Reading the naming-rights data without getting it wrong
Naming-rights numbers travel well. A figure quoted once in a press release turns up in a dozen articles, then in a chart, then in a slide, long after anyone can say what it measured. This post is about what the sponsorship fields on this site do and do not claim.
Deal value is what a publisher printed
The deal_value field is only populated when a source states a figure, and the source URL is
stored beside it. Most deals have no published figure at all, and a blank here means exactly that:
not published. It does not mean zero, and it does not mean small.
Where a figure exists, three caveats apply:
- It is usually the total, not annual. Where a source is ambiguous, the value is stored as printed and the term in years is stored separately, so you can divide if you want an annual figure rather than inheriting someone else's division.
- It is nominal, not inflation adjusted. A deal signed in 1998 and a deal signed in 2024 are not comparable without adjusting, and this dataset does not adjust them for you.
- The USD figure is converted at the start-year average rate from a committed exchange rate file, not a live API. That means a rebuild never silently changes a historic figure, which in turn means two exports of the same deal always agree.
A name change is not always a sponsorship change
The name history table counts every recorded name, and the name-change count in the table column is that list minus one. Some of those changes have nothing to do with sponsors:
- a sponsor renews under a new corporate identity after a merger
- a venue reverts to its original, unsponsored name when a deal lapses
- an operator rebrands the building without any money changing hands
So a high name-change count is a signal about instability of identity, not about sponsorship revenue. Cross-check against the sponsor column before reading anything into it.
Dates are frequently year-only
Sources almost never give the day a naming deal started. Where only a year is known, the date is stored as the first of January of that year with a year-precision flag. The timelines on venue pages draw those as whole years, which is why several of them line up suspiciously neatly on January boundaries. That is the honest rendering of a year-precision fact, not a coincidence.
What is out of scope
Jersey patches, kit deals and competition title sponsorships are not in this dataset, however related they look. It covers sponsorship of the venue: naming rights first, then founding-partner and pouring-rights deals where a public source exists. Anything else would need a different source base and a different set of caveats.