Last time we learned that while the almost perfect patent search seems possible (at least when looking at the ‘validity search’ use case), it’s not necessarily repeatable due to the prohibitive cost of the sheer man hours required. Perhaps instead, we can decode the DNA of the ‘perfect searcher’ to improve search productivity and quality within the field of patent searching and patent landscaping.
If indeed they exist at all, what kind of personality, experience and profile would a perfect searcher have? One spring morning last year, when discussing our IPR&D Search and Intelligence Solutions with the head of the patent department of a well-known European car company, I found out. In an office on the same floor, I was introduced to ‘Daniel,’ the internal patent search specialist – and the reason why the company believed it didn’t need external support on novelty or Freedom to Operate (FTO) searches.
You might expect Daniel to sit in a datacenter with all the latest technology, subscriptions and search resources at his command. Instead, his likely marginally-used computer gathered dust while he drew daily on his four-decade, highly industry-specific knowledge. His encyclopedic memory of the most-relevant prior art was already indexed in his brain by relevance to most search requests to which he had so far been presented. A lifelong networker respected and well-known within his field, he augments his own memory by picking up the phone and asking the internal experts themselves – and they take his call.
This makes Daniel seem like the perfect searcher to his employer. Re-using his know-how from earlier FTO or novelty searches, he connects the regular tech watches with search requests to answer questions within a few hours, if not on the spot.
However, Daniel is well into his 60s, and his expertise is rare (if you’ve ever met a similarly perfect searcher, please tell us in the comments). How would you begin trying to institutionalise Daniel’s talent for future use, transfer knowledge from his lifetime’s experience, contacts, and insights? In fact, after investing in creating such a great internal resource, how do you mitigate the risk that your IP and R&D information services asset leaves the building each night, perhaps never to return. What’s more, how do you amplify this skillset to ensure you are covering all future areas of relevance? No one person – not even a perfect searcher like Daniel – can span all necessary expertise these days.
How to acquire, distribute and store knowledge is, of course, something that challenges most management teams and educational institutions today more than ever before. In a world of automation, human experts will always have a pivotal role in patent searching or R&D Intelligence, at least for use cases where customers (e.g. patent counsels, R&D managers) require high accuracy and reliable insights. In a future post we will take a closer look at training, skillset, and other important elements that form great patent searchers.
However human beings are greatly limited in their ability to transfer knowledge from one individual to another.
Machines can support this process more effectively. Knowledge can and should be transferred to a technology and/or methodology-driven approach where all the knowledge collected is indexed, stored and quickly accessible for future searches by any member of the team. This enables more productive and high-quality results for the company, while being quick and easy for the human expert to draw on both their own experience, but also the ‘corporate memory’ of relevant past searches – for example drawing on prior art outside their normal field of expertise or someone-else’s great contacts.
So, having a perfect human searcher isn’t enough. However, by combining Mind+Machine you can get closer to creating the ‘perfect searcher.’
I’ve now shared two ingredients of a great search – parallel searching (in our last post) and knowledge management. Next time, we’ll be discussing if there is a clear definition of search quality and will find out that search quality depends on the use case. In some use cases a search output may be considered good, however in other cases the same output would be considered poor. Why is this so?
By the way – Daniel is now retired.
