Earlier this afternoon I read an interview with Loren Brichter, the man behind Tweetie and Scribbles, on The Setup. In the interview, Mr. Brichter notes that, while the Mac OS X flavor of his Twitter client has advertising unless users pay $19.95 to register the application, many of the people who opt to register their copy of Tweetie don't bother to turn off the advertising that in effect "pays" for the free version. At first, I this seemed very strange to me. I usually despise advertising. (Long live the TiVo!)

But as I thought about it, I realized that not only do I not mind the advertisements injected into my Twitter stream in Tweetie, I actually like having those ads in my Twitter stream. For unlike most other advertising, the Fusion ads in my Twitter stream are informative; I've learned about a variety of useful software products and web services, such as WuFoo. And when I've seen a particular advertisement before, or I find the good or service it promotes irrelevant to me, the injected ads don't blink, honk or otherwise try to take my attention of off my Twitter feed.

Which leads me to propose these two criteria for determining the goodness of an advertisement:

(1) A "good" advertisement conveys novel information
(2) A "good" advertisement imposes minimal costs to the consumer

Note that a given advertisement can meet and fail the first criterion at the same time, as whether an advertisement conveys novel information (or not) depends on who consumes the advertisement.

For instance, a banner ad for the Kindle on The New York Times' website might satisfy (1) for a 40-something doctor who spends her weekends kayaking, but who can't tell the difference between a DIMM and SIM. The same ad, however, might fail (1) if it's consumed by a 30-something advertising executive who paid $399 for the original Kindle the day it went on sale, and who has since upgraded to Kindle 2. In the first case, the banner ad conveys something novel: the doctor had no idea that Amazon.com sold an electronic reading device with over-the-air purchasing functionality. But in the second case, the ad tells the advertising executive nothing he didn't already know.

It's also interesting how the second criteria manages to capture some of the differences between print and electronic media. Advertisements in print magazines generally bug me far more than advertisements online, partly because ads in physical printed magazines add additional mass. I've always wondered why people subscribe to magazines like Vogue, even if the content is interesting, just because every issues has so many advertisements that its weight is often measured in pounds rater than ounces.

If I were more adept at constructing economic models, I could easily see this developing into a nice little paper with a title like "Optimal Advertising Theory."