Thursday, September 24, 2009

Nambu Gets Better and Shortener User Tracking is Undermined


When I last wrote about the tribulations of tr.im and the business of bit.ly, our heroes had just stepped away from their nose-to-nose struggle, with Nambu founder Eric Woodward having announced the shut-down of tr.im, his URL shortener, only to vow its revival a few days later. Bit.ly, with a cozy relationship with Twitter, seemed to have taken a dominant position in the URL shortening business, whatever that turned out to be. I speculated that Bit.ly would use its position in the ocean of usage data to build psychographic profiles of users to help target advertising.

Since then, Woodward decided not to sell the tr.im business and has instead released the tr.im software as free open source, making it that much easier for websites to do their own URL shortening. He's also focused his company's attention on its Nambu Twitter clients (Mac and iPhone), the development of which suffered a major setback when his Chinese developers left for richer opportunities as soon as their contract was up. Nambu for Mac OS X has been my preferred Twitter client; it has a much more Mac-like user interface than others I've tried. When I updated my system to Snow Leopard last week, I was disappointed to find that Nambu had not survived the system update.

After unhappily revisiting Tweetdeck, I decided to try the beta version of Nambu, even though it's described as being not quite done. So far, it looks pretty solid. One change in particular pleased me, and that's the way the new Nambu works with URL shorteners. It seems that by surrendering URL shortening to bit.ly, Nambu is now freer to innovate in the user experience. Nambu now pre-expands all the shortened links so that the user can see the hostname that the links are pointing to. This has a number of consequences:
  1. The user can tell where a link will go. This will help avoid wasted clicks, and will help the user avoid spam and malware sites.
  2. Because all of the links are dereferenced before use, the URL shortening sites will no longer be able to track the user's reading preference. The business model I previously suggested for bit.ly will be defeated, and the user's reading privacy will be protected.
  3. The URL shortener will have to deal with an increased load. Nambu's going to make bit.ly work harder for the privilege of domination the URL shortening space.
Now I understand why bit.ly has been registering a bunch of instantaneous hits whenever I tweeted a link- it was robot agents, not people, that were clicking the links.

I was curious to see if Nambu was querying the URL shorteners directly or whether Nambu was trying to aggregate and cache the expanded links. I installed a nifty program called "Little Snitch" to see the outbound connections being made by programs on my laptop. It turns out that Nambu is doing a direct check for redirection on ALL of the links that it shows me, not just the shortened ones. Although this could break links that are routed as part of a redirect chain, I imagine that sort of link occurrs rarely in a Twitter stream.

The new behavior of Nambu and its effects on usage tracking points up a general problem faced by any system designed to measure and track internet usage. In my post on "bowerbird privacy", I mentioned that I use StatCounter to measure usage on this blog. StatCounter works quite well for now, but I imagine that its methods (based on javascript) might well stop working so well as web client technology evolves. That's one reason I expect that efforts to standardize measurements of usage in the publishing community, such as Projects "COUNTER" and "USAGE FACTOR" are doomed to rapid obsolescence.

Will bit.ly ever get a business model? Will Nambu find peace with the chilly kitty? Find out in next months installment of... As th URL Trns

0 comments:

Contribute a Comment

Note: Only a member of this blog may post a comment.