Methodology

What counts

The index records four things. Deals and bids that a buyer, a seller or a named publication has disclosed with a dollar figure. Price bands that a buyer publishes on its own pages. Commitments and payout totals a buyer has announced. Referral bounties a buyer publishes for introductions. Each row is one figure from one source.

Sources

We prefer the buyer's own page or terms document, then court filings and named press reports, then registers such as Who Sells Data (whosellsdata.com/buying), which cites its own sources. Every row links to the page we read and gives the date we read it. Where a figure reached us second-hand, the source label names both the original publication and the page that cited it.

How ranges are recorded

Low and high are in US dollars as published. An open-ended figure such as "$1M+" is stored as a low of $1,000,000 with no high. An "up to" figure is stored as a high with no low. Percentages are kept as percentages, with any floor or cap in the dollar columns. Figures the buyer labels indicative or illustrative keep that label in the description. We do not convert annual figures to one-time figures or the reverse.

What is excluded

Calculator and estimator outputs, because they depend on the inputs chosen. Offers that are not public. Programs that buy consumer or personal data. Valuations and funding rounds of the buyers themselves. Any figure from a Marin Data engagement; once we have ten closed deals we may publish an anonymized distribution, labelled as ours.

Update cadence

Every row is re-checked against its source once a month, and the updated date at the top of the index changes only when a row is re-checked. When a buyer changes a published term, we add a new row with the new date and keep the old one, so the record shows the change.

About Marin Data

Marin Data helps companies license their data to AI labs. It is not an agent of any buyer listed here.

Back to the index · CSV · JSON