Tuesday, 29 September 2015

Understanding IP: The Big Data solution

In a contribution to ITPro Portal this weekend, Aistemos CTE Steve Harris writes:
A great deal of intellectual property data is available today, including patent, trade mark and design data from the 200 IP registries around the world and litigation data from the courts [This sounds really useful, but ...].

Time pressure can lead
to record-keeping errors
Patent registries across the globe work more-or-less independently, so there is no one view of individual inventions protected in multiple countries, and no clear view of ownership [since the same IP right may be owned by different entities in different jurisdictions]. The information relating to ownership of IP assets is typically poorly normalised, with no consistent identifiers [unsurprising, if you consider that some corporate names cannot be easily replicated or transliterated into other languages], and is often out of date [an IP right may have been invalidated, assigned or licensed a considerable time before the official records are updated]. More than 1,500 patent applications per day were sent to the US office alone in 2014 [even assuming a hour working day, that's 150 applications an hour, or one every 24 seconds], and any of them could have a profound effect on a sector, so tracking changes in this area can be critically important.

Identifying IP assets is only the starting point. To build up a complete picture requires the consideration of all data that impacts the value or risk of the assets – for instance ownership, licensing and litigation data. Once this data is assembled we can apply machine learning and data science to summarise, identify trends, spot anomalies, and find stable and reliable indications of risk and value [the downside being that this data assemblage has a limited shelf-life and may decay quickly if any one of the components that help make it p turns out to be volatile]
Turning to analytics, which is a polite way of saying "the way you squeeze a bit of sensible meaning out of Big Data (whatever that means)", Steve continues:
Analytics should be treated as the starting point for human interpretation, not a replacement for it and, by extension, the better the data and analytics, the better the basis on which human interpretation of the facts can be built [it cannot be strongly enough emphasized that analytics isn't a panacea; it's a process that leads to a data-driven product that helps humans make decisions but doesn't do that for them].

A Big Data solution provides immediate access to business intelligence related to IP, leaving it for users to make informed decisions based on those same facts, circumstances, needs and requirements. Knowing where your competitors are investing lets you counter future changes of direction, and knowing where your IP portfolio is strong or weak lets you know how you sit in the marketplace, and allows you to foresee future litigation risks.
You can read Steve's article, "Understanding IP: The Big Data solution", in full on ITPro Portal here.

If you want to do something about making patent records more reliable -- you can!  Visit the ORoPO Foundation website and see for yourself.

No comments:

Post a Comment