Ask a team what data it has and the answer depends entirely on who you ask. The analysts describe tracking and video. The medical team describes something else. Nobody describes the whole thing, which is how a team ends up with a dozen partial pictures and no complete one.
This is the whole picture, and what each part demands of you. Volumes vary so much between setups that quoting numbers would be misleading, so what follows describes shape rather than size.
Positional and tracking data
What it is. Where every player was, many times per second, from optical camera systems, wearable devices or both.
Shape. High frequency, small per record, enormous in aggregate over a season. Compresses well. Almost useless without the metadata that says which session, which player, which calibration.
Who uses it. Performance analysts, sports science, increasingly recruitment.
What it demands. The raw feed, not the vendor's summary. Aggregates answer last season's questions; the raw positions answer next season's. This is the single most common thing a team discovers it cannot export.
Video
What it is. Game footage, training footage, tactical angles, and the clips and tags your analysts have cut from all of it.
Shape. Dominates every other category by storage volume, usually by an order of magnitude or more. Expensive to move, which is exactly why it tends to stay wherever it first landed.
Who uses it. Coaches, analysts, recruitment, media.
What it demands. Somewhere with room to grow and a realistic plan for moving it. The tagging is often more valuable than the footage and is frequently locked inside the tool that produced it.
Load and availability
What it is. Training load, session ratings, sleep and readiness inputs, rehabilitation progress.
Shape. Small, frequent, and strongly personal. Often partly self-reported, which makes consistency a human problem rather than a technical one.
Who uses it. Sports science, medical, coaching staff selecting a roster.
What it demands. Careful access control. This is where performance data starts shading into health data, and staff expectations about who can see it are usually stronger than whatever the system enforces.
Medical records
What it is. Injury history, screening, imaging, treatment notes.
Shape. Low volume, long retention, highest sensitivity in the building.
Who uses it. Medical staff, and a much narrower group than most teams realize once they look.
What it demands. In most jurisdictions health information attracts stricter protection than ordinary records, and the obligation sits with the team as the organization the athlete has a relationship with, regardless of which supplier holds the files. Treat the retention period as a decision to be made deliberately rather than a default to inherit.
Scouting and recruitment
What it is. Reports and assessments on your own players and on targets elsewhere.
Shape. Mostly written judgement rather than measurement. Small, and the most commercially damaging thing you own if it leaks.
Who uses it. Recruitment, the sporting director, the board at transfer windows.
What it demands. Access control that survives staff turnover, since scouts move between teams more than most roles. Ask where this data lives with the same seriousness you would apply to payroll.
Game and event data
What it is. The coded record of what happened: passes, shots, duels, set pieces, with timestamps.
Shape. Modest volume, high analytical value, usually bought from a provider rather than generated in house.
What it demands. Read the license rather than the invoice. Purchased event data often comes with real restrictions on retention, redistribution and derived works, and those terms outlive the contract.
Venue and operations
What it is. Ticketing and access, crowd flow, and increasingly the network, sensor and radio telemetry of the stadium itself.
Shape. Bursty, concentrated around games, and the category growing fastest.
What it demands. Attention, because it is usually owned by a different department from the performance data and nobody has drawn the two together. Where it touches identifiable supporters it carries the same obligations as any customer data.
What the inventory is for
Listing it is not the point. The point is that for each row you should be able to answer three questions without ringing a supplier:
- Where is the primary copy? Not a copy. The one that would still exist if a contract ended tomorrow.
- Who can read it? Named roles, and a record of who actually did.
- Could you get it out? In what format, at what granularity, and how long would it take.
If any row leaves you guessing, that is where to start. The questions worth asking a supplier turn each gap into something you can put to a vendor, and sports data sovereignty covers why the answers matter commercially as well as technically.