The Perils of Short Links

Earlier this week the people behind a Web service called Tr.im announced that they were going to shut the service down. Then, a few days later, they announced that they were going to put the service back online and keep it running "indefinitely." It was a minor blip on the Web for most people, but it teaches several lessons about the Web. I'll get to those in a minute.

First, some background on Tr.im. The service is a Web link shortener. That is, it takes long links, and converts them into short links, like "http://tr.im/wmGS." This class of service (it has many competitors) arose a while ago, but really took off in the last year as Twitter became a popular way for people to share Web links. Twitter's 140-character limit on posts means that without link shorteners, it'd be nearly impossible to share links and have any room left for cogent commentary.

It's a clever solution to a Twitter-imposed problem. But it's a tenuous business, for a few reasons, and made even more so by Tr.im's own actions. It's illustrative to examine this.

Lesson one: Do not rely on the kindess of strangers

As the official Tr.im blog has stated more than once, the Tr.im product, while it serves the growing audience of Twitter users, also exists at the mercy of Twitter itself, and Twitter does not favor Tr.im. In fact, Twitter itself uses the competing URL shortener Bit.ly for shortening links. If you put a long link into a post on Twitter.com, it will appear as a Bit.ly link once you enter it.

If your product's success relies on the ongoing blessing of one company, and you have many competitors that do much the same thing, you might want to consider an expansion plan, some way to lock in a contract with the company in question, or -- just a thought -- leaving the business, since you really don't have any way to ensure ongoing success.

(Previuosly, Twitter had used the original famous URL shortener, TinyURL. So this lesson applies to Bit.ly, too.)

Lesson two: If you fall down, stay down

If you're going to stand between users and their destinations, as all link shortening services do, there's one thing you must be above all else: Reliable. Tr.im is not. I could not use the for about 30 minutes when working on this story. I recognize that inadvertent outages are unavoidable, but shutting down the service, as Tr.im said it was going to do, and then bringing it back two days later due an outpouring of public support, will eventually be seen as what killed this product. With so many competing link shorteners, there's just no reason to use one whose founders are capricious enough to announce they're taking it down and then say they're not, as they see fit.

As in other businesses, on the Net, it's a game of guts. You need them. Your users, and your competitors, can smell fear. Tr.im blinked, and it's not coming back from that.

Lesson three: Have a backup

This is just the right thing to do: If you make a mission-critical product, or one that, if it fails, will inconvenience a lot people, have a continuity plan that will survive your business failing. It will serve you in the here and now, because it makes people more likely to trust you and thus do business with you.'

There are emerging efforts to give companies like Tr.im a place to archive their databases, so that even if the companies fail, the data will be recoverable by others, like the people who are relying on clicks through the Tr.im services.

Lesson four: A great product doesn't mean a great business

Like Twitter, a great service that has yet to make money, the short link services are great products that don't yet generate sizable revenues. Chances are that one or two might -- Bit.ly, for example -- but the rest of the URL shorteners are rather doomed as businesses, even though people may like them a whole lot.

That's not to say that you need absolutely need a strong revenue model from day one. Twitter, Google, and YouTube are proof that founders of tech companies can figure out revenue, or just sell out, after they get their product right. (I'm making an assumption on Twitter here, but I'm confident about it.) But for every Google, there are dozens of Tr.ims. Moving from free to paid is a rough transition, especially if you fail your users' trust. Call me crazy, but I generally think it's best to start making money when people start using your product.

By Rafe Needleman