and data-science partners built the production services and models.

Start integration while live markets are still being validated

I separated the work customers could start from the work that still needed validation. On May 22, 2025, the Test API and documentation let customers begin integration before live markets were ready. I kept its request and response contract aligned with production so that work could carry forward.

The August 12, 2025 production release was a suspended beta. Customers could access the interface while markets remained unavailable for betting during validation. This let integration progress without treating an accessible API as a ready betting product. An operator used it during the 2025 season; multiple major operators are live in 2026.

Letting customers price and settle combined bets

On October 14, 2025, I led releases for combined bets using live player selections and for live team/match markets. A combined bet joins multiple selections into one request, requiring a shared price and results for settlement.

On January 7, 2026, I expanded the shared pricing and settlement APIs to support 26 existing CFB team/match markets before and during games. The request and response structures stayed the same. These were newly supported markets in combined requests, not 26 new models or separate releases.

One operator needed alternate lines

One major global operator needed live alternate lines before it could launch. These offer different thresholds for the same statistic, such as different passing-yard totals. I made them a 2026 offseason priority and preserved existing identifiers and API structures.

I defined player and team/match behavior, including observed values, remaining and final projections, probabilities, suspensions and incremental updates. I waited to announce the APIs until customer access and console support worked.

We shipped the capability in August 2026, and that operator went live afterward. This was one customer’s launch requirement; the product was already in live use.

A release program that enabled adoption

The delivered scope grew from player APIs to team/match markets, combined requests and alternate lines. Adoption followed different milestones: one operator used the product in 2025; another needed the 2026 alternate-line expansion before launch. Multiple major operators are now live.

I also scoped 100+ pre-match combinations that reused shared pricing, suspension and results, with Engineering implementing generation. Further releases expanded live match markets and prebuilt combinations. Those additions extended the product around the same shared interfaces.