Pricing Methodology

1. Purpose of this page

A buyer should be able to see why a dataset costs what it costs. This page describes the principles behind our prices. Where a dataset has its own pricing methodology document, the button on that dataset's page opens it instead of this page; that document names the comparable offers found, with links and the date of the search.

2. What the price covers

Every price on this site is a one-off payment for a downloadable dataset, with no subscription and no annual contract. Every dataset has a free sample and a passport (technical documentation) that you can review before buying. The terms for using purchased files are in the Dataset License; the policy on returns is in the Terms of Sale.

3. How a price is set

For datasets published from 17 October 2026, a price is set in three steps.

• Product class. The class of the product sets a price corridor: a constants package (model parameters without patient rows), a patient-level dataset, or a generator (source code that reproduces the tables and produces new cohorts).
• Four verifiable features place the product within the corridor:
  • ground truth stored next to the observed values, or re-simulated counterfactual versions of the same patients;
  • a validation report with a benchmark and published negative results;
  • scale: at least 100,000 simulated subjects or at least ten tables;
  • every model constant traced to a published source.
  The more of these a product has, the higher it sits in its corridor.
• Market check. The price is not set above the cheapest comparable paid offer found on the date of the check. If no comparable paid offer exists, the methodology document says so.

Prices are rounded to a whole ten or to a figure ending in 90.

4. Datasets published before 17 October 2026

Datasets published earlier were priced individually, before these steps were written down. This page describes the principles used today and does not recompute the price of an earlier dataset.

5. Questions

If you want to know why a specific dataset is priced as it is, write to code@sentineldata.com.ua.