VentureBeat

Baynote, a company that provides recommendation technology for web publishers, businesses, and others, is launching a new software service — what it calls its Collective Intelligence Platform.

The Cupertino, Calif. company has been developing software that tracks user behavior on a web site or other communication channels like email, identifying then sorting the variations in visitor behavior into a bunch of different categories. So if a user comes to a site and starts randomly clicking around on front page articles, those actions will indicate to Baynote that they’re new to the site and looking for general information. Baynote then serves up what it considers to be the most relevant articles from the site, in an interface like a sidebar window of recommendations.

On the other hand, if a person comes to a site and clicks only on articles about a particular topic, or does a search on the topic, it indicates that they’re on the site for a closely-defined reason. Baynote will serve up closely related articles and other information.

See screenshot below for an example of user behavior differences within a site about children’s health.

Today’s news is basically that the company is taking this technology and turning it into a platform for publishers to analyze a wide range of services. Its new email recommendation services let a Baynote client display targeted ads or products in emails to customers. Its new mobile service, like its web service, makes content recommendations based on visitor behavior.

Another part of the new platform, “Baynote Insights,” promises to give more detail about a site’s community. An analytics service, it lets publishers see specific behaviors of new groups of visitors on a site, behaviors of groups of existing users, and more. It also lets publishers track how Baynote targeting affects pageviews, engagement, clicks, and changes in revenue. Finally, a publisher can also use the platform to customize Baynote recommendations on a site to emphasis specific types of content.

Baynote competes with Aggregate Knowledge and Loomia, both of which use their own differing methods of targeting content — all of these companies have been reporting a strong reception from large clients over the past year or so. However, chief executive Jack Jia tells me he sees the most direct competition coming from “homegrown solutions” that companies build on their own.

Baynote wouldn’t disclose revenue to me, but said it was “significant” for a 50-person company. It already has many brand-name clients, including Expedia.com, NASA and Symantec.

Aggregate Knowledge, which tracks Internet user surfing patterns within a site in order to sell them more, has raised $20 million in venture capital, as expected.

We reported about Aggregate Knowledge’s progress in February, including early details of this round. Venture capitalists valued the company higher than $70 million (post-investment), the company’s chief executive Paul Martino confirmed. DAG Ventures led the round, and was joined by Kleiner Perkins. Martino, an affable fast-talker, also co-founded social networking company Tribe.net. Unlike that company, which struggled, Aggregate Knowledge is making millions of dollars after less than two years of business. It serves 15 customers, including Overstock and Washington Post.

It works by offering products, news content or other material to users based on what previous users expressing similar behavior have chosen to buy or read. Competing start-ups like Wunderloop and Baynote pursue similar strategies (see VB coverage of Wunderloop and Baynote). Aggregate’s Martino says he is targeting industry leader Omniture, a public company that bought Touch Clarity just as Aggregate Knowledge was closing its venture round several weeks ago. Omniture uses both Web analytics and behavioral targeting — and that’s where Aggregate Knowledge also wants to go, said Martino. Unlike Omniture, he said, Aggregate Knowledge has negotiated the right to use customer data across sites. So if you’re reading an article about a baseball game on one site, AK can recommend a ticket from another site.

Leading the round is DAG Ventures, which has become known as “coattail ventures” because it loyally invests in companies already backed by successful firms like Kleiner Perkins, Sequoia and Benchmark. Kleiner Perkins led AK’s first round. The latest round brings total investment in AK to $25 million.

Wunderloop, of Germany, recently unveiled its WunderLOOP’s Connect, a behaviorally targeted online advertising exchange. It allows publishers and advertisers to access user profiles to maximize ad targeting possibilities.

Baynote is yet another company boasting that its customer behavior-tracking technology can boost revenue for Web sites.

It says it has increased revenue for some customers by up to 20 percent.

It recently raised $10.75 million in venture backing, (see our recent coverage), but we only yesterday got to see a demo of the site.

Baynote’s technology seeks to quickly get people to pages they’re most interested in. Example: One of Baynote’s customers is StressCenter.com, a site centered around advice given by stress counselor Lucinda Bassett. If you type “StressCenter.com” in your browser (see first screenshot below), there’s a high likelihood your interest is of a general nature. Baynote knows that you’re more likely to first seek a self assessment test, or a tutorial about how the StressCenter program works. Baynote knows this by analyzing behavior of past visitors, watching how much time they spend on each page, for example.

However, if you type in “Luncida for stress,” in Google, you’re more likely to be interested in things directly related to Lucinda. Indeed, Baynote finds that people using this search are more likely to spend time with things such as Lucinda’s testimony about how she overcame anxiety and depression (see the screenshots below, for how Baynote produces a menu box that caters differently to each visitor).

Baynote is also applying this to e-commerce. If you’re buying a washer, for example, Baynote will show you washers that are similar to the one you’re browsing on your retail site. It will show you washers that other people have spent time considering (Baynote tracks customers as they surf back and forth between washers). Baynote says it observes 20 types behavior patterns. Aggregate Knowledge is a competitor, showing fast growth, and doing something similar. Baynote offers some features that Aggregate Knowledge doesn’t. With some customers, such as PoliceOne.com, Baynote’s software gives you a menu page with suggestions of pages you may be interested in — and pops up previews of the pages when you scroll over the menu, thus saving you time. (If you go to PoliceOne.com, you may not see this feature; Baynote splits visitors into two groups, one who get to use the feature, and another that doesn’t, for study control purposes).

Our conclusion: This provides a lot of use for retail and information sites that until have relied on regular old (and dumb) key-word search technology (whether Google, Verity or other corporate search product).