Enterprise API products, from roadmap to customer adoption.

Own the product beyond its first response.

Swish Analytics supplies predictive sports data and APIs to enterprise customers. My portfolio spans NBA and college football: new market families, shared pricing and bet composition, results, and operational controls. In-play means during the game. Bet composition combines supported selections into one request; settlement determines their results when play is complete.

I own the roadmap, product requirements, release decisions, and customer readiness. Data Science, Data Engineering, Software Engineering, Trading, and Customer Success contribute their specialist work. My job is to connect those dependencies into a coherent product contract, then follow it into use. Dedicated NBA and CFB roadmap ownership began in early 2025; earlier releases sit within my Product Manager tenure, which began in December 2022.

The flagship program

CFB In-Play: sequence the path to adoption.

I made a missing live college-football product a top roadmap priority. Customers could start integration through a non-actionable Test API before live validation. Production followed in stages, then player composition and wider market coverage. A major global operator needed alternate lines—different thresholds for the same statistic—before launching. I drove that prerequisite as a shared product capability; the operator subsequently went live.

Decision 01 — Deliver the whole market lifecycle

Make new NBA offerings complete client products.

I owned NBA product releases spanning period-based offerings, team and match alternate lines, and 16 Leaders markets. Leaders offered rankings, such as where a player finished relative to others, rather than only raw statistics. Delivering that family meant connecting pricing and composition to upstream data, corresponding rank results, and participation-aware settlement.

A player’s rank is different from the number of points they scored. A player who did not participate is different from one who played and recorded zero. I worked through missing data, result properties, and those distinctions so clients could consume the offering consistently across its lifecycle. Engineering and quantitative specialists implemented the underlying services and models.

Scope decisions were part of delivery. For live first-quarter composition, I released three ready markets while deferring team points because its calculations did not reconcile with related markets. Customers received usable scope while the inconsistency stayed out of that release.

Decision 02 — Separate the interface from the behavior change

Give clients finer control without forcing a migration.

A market-level pause couples the availability of its individual selections. I led a staged NBA selection-level suspension rollout so an adopting client could keep an eligible selection active while another stayed suspended.

In November 2025, player pricing and alternate interfaces gained selection-specific fields alongside the existing market-level behavior. In January 2026, differentiated behavior was enabled for live player props, live alternates, and in-play composition. The first stage introduced fields across pre-game and live interfaces; the second enabled behavior for the documented live products.

Existing clients could continue conservative market-level handling. Adopting clients could use the finer controls. The product result was an additional client capability with the legacy contract preserved, not a mandatory migration.

Decision 03 — Reuse the composition and results contract

Expand shared capabilities instead of duplicating them.

In January 2026, I expanded shared pre-game and live bet-composition support to 26 existing CFB team and match markets, with results through the existing path and no structural request or response changes. Customers gained additional composable offerings through familiar interfaces. These were new supported combinations of existing markets, not new underlying models or endpoints.

The same reuse principle appears in the 100+ CFB prebuilt combinations delivered for pre-match use. I led scoping and the generation handoff to Engineering. Templates use shared pricing, suspension, and results capabilities rather than a duplicate lifecycle. Clients still need to configure and test the offerings they adopt.

Follow the release into the customer’s workflow.

NBA Fast Props—short-horizon offerings during a game—also needed dependable consumption. Releases added score and clock context, selection identity, and a current-market filter that retained concluded markets for results. Narrowing the live inventory should not remove the information needed to settle it.

A separate production incident involved selections disappearing in a customer integration. I investigated requests in Datadog, coordinated staged engineering fixes, and followed up until the customer confirmed the selections stopped disappearing. That confirmation concerns the specific incident; the context and filtering releases are separate product work, not the established fix for it.

A portfolio customers can build on.

The outcomes take different forms: a major operator launching CFB, newly available NBA product families, additional composition options, optional finer controls, and a customer-confirmed production correction. Together they show sustained product ownership across launches and the shared capabilities that make those launches usable.