Twitter Quitters: API Sucking Away Traffic

by Eric H. Doss on 11 May 2009

A few weeks ago, Mashable.com ran an interesting story about returning Twitter users.  The original story reported that only 60% of Twitter users returned during the same month they first registered an account.  More interesting was an update added to the story later that noted that Nielson, the ratings and stats company and author of the report, can only track people actually visiting Twitter.com.  They cannot track people that use Twhirl, TweetDeck or other desktop clients.

I find this really interesting…  Advertising is the lifeblood of the internet.  Advertising generates revenue in a few ways: you can sell clicks or you can sell impressions.  If you sell clicks, you’re paid when a visitor clicks on the ad on your site.  If you sell impressions, then you are paid a flat rate for an ad based on the number of visitors to the site.  This is measured in CPM, or Cost Per Thousand Views.

One of two things are happening with Twitter.com.  Either the site is truly losing 40% of their new users every month, or people are migrating to desktop clients and Twitter has no way to track users.  Twitter desktop clients are so ubiquitous because Twitter has a very open “firehose API” that allows developers to access all user generated data for all members.

The big question is how do you monetize this traffic?  Though I’m not a programmer, my first response is to insert the advertising through the API.  I see two immediate problems: First, you might have to attach the advertising directly to the user’s update, which will be a bit processor intensive.  Next, the real question is will developers want to use an API that sends uncontrolled advertising to their products?  Would Twitter have to provide controls for advertising to developers?  Would they simply block access to programs that strip out the advertising?  Share revenue with developers?

Anyone wanna weigh in here with a few ideas?

Reblog this post [with Zemanta]

Leave a Comment

Previous post:

Next post: