|Created as a hobby Web site in 1996, PatentCafe quickly became a must-visit
location for anyone interested in intellectual property (IP). PatentCafe
CEO Andy Gibbs talked with me about the service and said that with the
launch of IPSearchEngine, it now poses a direct threat to heavyweight information
providers like Derwent, MicroPatent, and Delphion.
Q When and why did you set up PatentCafe?
A It began as a hobby in 1996. While running
a previous company—which manufactured ski and bike racks for the auto industry—my
brother and I had accumulated a number of patents. To support our patenting
activities I began collecting IP-related data as part of a personal Web
Q When did the hobby become a business?
A Initially the service was just a Yahoo!-like
directory that cataloged reliable resources of information. But over time
the content and services grew significantly, and by October 1999 it had
become the go-to place for patent and intellectual-property information
on the Web. It was also occupying around 15 hours a day of my time, so
I decided either to make it a business, or back it down to a manageable
hobby. As it turned out, I had the company funded, staffed, and up and
rolling within 90 days. It was launched officially as PatentCafe in January
Q What does PatentCafe consist of today?
A Currently we have seven Web properties.
This includes the original directory, which has become the largest IP-related
Web directory in the world; an IP magazine; a news service; a bookstore;
and a portfolio of IP-specific Business Centers dedicated to patents, copyright,
trademarks, trade secrets, etc.
Q Despite the name, PatentCafe covers more
than patents then?
A Yes. Most of the emphasis is on patents,
but we cover all types of intellectual property. After all, competitive
business positioning in the global economy has to be based on intellectual
property, not on patents alone. Being broad is easy. Being deep is difficult.
Being broad and deep—in content and resources—is the most difficult of
all. But that is what PatentCafe aims to do.
Q So PatentCafe is an IP portal?
A I prefer to view it as an IP products-and-services
distribution channel or, so to speak, the "railroad" for intellectual property
on the Web. Today we have a channel to most of the patent offices worldwide,
and to inventors and to universities globally. As such, we can deliver
any IP-related product, any service, any news or information, and act as
a facilitator of commerce in both directions. In short, we are now the
largest IP distribution channel on the Internet.
Q You also recently entered the patent information
A Right. IPSearchEngine—which was officially
launched in September 2000—is probably our most important new launch since
the company was founded. Among other things, it offers a single point of
access to the world's leading patent and trademark databases, including
the databases of the USPTO, EPO, JPO, Canada, etc. Users can search on
more than 40 million patent and non-patent prior art database citations,
and we have just added another 2,000 more collections specific to the genetic,
energy, and chemical-research areas—bringing another 60 million patent
Q A patent metasearch engine then?
A Not a metasearch engine, per se, since
that implies more than one database is searched simultaneously, and multiple
sets of data retrieved, sorted, and presented. With IPSearchEngine, a collection
of databases can be searched individually—with search results being returned
in separate windows, one from each database searched. As such, patent and
non-patent art can be searched using independent search queries, with the
results displayed in multiple windows and then compared, analyzed, etc.
Q How does PatentCafe make money?
A IPSearchEngine was relaunched as an ASP
service in April. At that point we moved to a subscription-based model.
We also sell patent analytic models, and we sell books, tapes, and software,
etc. We also license the IPSearchEngine software as the intellectual-property
backbone to enterprise-level intranets.
Q You liken PatentCafe to a railroad. As
online hosts like Dialog and STN International integrate themselves into
the Web and link out to third-party data, are they not also becoming railroads?
A IPSearchEngine doesn't only provide access
to data, it also provides a comprehensive collection of products and services.
In this sense it is an IP management tool, rather than simply an IP data
content location service. Unless and until companies like Dialog make a
significant change in their method of business, they will never really
be a railroad in the sense I am talking about.
Q So, for instance, as well as search you
offer analytical capabilities?
A Right. We offer an analytics tool called
CafeDoors. This allows users to analyze a particular patent with regard
to prior and subsequent art, both cited and non-cited. Further, it will
assess concurrent art—that particular collection of issued patents that
were simultaneously pending and remain in "citation limbo." No other search
tool provides this level of real-time analysis.
Q Is this not similar to the patent mapping
services offered by companieslike Aurigin?
A Sort of. If you were to take the most
important, salient features of Delphion's PatentLab-II, BizInt Smart Charts,
and Aurigin, and then add in the ability to analyze concurrent art (none
of these competing products can do a validity analysis on a patent that
assesses other non-cited patents that were pending at the time the subject
patent was in prosecution phase)—then you have a more definitive comparison.
But the biggest comparison point between PatentCafe and all of the competition
today is that we can offer what they offer for thousands of dollars less.
Q What is your pricing structure?
A IPSearchEngine costs just $150 per seat,
per month, which is considerably cheaper than most other information companies.
Compare this, for example, to Derwent's pricing, which is arcane in its
complexity, unpredictable, and significantly more expensive. Similarly,
it costs just $35 for each patent analyzed with CafeDoors, and they can
be analyzed six different ways. Contrast this with a minimum $7,000 access
to BizInt, or the $100,000 licensing fee required for the Aurigin patent
Q Who uses PatentCafe, and how many users
A Everybody from independent inventors,
engineers, and scientists, to researchers, government organizations, and
the small independent invention shops. We have on average about 350,000
discrete users per month.
Q Is PatentCafe a service of interest to
information professionals and librarians?
A Absolutely. It provides a unique collection
of more than 600 databases. While it will not replace Dialog or LexisNexis,
it does a fine job in augmenting them with more than 100 million searchable
citations and data records in four categories—for a small fraction of the
Q Whom do you see as your primary competitors?
A We don't have competitors in our core
model of being an IP products and services distribution channel, or railroad,
so they have to be defined in terms of each of our different Web properties.
Q So who are your main competitors in the
patent information business?
A In terms of what we do with IPSearchEngine,
clearly we are talking about companies like Delphion, MicroPatent, Chemical
Abstracts, and Derwent—all of whom provide patent information in one form
or another. However, when we compare ourselves with these kinds of companies
we view them as mini PatentCafes, in the sense that they are little more
than special-purpose cog railroads that service a regional line (albeit,
a very high-quality, specialty regional line). We, on the other hand, own
Q Unlike these companies, however, you don't
own, or even host, content; you merely point to third-party data?
A No matter how much data you own today,
you cannot ever own enough to become a sole source, which is the basis
on which traditional information providers operated. Since you can't own
all the data, you inevitably have to provide access to all of the rest
of the data that you don't own. We don't want to be in the business of
owning data, as it is merely an anchor around your neck.
Q Nevertheless you don't offer access to
a number of key sources of patent information—Derwent's World Patents Index,
for instance. Is this not a significant gap in your coverage?
A It's true that we don't provide access
to Derwent's content today.
Q IPSearchEngine's strength, then, lies
in its ability to search on third-party data (much of it raw patent documents)
and then to analyze that data on the fly. But can this really match the
precision of a search on, say, Derwent's manual codes using the complex
search commands of a traditional online host?
A Absolutely. CafeDoors will search and
analyze in a very sophisticated manner and, as you say, in real time. It
actually de-prioritizes the citations and international patent classifications
in patent documents, and reads and analyzes the claims themselves using
sophisticated semantic language filtering. As such it can also find mis-classified
patents, and it will find patents in completely different industry segments
that show the same claim language being applied.
Q So there is no longer any need for companies
like Derwent to manually index patent documents?
A That's right.
Q What are the main challenges faced by
patent information providers today?
A Clearly the fact that patent offices are
now freely offering their data over the Internet has made patent information
a commodity, and so is a significant threat to traditional providers.
Q What trends can you see in the use of
the patent system?
A The patent process is getting increasingly
complex and expensive. Consequently, what companies are discovering is
that patenting is not always the answer. While every company would love
to patent every idea that every engineer in their organization comes up
with, the economicssimply don't support it anymore. For this reason there
is a growing trend toward putting inventions into the public domain through
technical disclosures, rather than patenting them.
Q Which is presumably why new Web-based
services like IP.com, Delphion, and PatentCafe are keen to provide access
to technical disclosures as well as patents?
Q Yet traditional providers have not rushed
to add non-patent prior art to their services. Have they missed a trick?
A They have. The explosion in Web content
and non-patent prior art means that companies like Derwent now provide
a very small window to a huge mass of data that they simply can't analyze.
Even if you use Derwent 100 percent, it is only 20 percent of what you
have to access if you are going to do a proper prior art search.
Actually, the slow response of traditional players would have been predictable
given their conservative nature. They just don't have the fire, the spark,
or the real need to break outside of tradition. The fact is that PatentCafe
is moving far more aggressively than large established players like Derwent
—where managers enjoy 8-to-5 work lifestyles—ever could.
Q In the wake of the dot-com crash, however,
some might argue that new entrants like PatentCafe will turn out not to
have staying power.
A I've been around the block a couple of
times. This is my seventh successful company, so I have absolutely no doubt
Q Bottom line: What is it about PatentCafe
that will ensure its survival?
A CafeDoors and IPSearchEngine provide the
single most complete access to the broad patent and non-patent art when
compared to any other service, and our search and analysis produces results
that are more precise, for less money—and without any complicated learning
curve—than any single prior art analytics tool available on the Web. And,
as I said, we own the railroad now.
Richard Poynder is a freelance journalist based in Oxfordshire, U.K.
He writes for numerous online publications as well as the London Financial
Times and The Wall Street Journal Europe. His e-mail address