A data licence is the contract that gives a buyer the right to use a dataset and sets its limits: which data, for what purpose, by whom and for how long. It governs use, not only access, so holding the files does not carry the right to build them into a product or pass them on.
The three uses
- Internal use: staff and models inside the buyer's organisation use the data, and nothing leaves it.
- Embedding: fields from the data appear inside the buyer's own product or service, shown to its customers.
- Redistribution: the buyer passes the data itself, or a substantial part of it, to third parties. See redistribution licence.
Each use needs the licence to say so. A licence for internal use does not cover embedding or redistribution.
Terms to settle
- Scope: the datasets, the set of companies and the fields.
- Term and renewal, and who may use the data, including affiliates.
- Derived data: whether scores, models and indexes built on the data may be kept or shown after the licence ends.
- Attribution, delivery and versions.
- What must be deleted at the end, and any audit right.
How terms travel with the data
Terms that bind a source pass through to the buyer. Where a licence requires a credit, the credit sits in every row of the data, so it travels with the data into the buyer's tables.
Fokals bulk exports come with a manifest that names the sources, the period, the label versions and the licence, so a data team can store the terms beside the files.
In Fokals data
Fokals licenses data by written agreement, for internal use, for embedding in a product or for redistribution, with scope and terms set in the agreement. Sample data for the companies you track is sent on request with the data dictionary and methodology. The data licensing page is written for teams that build on the data, and what a data licence covers goes through the three uses in detail.
Related terms
- Redistribution licence: the licence to pass data on to others.
- Data manifest: the file that records what a delivery contains.
- Third-party data: data a vendor supplies from outside your own systems.
- Data marketplace: a venue where providers list datasets and buyers acquire them.
Frequently asked questions
What is the difference between a data licence and a data subscription?
A subscription describes how access is delivered and paid for, usually by period. A licence defines what you may do with the data. A subscription without clear licence terms leaves open whether you may embed the data in a product, share it with an affiliate or keep it after the term ends, so read the two together.
What is a data redistribution licence?
A redistribution licence lets the buyer pass the data itself, or a substantial part of it, to other parties, for example by selling it on or including it in a data product. It is wider than a licence for internal use or for embedding, and it typically adds terms on attribution, versioning and what the recipients may do.
Can you train a model on licensed data?
Only if the licence allows it. Training creates derived data, and many licences deal with it separately: whether a model may be built, whether it may be kept after the licence ends and whether it may be shown or sold to others. Settle derived-data terms in writing before the first training run.