Quigo: An Israeli Start-up Worth Watching?

Quigo's offering span the entire gamut of online advertising and intelligent information retrieval markets. Quigo has traversed a diverse technology path to come to this present position. They have worked in automated information retrieval, document clustering and domain parking and recently have shifted to text analysis based contextual advertising. And talking of start-up-resilience, the Israeli Daily Ha'aretz reports (Overture picks Israeli startup Quigo ) that the company has just started seeing revenue in the past year, and in the Internet bust period the staff of this company worked without any pay.

So what is the Quigo-advantage? Quigo started with deepweb along with other players like BrightPlanet, Invisible Web and ProFusion. Since then they have re-christened its deepweb crawling and information extraction technology to IntelliSonar (it was called DeepWebSonar earlier). IntelliSonar now offers a range of business-oriented deepweb services and solutions. With IntelliSonar you can define the sources you want to crawl (both free-public or paid), you define the kind of information you want to extract and then Intellisonar will automatically extract, normalize and index that information and deliver that information to the user. So Intellisonar allows to automate your information extraction tasks.

Interestingly the whole idea of deep web has receded from the limelight as even the main-stream search engines have started incorporating some of the capabilities (indexing and searching dynamic html pages, retrieving different documents formats, retrieving information from public databases) offered by deep web searches. So what remain un-searched are the private information databases - LexisNexis and the Dialgos of the world. This can explain the transition of DeepWebSonar to Intellisonar, where the unique proposition for Intellisonar seems to be intelligent and automated information extraction rather than indexing.

But Quigo has been in the spotlight recently not because of its ability to ferret out the deepweb, but for its ability employ the very same technology to provide "better" contextual advertising. AdSonar platform is marketed by Quigo as having the capability of automatically identifying content on the page, automatically matching a relevant ad (banner as well as text based ad) and serving the ads up to the user dynamically.

AdSonar uses a combination of Bayesian-Statistical training for information extraction and also provides support for manually edited keywords. The machine extracted keywords are weighted by relevancy scores, humans can edit and re-calibrate the weights of these keywords or add/delete keywords. After having done the initial training the entire task of relevancy detection can be automated, AdSonar then matches these keywords to the actual listing of ads being served.

The third and final offering from Quigo called FeedPoint is unique in its own ways. This offering comes very close to the idea of "super-imposing" XML-RSS feeds on dynamic websites - idea promulgated in an article that appeared on K-Praxis in June 2003 (Intelligent Information Extraction and Discovery through RSS). Quigo offers FeedPoint as a platform as well as a service-based solution that can help e-commerce and dynamic catalog based sites in automating their search engine marketing efforts. FeedPoint can extract all relevant information from product pages or product catalogs, index these pages in an XML format, automatically generate XML feeds at periodic intervals, these feeds then can be picked up search engine or paid listing sites. FeedPoint is an interesting solution and gives a new twist to the idea information discovery.

Quigo and Applied Semantics: A Comparison

Quigo's partnership with Overture creates a unique contextual advertising combination: Quigo -> Overture -> Yahoo vs. Applied Semantics -> Google. Curiously enough, trajectory of Quigo's growth is very similar to Applied Semantics. Quigo started with information retrieval and unstructured data management technologies, went into deep web, went into domain parking, and now has finally landed in the area of contextual advertising and search engine marketing.

Applied Semantics, as the very name suggests, is well-known for its ontology based semantics technology Conceptual Information Retrieval and Communication Architecture (CIRCA). Applied Semantics, now a Google company, offers solutions in the field of contextual advertising, domain parking along with a broader range of enterprise solutions for automatic categorization and taxonomies. But the company has been more successful in domain parking and contextual advertising than in the enterprise solutions. Its contextual advertising solutions in a way do not require the heavy ontology-based technology, so in a way Applied Semantics was "over-qualified" for search engine marketing but has earned its living out of these businesses and after having being acquired by Google now it is in direct competition with Overture/Yahoo for this space.

Quick Survey of Other Contextual Advertising Technologies and Solutions

There are other companies that are attempting to offer similar technologies and solutions. Notably among them are :WebRelevance and Sprinks . WebRelevance offers solutions that it claims to offer similar kind of contextual targeting - similar to ones offered by Quigo and Applied Semantics. Sprinks' offerings, on the hand hand, are more focused on keyword bidding for advertisers. Sprinks's tools can help advertisers in placing targeted ads on various sites. It is important to note that Sprink is a part of Primedia - a targeted media company and a leader in niche markets. Both of these solutions - technologically speaking - come nowhere near to that of Quigo and Applied Semantics's. In the keyword marketing space, besides Sprinks and WebRelevance there are a number of other players (GoToast, RichLinks, eZula, Connextra, etc.) who offer services-based offerings to this target segment.

Conclusion

The future of contextual advertising is likely to be dominated by these players for the time being. And despite the limitations of language processing technologies, keyword-based approaches allied with some machine learning methodologies seem to work quite well for the present needs of the advertisers and the companies like Google and Overture. Nevertheless, an interesting space to watch in near to immediate future!