A practical guide to selecting per‑hour, per‑record, or flat‑fee pricing for proprietary data assets licensed to AI developers.
Read note[§0] Dataset valuation research
A range is only as useful
as its assumptions
Put licensing and sale scenarios side by side. See what drives the result—and what you still need to know.
Net proceeds by structure
USD / transaction[ Low — Base — High ] · scenarios, not confidence intervals
A range starts with your assumptions
Enter your own prices, or explore the clearly labelled illustrative preset.
| Structure | Low | Base | High |
|---|
[§1] Scenario output, not an appraisal. Assumptions listed below. One transaction per structure. No assumed buyers, renewals, tax, time value, or probability of a deal. Negative results are retained.
[§2] Assumptions ledger
The numbers behind the numbers
Total units × usable share
Usable volume × entered unit price
Gross × (1 − deductions) − costs
User inputs required. Prices are not independently verified.
[§3] Sensitivity analysis
Which assumptions move the result?
Change one input at a time. Ranked by absolute change in net proceeds. Percentage-point changes differ from relative percentage changes.
| Assumption | Change tested | Change in net USD |
|---|
The workbench calculates the consequences of your assumptions. It does not discover a market price or establish that a dataset can be licensed.
- Verify the rights and the exact scope of the transaction.
- Measure usable volume, not just raw record count.
- Record a source and rationale for every proposed unit price.
No comparable transaction database is connected. Illustrative presets are examples of arithmetic, not pricing evidence.
[§5] From the research desk
Read beyond the headline price
[§6] Analyst review
Request review of this scenario
Bring the assumptions. We’ll help identify the evidence needed to test them.
Your working scenario
Enter your inputs in the workbench above.
Calculator inputs stay in your browser unless you explicitly include them.
No source files are needed. Please leave out datasets, credentials, and sensitive personal information.