The rights attached to derived data – i.e. whether the licence to the first set of data allows you to create derived data, and what you can then do with that derived data – can be critical both to the licensee and the licensor.
In a recent piece of (very expensive) UK litigation, the Ordnance Survey (owned by the UK government, it’s the main map-producing body in the UK) sued a start-up called 77m whose projected business was the creation of an address database that would compete with Ordnance Survey’s.
Ordnance Survey’s licensing terms prohibited commercial re-use of its data but didn’t prohibit the use of the OS database to validate the data of 77m’s rival database – a very expensive mistake that could have been avoided by better drafting of the licensing terms.
Given the uncertainty around the property nature of data, the only secure way to arrange for the licensing or purchase of data – whether incoming or outgoing – is to provide for it by contract. But this means that the contract – and what the contract actually says – becomes the linchpin on which your approach to licensing turns.
This is made doubly difficult by the fluid nature of data, a fast-evolving market, and the appearance of use cases which a few months ago no-one would have thought of. Defaulting to stock phrases like “internal business use” or (as happened in the 77m/OS case) “use within your organisation” is, with something as fluid as data, just asking for trouble. This is where it’s really important to pause, think, and bring in some serious expertise if you don’t already have it in-house.
This is the last of the “Making Sense of Data” series – I hope it’s been useful. Let me know if you have any questions.
Still interested in data? Here’s a good article on pricing datasets: https://pivotal.substack.com/p/how-to-price-a-data-asset.