These facts are nothing new and, I’ll bet, do not allude those that built these services. But, they see a general use and purpose for these services and decided to provide their own solution to the problem.
The problem is that, in some cases, you need a shorter URL than the one provided by a particular Web site. Web sites with incredibly long URLs (like Amazon, Google Maps, or search results on a site) can be cumbersome to deal with in situations like writing email, Twittering (I’m cdevroe by the way), and sending SMS messages. URL shorteners attempt to solve this problem by creating links to these pages much easier by providing a significantly shorter URL that simply redirects to the URL that you chose.
Seems innocent enough. Seems simple enough. However, by creating a shorter URL that represents a longer one you’re, as Joshua states, adding unneeded layers that could potentially fail overtime. If the URL shortening service manages 1,000,000 redirects, and suddenly goes down, those redirects no longer work. This is a big problem.
For services like Twitter, which benefit greatly from these URL shortening services due to their short message limit, they stand to have millions and millions of dead links. Right now, by default, Twitter uses TinyURL to automatically shorten URLs to help them fit into the 140 character limit for SMS messages. Jason Kottke suggested that Twitter create its own URL shortening service so that they can guarantee it be around forever and to replace all of the short URLs it had created in the past. I’m going to go one step further and suggest that they buy Bit.ly.
While Twitter has chosen to use TinyURL I believe this was because Bit.ly wasn’t around when they added the TinyURL functionality. Bit.ly is more on par with Twitter’s real-time efforts. Twitter would immediately get their own URL shortening service that has, on top of it, a very good statistics package to show how those links are being used, where they are clicked on from, how many people clicked them, and a service that has a good API.