The same question gave three answers depending on who you asked. The blocker was never technical. It was that no team would agree on what a word meant.
At Sportradar the same business question could give you three different answers depending on which team you asked. Pricing inputs, content decisions, federation reporting, all of it ran across a patchwork of legacy systems and local data stores, and work that should have taken minutes took days or months.
The catalogue to fix it had never been built, and the reason wasn't technical. No team had ever agreed on a shared definition of the basics, and nobody wanted to give up their own store. So I started with the people. I ran discovery with each team separately, engineering, data, business operations, surfacing what each one cared about most and what they were quietly afraid of losing. Then I put the team heads in one room and we negotiated a shared ontology in the open: one definition of a match, an event, a market, with the disagreements documented rather than buried. The rule was simple, each team names its two non-negotiables and we honour those, everything else is on the table.
The people work took about 40% of the timeline, more than I'd budgeted, and it felt uncomfortably slow in the moment. It was the right ratio. Once the ontology was agreed, engineering had a clean contract to build against, and I made sure the framework was credited to the working group, not to me, because the system only survives if the teams feel like they own it.
The reason this hadn't been built before was never the technology. It was that nobody had done the slow, unglamorous work of getting rival teams to agree on what a word meant.