The takeaway
A first inventory describes the data. It does not need to contain the data itself.
Give the assessment a useful starting point
You do not need to assemble a giant export to begin exploring licensing. A concise inventory can help your team explain what exists, where it lives, and who knows it best. Start with systems that record a complete operational process, such as support, order management, or project delivery.
Give one person responsibility for the inventory and invite system owners to validate their rows. The goal is a shared picture of the opportunity and the open questions, not a promise that every collection is ready to license.
Capture eight fields for each collection
| Field | What to record |
|---|---|
| System and owner | The application, department, and responsible person |
| Workflow | The business task the records describe |
| Record types | Cases, orders, events, documents, or other units |
| History | Date range, known gaps, and migrations |
| Approximate scale | A count, its unit, and whether it is verified |
| Connections | How related records are linked across systems |
| Restrictions | Known contractual, privacy, or access questions |
| Export readiness | Available formats and who can check them |
Use consistent units and honest uncertainty
“A million records” could mean messages, attachments, field updates, or complete cases. Use a unit a reviewer can understand. If one support case contains many messages, distinguish the number of cases from the number of individual events.
Mark estimates and unverified assumptions clearly. A system owner might know the account began in 2019 but still need to confirm whether all history is retained. Record that as a question instead of representing seven years of complete data as a fact.
Describe structure before sharing samples
A field list or data dictionary can explain a collection without revealing actual customer content. If an example is useful for an initial discussion, create a clearly labeled synthetic example or agree on a reviewed sample process. Do not pass a made-up record off as evidence of your actual data quality.
Dataset documentation research emphasizes making a dataset's origins and intended uses understandable. For this inventory, add short notes that only your team would know: what a status means, why a period is missing, or how a system migration changed identifiers.
Finish with a shortlist and next actions
Choose a few candidate collections and assign an owner to each unanswered question. Your next step might be checking retention, locating a customer agreement, or confirming that an export preserves event history.
Bring that shortlist to the assessment conversation. It helps everyone discuss scope and feasibility while keeping the work proportional to the opportunity. A detailed extraction project can wait until the proposed use and review process are clearer.
Sources & further reading
Sources checked October 2, 2026.