More in the Email Buzzosphere

Email has reentered the blogosphere in a big way over the last 6 months or so. There are so many posts now that it’s impossible to keep track of them all. Dave McClure posted another entry in his recent series on email a few days ago.

In his wild-man style, he covers everything from open payment systems to friend lists and by the end of the article seems to have transformed decade old eMail systems into brand new billion dollar business franchises that will dominate eCommerce and displace the social networks.

So he can be dramatic. So what.

The main point he is making at the bottom of it all is that the social information implicit in email is being badly under leveraged. He gets into a little more detail on it over here.

There are a few interesting ideas in there, and I think that for many people email is a good source of social information. His vision is that email patterns will be analyzed and used to come up with a list of contacts who should become the short list for new social sharing.

PROBLEM: when i’m on a website or social network or social app, it wants me to help refer other people / help with some lame-o viral marketing scheme. however, i’m only gonna do that for a select few who really share my context / insanity, not everyone in my network (unless my name is Robert Scoble).

what I DO NOT WANT:

• select from a god-awful list of 100 faces in your pop-up face listbox
• upload my entire address book of 2000+ contacts for you to spam the world
• wait for your data-loading / selection function to crash horribly, after taking forever to load

what I DO WANT:

• popup the MOST RELEVANT 5-10 peeps who meet certain key criteria
• use an intelligent combination of shared interests & messaging frequency to figure out who these “TOP” friends are (for the given context)

let me select 1-3 of them to invite & checkout an awesome [video game | baby stroller | new book | really good pr0n] i just found

Sounds cool. I think the problem with the thinking is that people communicate for vastly different reasons and the algos needed to guess at which contacts might be relevant for which purpose are way-hard to write. We tried to use prior behavior to guess at what products might be relevant for users at Shopping.com. If our experiences with behavioral relevance at shopping.com are any indication, the best friend-guessing in the world will only get you 80% of the way there. Thats not good enough to automate the job, and it’s not like its the most painful thing in the world to check off a few boxes each time you want to do some medium-scale sharing.

Although nice to have, I doubt if they will be worth in revenue terms the investment they would take to write well…