This article originally appeared in the July 2000 edition of Information World Review. It is reprinted here with permission.
Patent researchers are out to map the white space. To go where no one’s gone before. The phrase may send Kathy Flynn of Procter & Gamble running into the ladies room to scream but, as Susan Mendelsohn reports, she’s doing everything she can to ensure that end users at P&G are out there mapping with the best of them.
Kathy Flynn describes herself as "one of those twisted people that actually likes to read patents". A process she herself developed was kick-started when she read a 1968 German patent. Once that happens, she says, you no longer see patents as schlock literature. "For a traditional scientist one of the hardest things to learn is to look at patent literature without feeling that someone should be arrested for writing it. It doesn’t have that nice clean feel of peer-reviewed chemical literature. But master it and you get a bird’s-eye view of what your competitors are doing and how things are moving."
A more prosaic description of Flynn would include her job title, technical intelligence manager, senior scientist, and a description of the company she works for: Procter & Gamble, a diversified consumer products company established in 1837 which markets over 300 brands such as Ariel, Pringles, Pampers and Oil of Olay to nearly 5 billion consumers in 140 countries.
She has been rewriting her job description over the last three years by leveraging the patent searching skills she learned working on agricultural chemicals at Rohm & Haas before joining P&G, and her training in competitive intelligence. One of her specialities is patent analysis from an end-user standpoint, and her day now consists of trying to raise the average skill base of P&G scientists as well as keeping track of the greatest and the latest on the consumer products scene.
From her perspective as an end user, Flynn has seen the traditional boundaries between the legal division, the business information specialists, and the technologists (research scientists) at P&G blur, and she currently liaises between the three groups. "The speed that’s required now has broken down these lines of demarcation," she says. "You’re not just having specialists looking at the data anymore. Those days are over. The job of the information scientists is no longer to be the portal, it’s to be on the bleeding edge with the better tools to mine the data, to make it more accessible."
End user accessibility has become the great defining criterion for electronic patent information, she believes. "The classic thing we have at P&G is to ‘map the white space’, the place where there isn’t any patent art. That’s not what electronic databases do best, it’s what the human brain does best. We need the databases to free us up so that we can see what is missing and make the connections."
Scientists spend too much time gathering information and not enough time doing good analysis, she adds. "So often we default to ‘give me everything you’ve got’, as if someone would deliver this huge vat of information and somewhere in there we’ll find the answer. This is like getting a huge vat of beer. And there’s only one thing to do with that – drown in it. Good analysis intelligence, whether it is from patents, business, or a combination of the two, is more like a single shot of good malt whisky. You take it and you’re not the same person afterwards. You know exactly what happened to you – you can feel it hit. That is what you need.
"The biggest problem for end users is that patent information is very esoteric, very specialised knowledge. The determinant for me is no longer the information, it’s efficiently finding the data you want and doing the evaluation and analysis to influence your strategy and decision- making."
She places the onus to change things on information publishers. "I would fall on my sword for Derwent, but a printout from its World Patent Index (DWPI) makes the Gutenberg Bible look like a paperback novel. End users weave their way through the thing trying to figure out which way is up and, frankly, couldn’t recognise their own mother’s address if they saw it because they get so brain dead after page 3. This is true for all the data. I see it in the business information files, too."
The challenge for the publishers, then, is to better use the tools to do what tools do best, i.e. to free up the human brain. Then the human brain can do what it does best, which is to find unique connections. "If I had a magic wand I’d conjure a product that would give the average technologist a macro view of the data very quickly and very easily. Specifically, I’d like Derwent to partner with a search engine, and accelerate their movement in the direction of accessibility."
Derwent’s Lotus Notes Solution is a step in the right direction, Flynn believes. This allows users to access and manage their own set of data from DWPI or from Derwent’s Industry and Technology Patents Profiles (ITPs). Derwent claims that the interface gives end users the ability to search for and store data productively from day one of access to the system.
In addition to Lotus Notes, Flynn encourages P&G technologists to use text and data mining tools. "I do not read more than 10 hits on a Derwent printout without putting it in a tool," she says. "DWPI is chock full of in-depth, rich information but you cannot read linearly and see patterns. You have to go back and forth, back and forth. I once analysed 1604 abstracts – the number is engraved on my mind – and I vowed that I would never do it again. I don’t care what’s at stake. With the data mining tools, you still need to read the full text of the patent but you know that it is one of the 20 that the top company for the last ten years in your search published and, by the way, it is from their top inventor who has 35 other patent filings. When you can read a patent this way, you have a whole different mindset."
Two of the text and data-mining tools Flynn has recently seen are Aurigin and VantagePoint, both of which were shown at the Patent Information Users Group (PIUG) Conference in May. VantagePoint, developed by Search Technology, comes out of ten years of research at the Georgia Institute of Technology in conjunction with the US military. It is a text and data mining tool for field-defined datasets that mines the text and data side-by-side.
Says Flynn: "You can build correlation matrices between any fields. You can parse data out of fields, for example, in DWPI you could search company code by year, a priority filing, year of publication, a company code to company code, and find out who’s co-filing on the same patents. At any point you can click on a hit to see the full Derwent record. I especially like its ability to group words so you don’t have a separate nodes for ‘polymer’ and ‘polymers’ as in some data mining tools.
Aurigin, from Aurigin Systems, Inc, is the other tool she likes. "Aurigin allows you to do some basic statistics on some of the data, some time slicing. It mines the text and makes some sexy maps that are great when you’re presenting to upper management."
Much, but by no means all, of the patent research at P&G is done using electronic sources. "It’s become very high in electronic very fast, but I have to keep reminding people that electronic databases only go back to 1976," says Flynn. "Typically, if a competitor’s patent has been granted in the European Patent Office you have nine months to oppose it so everyone scrambles around trying to find prior art. If you just stick to patents you’re shooting yourself in the foot. A poster at an exhibit in Australia will do. So, as much as I’m trying to train people on electronic sources there is still a need to master the paper ones."
The favourite sources of electronic patent information in the company are, without doubt, USPTO, the US Patent and Trademark Office web site, and the Intellectual Property Network commonly known as the IBM Patent Server. Why? Because they’re so accessible. "The average technologists like these products because they come to their level. They don’t say, we’re going to wait for you on the mountain-top and unless you talk to your Boolean searchers and do your nested parentheses right, you’ll get lost." She’s especially happy with the USPTO: "It has all the functionality, it’s full text, it’s available online, it’s totally searchable, you can search on claims only, and it goes back to 1976. I find it excellent."
Flynn uses EPAT from Questel Orbit, but wouldn’t recommend it to the average end user because, she says, "it’s still very archaic". And for what she calls "scorch the earth searches", when P&G is starting a project or filing its own patents, users always go to Derwent. The legal division will not accept anything less.
Although USPTO is free, its availability has not adversely affected the use of for-fee services such as the IBM Patent Server and DWPI. In fact, the opposite is true, says Flynn. "My impression is that the availability of electronic information is awakening people to the value of intellectual property. It’s no longer a few of us running around saying we should be reading patents. It’s probably making everybody voracious."
Business information products are an important part of any patent search, Flynn adds, and her favourite online products for this are Wall Street Journal Interactive (www.wsj.com), Financial Times (www.ft.com), Wright Investors’ Site (www.wisi.com), Les Echos (www.lesechos.com) and Lexis-Nexis. "If you aren’t aware of acquisitions or strategic alliances, you can’t do patent analysis. You can’t do this in a vacuum."
As an advocate for end users, Flynn is not as interested in whether suppliers are listening to end users. The question is whether end users get enough opportunities to listen to the suppliers. "End users can look lazy to an information professional, but my argument is that they can’t ask for something that they don’t know exists.
"The biggest challenge to the information publishers is not to do with pricing or service," she stresses. "It’s that all their highly valued information is not being accessed. You can’t continue to have end users who haven’t a clue running amok on the web. You have to figure out how to deal with that. You have to come to their level. You can’t sit on the mountain-top and scoff at their ignorance. Information professionals lament that end users don’t know what they’re looking at, and I have to reply, ‘well, whose fault is that?’"