The Net ad you're about to see has read your e-mail

Google's new sponsored e-mail service turns out to have some interesting self-imposed constraints.
The New York Times

The story "The Net ad you're about to see has read your e-mail" published June 20, 2004 at 9:20 AM is no longer available on CNET News.

Content from The New York Times expires after 7 days.

[hadaso]

What's the problem?

Gmail only displays ads to a user that opted to see them. No one is forcing anyone to use Gmail. No one is reading the mail. The only thing that happens is mathematical analysis of the message that uses past statistics collected to display ads. No one but the user seeing the screen reads the email. If the ad is clicked, the only thing that the advertiser would know is that it is refered to from Google(Gmail).

And about relevancy of the ads: you cannot expect a service in beata with perhaps a few tens of thousands of users active for a couple of months to be perfect. Especially not when the method used is a self learning algorithm. It will take a few million users and several months of use for the matching algorithms to self adjust.

Too much attention is given to the "privacy issue" related to Gmail. In my opinion it's a non-issue. It's just a matter of some mathematical formulas applied to the contents to produce some "related" links. As I see it the only information Google will have in the end is a collection of parameters that whan applied to a message maximize the probability that a link would be followed. Not something that can be "reverse-engineered" to reveal actual facts on the reader or sender (especially not the sender, as they would be aggregated by recipient, if at all).

There are far more interesting aspects to Gmail that are hardly discussed, such as how it would impact the rest of the industry: will other webmail provider or email client producers switch to the "conversation+search" paradigm? It seems to me that it can be accomplished using the existing IMAP specification or with slight revisions to IMAP, by trading the use of IMAP "mailboxes" as folders in favor of using IMAP flags as labels, and using IMAP's built in search. Of course this would probably make demand for better support of search in IMAP servers, and different IMAP clients.

[TV James]

This is still a non-story

Slow news day?