Since the Web entered mainstream usage, retailers have been intrigued -- some would say obsessed -- with the idea of using the Internet to attract customers and sell goods. However, companies quickly realized that customers couldn't always find the right products using site navigation, drill-down menus, and search. Many turned to the "recommendation engine" developed by Net Perceptions. In this installment of Where Are They Now, we'll examine the history of the company, from its early Web-focused roots to its eventual shift into manufacturing.

Founding: Net Perceptions launched in 1996, with a mission to sell its recommendation engine technology to e-commerce websites. It was co-founded by CEO Steven Snyder, previously a vice president at employee talent management company Personnel Decisions International. Other co-founders included John Riedl, Brad Miller, Joe Konstan, and David Gardiner. The company raised $12 million from investors including Hummer Winblad Venture Partners, St. Paul Venture Capital (now Split Rock), Jafco (now Globespan Capital Partners), Paul Allen's Vulcan Ventures, and Berkeley International Capital.

History: During the infancy of e-commerce, vendors would place products in an online store and hope that buyers would be able to find and buy the products that best suited them, using navigation, search, or promotions. As the field became more competitive, a new technology appeared. It was called "collaborative filtering," and was created by John Riedl through a research program he started at the University of Minnesota in 1992. Collaborative filtering let websites make product recommendations to site visitors. Riedl was a co-founder of Net Perceptions, and became known as the "father of collaborative filtering" according to its former CEO, Steven Snyder.

Snyder describes the technology in an email interview with The Industry Standard. "We called our product a 'recommendation engine' because it allowed a site (like Amazon) to make personalized recommendations to each of their customers," says Snyder. "These recommendations were uncannily accurate -- and customers found them very useful. Recommendations were made by finding a group of 'similar others' within an online community. The engine would use these similarities to make recommendations of what a person might find interesting or appealing. This facilitated the type of one-to-one marketing that was greatly discussed in the early days of the Internet."

Snyder says the technology offered two important benefits -- increased revenues, and improved customer experiences. The company's recommendation engine was sold to over 250 sites. "We also expanded our customer base to large catalog merchants who used our technology to achieve add-on sales in their inbound call centers," he adds.

The company's pioneering technology led to a successful IPO in April 1999, which raised $40 million, along with another $85 million in a secondary offering in March 2000. The company reached its highest valuation of $1.5 billion in February 2000.

According to Snyder, the biggest challenge was growing Net Perceptions' internal infrastructure to support its growing base of customers. For e-commerce sites, integrating personalization technology with legacy systems often proved challenging.

While hype helped boost Net Perceptions, the fallout from the bursting of the Internet bubble was brutal. The price of the company's stock plummeted from $57 in March 2000 to less than $1 in March 2001, and the company cut approximately three-quarters of its workforce. Snyder states that the company's problems mostly arose from market forces, and is sharply critical of the investor mentality at the time. "I think that the biggest ‘failure' was a failure of the capital market system," Snyder says. "Many stakeholders contributed to this failure -- overhyping by investment bankers, statements by Alan Greenspan, and of course, the desire of investors to ‘get in' on a hot investment. All these players contributed to what is now called the ‘Internet bubble'. Essentially, the market became so overhyped that the system was destined to collapse. That's what happened beginning in October 2000. This was not specific to our company -- but, instead was systematic throughout the entire industry."

What Happened: Net Perceptions never turned a profit in the early years, but it did manage to conserve its cash and investments, which helped it survive the bursting of the tech bubble. According to Snyder, the board considered multiple acquisition offers as late as 2003, but ultimately decided to do the following: Return the remaining money to the shareholders, sell its patent portfolio to Intellectual Adventures, sell the corporate shell of the company for the tax loss that could be carried forward, and allow a core group of developers to continue to service customers on a consulting basis.

Where Are They Now? Steven Snyder now runs Snyder Leadership Group, which provides leadership and organizational consulting. He also teaches Business Ethics at the Carlson School of Management at the University of Minnesota. John Riedl is back on the faculty of the University of Minnesota, where he still does research on collaborative filtering along with Joe Konstan. Brad Miller now teaches computer science at Luther College in Iowa.