To find a Web page you wanted in the pre-Google era, you often had to guess at its address. Was General Motors generalmotors.com, general-motors.com, or gm.com? This led to all kinds of trouble—a speculative bubble in domain names, huge legal battles over valuable addresses, and unseemly attempts to benefit from the confusion. (Accidentally visiting whitehouse.com would bring up stuff that was much less safe for work than the Starr Report.)
In 1997, an entrepreneur named Keith Teare came up with what he considered a simple solution to the problem. His company, RealNames, sold natural-language keywords that would work as Web addresses. General Motors, for instance, could buy "General Motors," "GM," "Chevy," "Chevrolet," and so on. RealNames signed a deal with Microsoft to build its keywords into Internet Explorer, and RealNames became a dot-com darling. Microsoft bought a 20 percent stake in the firm, and there was talk of an IPO. But even though it espoused simplicity, RealNames was a bit kludgy; it didn't work across browsers, and it created a single point of failure on the Web (if RealNames went down, all keyword-based addresses would stop working). Soon Google came along with a much handier way to find things online. In 2002, Microsoft dropped RealNames from IE. Just like that, the company was toast.
I got to thinking about RealNames as I was looking into the recent boom in URL shorteners—sites that turn unwieldy links like www.slate.com/id/2220818/ into digestible strings like bit.ly/zfFbA. Thanks to the popularity of status updates—not just on Twitter but also on Facebook, Gmail, and other sites—shortened URLs are increasingly prevalent. By some counts, there are more than 100 shortening services online. In March, Bit.ly, the most popular URL shortener, raised $2 million in venture funding. The news earned its investors some derision, because URL shorteners have a big problem—none of them has figured out a way to make much money.
Shortened URLs also don't sit well with people who worry about the health of the Web. In a much-cited blog post, Joshua Schachter, the founder of Delicious, recently accused these services of disguising spam and phishing sites and littering the Web with loads of perishable links—if Bit.ly goes under, for example, all of the links it ever generated will stop working. Schachter echoes the criticisms that people once hurled at RealNames, and I think he makes a good case. URL shorteners seem like a hack, a short-term solution to a problem that could likely be addressed in better ways. The most obvious fix: Twitter could change its system to allow for links of any length. If it did that, URL shorteners would meet the same fate as RealNames—they'd be rendered unnecessary overnight. But Bit.ly, the URL shortener with the biggest dreams, doesn't see that as much of a danger. The firm says it has gone to great lengths to address its critics and that it has ways to crush spam and create tiny links that won't rot over time. Bit.ly also has a few plans to turn small links into big money. The ideas don't sound too crazy, either.
Shrinking Web addresses isn't a new practice. TinyURL, the first shortener, hit the Web early in 2002; its creator, Kevin Gilbertson, is a unicycle enthusiast who came up with the idea after many frustrated attempts to pass around unicycling links. (Early e-mail clients and newsgroup readers would break up very long links.) Several upstarts have tried to take on TinyURL over the years, but shorteners have never been seen as a hot business. TinyURL makes money from Google ads, which yield enough to cover the costs of the site; so far, no shortener has come up with a more lucrative revenue stream. Eric Woodward, the CEO of Nambu, says his firm created the popular shortener Tr.im mainly as an adjunct to Nambu's main business goal—building a great desktop Twitter program. The company has just one developer working on Tr.im, and Woodward doesn't have any brilliant plans to turn a profit from the site. "If you want to tell me how to make money from it, I'd be happy to listen," he told me.
Tr.im is among a new generation of shorteners that do more than just squish links—they also give you statistics about who's clicking on the URLs you generate. Bit.ly does something similar, and it's in these stats that the firm sees its fortunes. Bit.ly was created last summer by developers at Betaworks, a tech "incubator" in New York that has deep business connections to Twitter. A representative for Bit.ly—which has since been spun out of Betaworks—declined to speak to me on the record but was willing to talk at length about the company and its plans if I agreed not to quote him.
Bit.ly prides itself on the stability of its infrastructure. When you shorten a URL, the system stores data for the link in several different places across the Web, a level of redundancy that addresses Schachter's concerns about link rot—links created by the service won't go down anytime soon, Bit.ly says. Every time someone clicks on a squished URL, the link shortener's servers must translate it into a full address and redirect your browser to the new site. Bit.ly claims that its system for handling these requests is much more durable than those of other shorteners. In May, Twitter changed its Web site's default shortener from TinyURL to Bit.ly; as a result, Bit.ly saw its market share rise sharply. Now, according to Tweetmeme, more than 60 percent of the links on Twitter are shortened with Bit.ly. TinyURL has about 30 percent of the market, and every other shortener has less than 5 percent. The Bit.ly rep told me that the service now gets more than 150 million clicks on shortened links every week and that it can handle many more times that level of traffic.
How do you make money from hundreds of millions of clicks? One plan is to sell stats. At the moment, Bit.ly gives you real-time traffic data for every link you shorten. Ordinary people find this fun—you can immediately check to see how many people clicked on what you just sent out. For advertisers who are using Bit.ly to send out campaign messages, these real-time stats have monetary value. Companies have asked Bit.ly for much more detailed stats—the ability to test which specific slogan results in the most clicks, for instance. In the future, Bit.ly could make these in-depth stats available only to those companies willing to pay a subscription fee.
Another potential revenue source could be a content site. Because Bit.ly knows which videos, news articles, songs, and photos are getting linked to and clicked on across the Web, it can automatically spot trends before human-edited aggregation sites like the Huffington Post. It would feature all this stuff on a single site—and, of course, it would sell advertising on the page.
Bit.ly's grand plans all rest on a huge assumption—that ordinary Web addresses will remain long and that people will always need a way to shorten them. It's not clear that's the case. About half of Bit.ly's users shorten links just for Twitter, which—because it began mainly as a text-messaging service—imposes a limit of 140 characters per tweet. But as Schachter points out, Twitter no longer needs to stick to that limit, as few of its users continue to use text messaging as a means to tweet.
Not many people are calling for Twitter to drop the 140-character limit—many argue that imposing such a cap keeps Twitterers sharp. But what if Twitter simply stopped counting URLs toward the limit? This wouldn't impose much of a technical burden. It would also make Twitter much more user-friendly—people who read your Tweets would see meaningful Web addresses rather than cryptic strings. And it would make links more stable, removing an unnecessary middleman from the process.
Nobody knows what Twitter will do; the company may very well decide to let shorteners run free, and Bit.ly may yet see its fortune. Of course, that's what RealNames expected. Investors, beware.