URL Shortening Service is to turn long links into short ones have suddenly come into their own.
Universal Resource Locators – URLs, or web addresses – looked really simple when research physicist Tim Berners-Lee thought them up in 1989. Then database-driven content turned URLs into lengthy strings of gibberish.
Enter the redirection services. They are just huge, simple databases. They trade on the combinatorial power of the alphabet plus the 10 digits: with a “redirection” link just one character long, you could represent 36 possible URLs. Use two, and you have 36 squared, or 1,296; by the time you’re using six characters, the possibilities are in the billions.
Small is beautiful too
So: take a huge, 120-character URL and feed it into a dialog box on the shortening site. It first looks in its database to see if it has processed that URL before. If it has, it gives you back the previous answer. If not, it generates a new shortened URL on the fly. Put the shortened URL in your browser’s address bar, and the site looks in its database for the matching URL and redirects you there.
The first of today’s URL redirection services were set up in 2001: MakeAShorterLink (masl.to) – which has since been bought by the biggest such service, TinyURL – and SnipURL.
Until the beginning of last year, URL redirection services saw steadily though modestly growing traffic. In March 2004, Wired reported that TinyURL was generating 8,400 shortened URLs a day and 80m page views a month. The Wayback Machine shows that by 2005, TinyURL had shrunk about 5.5m links. In August 2007, it was 42m. Now, it’s 90m.
It was the arrival last year of Twitter, the microblogging site, that transformed redirection sites from merely useful into absolutely essential.
Twitter is blogging for the SMS generation. It enables the “smart mobs” Howard Rheingold wrote about in his 2002 book, Smart Mobs: The Next Social Revolution.
Rheingold has said he’s hooked on Twitter because it’s “a rolling present”.
Marshall Kirkpatrick, a consultant based in Oregon and lead writer for readwriteweb.com explains:”There are lots of other circumstances where shorter URLs might be a good idea, but on Twitter you have to have them because of the 140-character limit.”
The experience of Shashank Tripathi, founder of SnipURL, backs this up: traffic through SnipURL has nearly tripled since Twitter went live. There are as many as 70 of these services, all free, all with increasing traffic. Some offer added features such as the ability to track how many people have clicked on your link or an open programming interface (API) so developers can build the service into the applications they write.
One problem for all these services is dealing with spammers, for whom hiding a link’s ultimate destination is useful for bypassing spam filters and trapping the unwary into visiting their sites. Letting users preview sites is one antidote; another is acting quickly to disable bad links when they’re reported.
Another problem about increasing traffic, however, is: how does a free service scale? Andrew Chapman, one of the four people behind UK-based qurl.com, is not convinced it’s possible to turn that traffic into money.
“This kind of service, people expect to be free,” he says. Qurl shortens links to a five-character string and also allows users to pick meaningful words. He says: “I thought originally I would monetise that as a premium service, but I never got around to it.”
TinyURL does have Google Adsense on its home page, but much more than that, says Chapman, “would just annoy people, and then they’d go to another service. It’s part of web culture that lots of things are free to use.”
In mid-July, Twitter bought Summize, an independent search engine, just for Twitter. Can the acquisition and consolidation of URL shortening services be far behind? And nearly simultaneously, Summize’s early investor, New York-based Betaworks, launched its own new URL shortener, bit.ly.
Betaworks’ chief executive John Borthwick says that while it’s not expensive to build a URL redirection service, it is expensive to “build one that scales, grows, and keeps billions of URLs”. Until now, he adds: “This market has been made up of developers who’ve needed something to solve this problem so they put out a solution.” Betaworks built bit.ly because, he says, “nothing was quite open enough or scalable enough”.
Betaworks had expected to launch bit.ly quietly and add to its functionality over the next few weeks to make it clearly different. Instead, people leapt on it.
Thinking of linking
Where TinyURL uses a single server to generate shortened URLs and an algorithm that generates each new one based on the existing count, bit.ly uses a distributed index and multiple servers: it picks the next URL randomly and then synchronises the index tables across the servers. The chances of collisions are remote now because the service is so new, but “we’re thinking hard about how to deal with it when we have billions,” says Borthwick.
An important distinguishing feature: bit.ly keeps copies of the destination pages, eliminating one of the key problems with the older services: “link rot”.
Bit.ly is also running the saved pages through Reuters Open Calais API, which, says Kirkpatrick, lets it “analyse content and semantic terms”. Identifying the concepts and subjects of interest and matching those with geographical data parsed through another API means that “if bit.ly ever got to TinyURL status it will have a pretty rocking database of machine-readable web pages”.
What Kirkpatrick is talking about is taking search to a new level: a human-selected collection of pages that matter to people. And you don’t have to rely on active tagging or pay for all that relatively expensive indexing.
Kirkpatrick concludes: “I always assumed that’s how somebody was going to make money through one of these services because they’re seeing all this traffic come through – links that are obviously important to people.”