Published: 03 Oct 2007

Proximic, which has offices in Munich and Palo Alto, Calif., is offering a way for website publishers to display related information from other sources and ads, as well as to syndicate their content to other sites.

As per the information available, Proximic uses a widget, a piece of embedded software for adding those little windows of features. Proximic's widget matches a particular web page to other content—from the same website, from news sites, from Wikipedia, and from the across the web—that is constantly updated. This is particularly useful for news sites, where information is constantly changing.

Rather than requiring advertisers to buy ads based on keywords or categories, the system automatically reads and matches relevant content according to "interconnected patterns" in the documents that "are not even based on words," whatever that means.

The company—founded in 2001 and backed by Munich-based Wellington Partners and Holtzbrinck Group, publisher of Scientific American—announced two major publishing partners Monday: London-based The Independent newspaper and Nature Publishing Group.

Proximic CEO Philipp Pieper says the company's technology is fundamentally different.

According to media, Proximic's matching technology is different from both Google's keyword system in its Adsense programme, as well as other hot areas like semantic search. Instead of matching an ad based on a particular keyword, Proximic's algorithm essentially processes the characters on the page, parsing out unrelated content on the page, to provide the best match. The widget also provides relevant ads—using the same matching system—together with the content, under the header "infotizements."

This on-the-fly processing is independent of the language on the page, allowing Proximic to match nonstandard language such as obscure gamer nomenclature or certain types of slang. This also allows Proximic to store much less data about each web page, allowing quicker searches, Pieper added.