The Web, Now Just for You

New online tools offer personalized recommendations to help users find what they like—and help advertisers target their audience

by Catherine Holahan

Kevin Rose likes to talk about his taste for oolong tea. But lately the founder of Digg, a popular bookmarking site that enables users to submit links to Web pages and articles they find noteworthy, has been more vocal about strangers' appetites for, well, everything. Rose has been paying particular attention to which Web sites, stories, and links are gobbled up by the 900,000 registered users who visit Digg—and which ones are cast aside. He plans to use what Digg knows of its members' tastes to help them find more media fare they might want to consume.

Later this year, Digg will launch a recommendation tool able to expose members to fellow Diggers who appear to have similar interests, says Rose. "Digg will be smart enough to know what interests you," says Rose. The site will identify those with like interests in part by the previous stories they have dug and "buried"—the site's term for voting down a story. "Theoretically, over time, you provide useful information," says Digg Chief Executive Jay Adelson. "It makes a personal version of Digg more possible."

Ad-Delivery System

Digg is not alone in working to leverage what it knows about users' tastes to suggest more relevant content. Internet giants Yahoo! (YHOO) and Time Warner's (TWX) AOL, as well as a host of startups, are testing or developing tools to do the same. Earlier this February, Yahoo quietly launched a recommendation feature for its bookmark tool that suggests content related to that previously bookmarked. Similarly, AOL began testing a tool in November that suggests news articles to individual My AOL customers based on what they have read before and what those with seemingly similar interests have selected.

Of course, tools that personalize the Web are not new. For years, people have been able to customize their home pages on sites such as Google.com (GOOG) and Yahoo.com. Similarly, e-commerce sites such as Amazon.com (AMZN) have long relied on recommendation tools to fuel additional sales.

So what's new? Previously, Web sites added automated recommendation tools to their e-commerce offerings to get you to spend more money. Now, it's about getting you to spend more time. The goal of the new generation of recommendation tools is to encourage visitors to stay as long as possible on particular sites and, in the process, deliver as many ads as possible to users—without alienating them.

The "Engagement" Quotient

In the world of marketing, the length of time spent on a site is called "engagement," and it's a key consideration when advertisers determine where to spend their ad dollars. Sites that have high user engagement can gain longer exposure time for ads, making it more likely that a user will click on an ad or absorb a branding message. Sites that offer higher engagement often charge higher rates to show the ads or collect more revenue from click-through advertising.

When sites know enough about visitors to suggest content, the additional ad-targeting capability they can provide is very appealing to advertisers. Theoretically, that knowledge can be used to suggest items users may want to purchase, making advertising more likely to translate into sales.

VideoEgg, a two-year-old company that provides video tools for sites such as social network Bebo, has developed a new version of its media player that recommends videos related to the one the user is watching. The new player, which is set to launch sometime in the third quarter, shows ads related to the video playing. For example, a music video could display an interactive ad at its base selling ringtones by the same artist. The hope, says CEO and co-founder Matt Sanchez, is that the recommendations keep the user engaged with the player, and potentially the advertising, longer. "We want to entertain users as long as they want to be entertained," he says.

Filtering Tools

Many companies have yet to aggressively deliver ads along with their suggestion tools. Sites are still trying to master the delicate balance between a helpful suggestion and an off-putting sales pitch. Too much advertising can defeat the purpose of the recommendation tool by aggravating users and unintentionally pushing them to a competing site, says Tomi Poutanen, Yahoo's senior director for social search products.

The potential for advertising is only one aspect fueling the recommendation trend. Companies see the sheer volume of Web content as creating demand for tools that help direct users to the sites they want as well as entertainment and information avenues they would want if they knew they existed. "With traditional media consumption, you trusted an editor to discover it," says Josh Kopelman, a managing partner at First Round Capital who invested in social-bookmarking site del.icio.us, now owned by Yahoo. "As more and more content gets put online in an unpackaged format, the importance of discovery is going to be extremely important."

StumbleUpon, one of Kopelman's investments, is among the companies tapping this trend. The company, which launched in 2001, has a toolbar that suggests sites relating to categories or words users have expressed interest in. Users can then rate the site by clicking either a thumbs-up or thumbs-down button on the toolbar. Over time, the toolbar learns preferences and can better tailor suggestions to particular users. It also suggests sites highly rated by fellow stumblers whom users identify as friends.

Discovery Engine

The company, which calls itself a discovery engine, has added 1.8 million users in the past two years since focusing on recommendations and social search. "As the amount of content has grown, the need for effective filtering and recommendation engines has grown," says Garrett Camp, StumbleUpon's co-founder and chief architect.

The challenge for companies offering recommendations is to balance that need—and their need to make money through advertising—with users' privacy concerns and changing tastes. It is clear from the millions of folks who personalize their home pages, subscribe to topic-specific news feeds, and engage with sites such as StumbleUpon that they do want the ability to both focus on and discover content they find interesting.

However, many people are also uncomfortable with companies knowing too much about who they are: their likes, dislikes, interests, and habits. In September, three AOL users sued after the company released "anonymous" search data that was, in some cases, traceable to individuals based on the information contained in queries. Similarly, consumer groups have filed complaints with the Federal Trade Commission against tech companies they feel silently collect too much data about their users (see BusinessWeek.com, 11/15/06, "Taking Aim at Targeted Advertising").

Fickle Users

Web companies have long been grappling with privacy concerns but what may prove to be a more daunting task is understanding and keeping up with Web surfers' ever-changing palates. Digg's Adelson has noticed that tastes change depending on the time of day. In the morning, people tend to be more interested in hot-off-the-presses news scoops, meaty analysis pieces, and what is at the top of Digg's home page. By lunchtime, many are looking for easily digestible Internet snacks—entertaining stories or video clips that can amuse them during a quick break from work. If the engine thinks, based on morning behavior, that a user wants in-depth, up-to-the-minute news stories during the lunch hour, the suggestions could become a nuisance.

Despite the challenges, many see recommendations and personalization as the next wave. If 2006 was all about community and user-generated content, 2007 and beyond will focus on individual desires. "I think that the long-term future is personalization," says Digg's Adelson.