SUBSCRIBE NOW!

Vol. 20 No. 11 — December 2003

Q: How did you get into the patent information business?

A: Out of desperation. I spent 3 years at DuPont as a research chemist. Within days, I had a stack of Japanese patents on my desk, but I could only read the figures. So I decided I had better learn Japanese. Later, I did a post-grad with the Japanese government.

Then in 1992, when the Japanese Patent Office released a CD-ROM of patent documents in text format, I got hold of an alpha copy. I wasn't sure what I was going to do with it, but one of the interesting things I noticed was that although anyone could buy the CD-ROM, the JPO wasn't selling back copies. So I started subscribing and built up a collection. I also began learning how to program computers, which eventually led me to machine translation.

Q: So your business is based on a software program you wrote that translates the JPO data into English?

A: Right. My wife and I started Paterra in 1996, and we announced the service on PIUG [the Patent Information Users Group mailing list] in February of that year.

Q: Is it a fully automated service?

A: Initially, there was always manual intervention. We would get an order, run it through the translator, look at it, see what terms were missing, put those into our glossaries, rerun it, and then do some post-editing. In 1999, however, I automated the service. Users now just go to the Web site, place their order, and the documents are translated and delivered automatically.

Q: And you have just launched a current-awareness service called Protys?

A: Yes. Last year, I rewrote the program so that it could automatically index the data as well as translate it. This has allowed us to develop Protys, which is an alerting update database covering 5 weeks of Japanese patents with a lag from publication of less than a week. As a result, users can now also run SDIs on the data.

Q: How much do you charge?

A: It's all pay as you go. A basic machine translation costs $39 per patent. Running an SDI on Protys costs $9, which provides a list of titles. Then, to look at the abstract costs $2 per record, to view the context is $9, and images are $6. To download the full document is $25.

Q: What about competitors? Both Derwent and the Japanese Patent Office offer machine-translation services in this area. Are there others?

A: There are about five players in the Derwent niche, including GLTaC in Michigan and Nick Spencer of U.K.-based Rising Sun Communications. Derwent, by the way, charges around $400 per document, which includes pre- and post-editing and has a 5-day turnabout.

Q: So a customer can pay Derwent $400 and wait for the translation or pay you $39 and get it online immediately?

A: Yes.

Q: Can you say something about Paterra's sales and its customers?

A: I can't give you any sales figures. What I can say is that we have over 200 corporate clients, numbered amongst which are eight of the 10 largest—and 25 of the 75 largest—chemical companies worldwide. I can also say that if you compared what our top four clients spent on us last year with what they would have spent for human translations, you would see a $4 million difference. So there is a huge cost incentive for most of our clients.

Q: It seems like a tough market for smaller players, however. At a recent ACS Meeting, you described the odd events surrounding the USPTO's decision to offer a single-source contract for machine-assisted translations to Derwent. What happened?

A: In May 2000, I got a call from the PTO asking for a trial password. I gave it to them and they placed orders for about 50 documents. What I didn't realize at the time was that despite ordering them, they didn't download those documents. Then in June, they announced a single-source contract to Derwent. The announcement was made on a Monday and the contract [was] awarded on the Wednesday. It was only on Thursday, my logs showed, that the PTO started downloading the documents they had ordered. In short, they hadn't looked at our product before giving the contract to Derwent.

Q: What did you infer from this?

A: I don't have a public interpretation, but I know that the Federal Acquisition Regulations specify some fairly strict rules governing single-source contracts. Basically, you can only do it in very special circumstances. So I hired a procurement lawyer and we worked up a protest. Just before we filed it, we called the PTO and asked, "Do you want to compromise, or do we protest?"

Q: What was the compromise?

A: That we would also take part in the pilot program. I assumed it was fairly likely, however, that the solicitation for the main program would be written in a way that would disqualify us from participation—which turned out to be the case since they included a requirement for human intervention. We had just moved to the fully automated system. So we developed a product that translation companies could use, then contacted the four companies then under contract to the PTO. This meant that while we couldn't participate in the main program, three of the translation partners we contacted could. In the end, one of these companies got one of the contracts, and Derwent got the other.

Q: So the USPTO stepped back from a single-source arrangement?

A: They had to. If there are two capable companies in the marketplace, they have to use them both.

Q: We know small players always struggle to compete against large corporations, but your experience with the USPTO contract and your frequent controversial postings on the PIUG mailing list suggest some of today's larger vendors may be unduly predatory. Last year, for instance, you complained that, as a major sponsor of PIUG's annual conference, Derwent was able to get preferential treatment. More recently, you publicly charged CAS with trying to gain unauthorized access to your Web site, an action you characterized as "an unfriendly attempt to gather information." Do you think the larger vendors are now so powerful that they are able to behave anti-competitively?

A: I could say I agree with you on that, but the question is, what do I do about it? As a small company, there aren't too many options. One, however, is to speak up—which, as you say, I tend to do. When the CAS incident happened, I thought: "There is no reason to be quiet on this. It won't benefit me at all. So don't be quiet."

Q: One reason why other small vendors keep quiet, perhaps, is that they believe public complaints are counterproductive and may alienate users. As the CEO of Univentio, Willem Geert Lagemaat, put it to me, "You don't catch flies with vinegar but with syrup." Does he have a point?

A: I grew up on a farm, and so I know that there are substances that attract flies better than syrup. It's true that some users might get upset, but I've learned that many also like you to be spunky. The fact is that it just doesn't pay to be quiet all the time.

I would note that I find my relationship with CAS and Derwent surprising.

Q: Surprising?

A: I'm surprised that they are so interested in us. The day we launched the machine-translation service in 1996, we received our first order within an hour, from CAS. Two hours later, Derwent placed an order. Whenever we launch a new product or make an announcement on PIUG, I can see from my logs that within minutes these two outfits will have accessed the site.

Q: Why do you think that such large organizations are interested in a two-person company?

A: One reason may be that their markets are saturated, and they view any participation by another company as a threat. The companies that are building market share—presumably by taking it from CAS and Derwent—do not pay as much attention to us.

Derwent, by the way, started its machine-translation business about 2 years after us, and CAS started supplying machine-translated patent titles in CAplus about the same time. To my knowledge, our Protys database is the first commercial online full-text cross-language database, and it competes directly with CAS's CAplus database.

Some independent searchers tell me that they suspect Derwent of trying to take their clients. This would be consistent with Derwent having few options for increasing or maintaining sales other than through smaller lines of business.

Q: Given their powerlessness, then, smaller vendors need to shout hard and loud when they suspect larger players are behaving badly?

A: You have to be strategic about it. You need to ask yourself: Is it appropriate to shout at this point, or would it be better not to? But I believe other small companies are being needlessly quiet. I also think that users are being needlessly quiet. And in the end, we have to rely on the user community to drive this. They need to see what is in their long-term interests and, if they see something they are not happy about, they should speak up.

Q: The main patent information user group is PIUG.

A: Right. PIUG is a very interesting organization. It was initially user-based, and so I think they could play a very constructive role invigorating the interests of the industry. Unless such groups are user-driven organizationally, however, they lose their dynamics and their vitality. PIUG is brushing with that danger today.

Q: By becoming too dependent on sponsorship from the large vendors, for instance?

A: Yes. The danger is that some vendors may start to pull out of PIUG. For instance, MicroPatent started its own user meeting last year, and my reading of the current supporter list for PIUG is that MicroPatent is saying, "OK, we will give PIUG some support, but not a lot." It would be a real loss to the community if MicroPatent pulled out of PIUG, as it is one of the quality players.

Q: Large competitors aside, how would you describe the current environment for smaller patent information providers?

A: There are a lot of opportunities, particularly in automated information handling. Full-text technologies barely scratch the surface today, so as technologies become available, you are going to see a number of small companies popping up.

Q: No shortage of opportunities for new entrants then?

A: Right. Whatever influenced the USPTO to start putting its data on the Web, it has been a very positive development. You now have all this PTO data coming on to the market as a raw material, and anybody can buy it. So anybody can buy USPTO data, and anyone can now buy JPO data.

Certainly, I was able to take advantage of Japanese documents becoming available in machine-readable form. The number of companies that were able to get started with inexpensive PTO data is rather large—although MicroPatent has shrunk their numbers in recent years. And if the EPO were to make its data more available to small businesses, it could unleash a whole new industry of small businesses. Right now, Europe seems to be missing that opportunity.

Q: So the patent offices have enabled a new generation of entrepreneurs to enter the market?

A: Right. They have opened up a whole new industry for smaller companies and allowed fresh blood into the business. Prior to that, companies like CAS and Derwent had established products that you couldn't duplicate.

Q: What's striking here is that while you believe the patent offices are opening up the market, the big vendors complain that their activities—especially the actions of the EPO—are a threat to the industry. Their concern is that by offering patent information freely over the Web, the PTOs will destroy the commercial market for patent information. Do you not see any threat here too?

A: Well, the JPO already offers a free machine-translation service, but I haven't seen much impact on our business. Perhaps because the way it has been implemented means that it works fine if you just want a handful of documents, but if you are in an information center and need to pull in a document, look at it, and then hand it off again, it is very cumbersome. The same applies to the EPO site: Its preview interface is not user-oriented. Moreover, it doesn't offer full-text searching. There is plenty that the industry can do to compete with that.

Q: Presumably then, the larger vendors are more threatened by the actions of the patent offices. I'm told that when the USPTO started publishing patent applications, Derwent had to recruit 20 more people to index and abstract the data. This would suggest that they are facing additional costs at a time when their revenue is likely falling, as users presumably now do some of their searching directly on the patent office sites?

A: Right. And the question arises: Since Derwent machine translates all its Japanese patents and then abstracts and indexes them, why doesn't it make them available online? The only two answers I can come up with are either that they don't want people to see what they are abstracting, or there are fears that putting them online would threaten their abstracting people.

I would add that coming into this industry, one of the things that has impressed me is what an incredible job companies like CAS and Derwent did in their age. It is really awesome how an organization can take millions of documents, index every compound, and put it up on a database in a matter of days or months. It is amazing, but it doesn't have to be the only option.

Q: Presumably, the key moment comes when next-generation search technologies enable users to obtain comparable results on the full text as are possible using the indexing of a CAS or Derwent? Or will that never happen?

A: I think conventional indexing will eventually become obsolete.

Q: So CAS and Derwent are in the wrong business for the future?

A: Well, they have to do something. Since CAS is technically capable, it will probably cope very well, but it will be affected. Derwent is another question.

Q: To come back to Paterra, several people have commented to me that while you are well-established in the Northeast of the U.S., you are not widely known outside the area. Would that be accurate?

A: It's true we are light on the West Coast, but around 15 percent of our sales are in Europe, and this has held steady year to year.

Q: Is there a danger, though, that you could be playing in too small a niche? By specializing in Japanese patent translations, are you not making yourself unnecessarily vulnerable? Why not also start offering, say, Russian, Chinese, or Korean patents?

A: I speak Japanese and Japanese is what I know, so I prefer to stick to Japanese. We are also a two-person company.

Q: You could get some VC money and expand?

A: Venture capitalists have to get their money back in a few years. We are able to do what we do because we don't have to worry about outside investors. Being independent also gives us more freedom to speak out when the situation warrants it.