Personalization and the web

by Eric Norlin on Jun.05, 2007, under defrag "theory"

Way back in the early mists of internet time (you know, like 1999), there was a company called Net Perceptions. They were the guys that came up with the algorithm for “if you like this, you might also like,” and it started a whole movement around “personalization.” Ironically, the first conference that I was ever involved with was the “Personalization.com Conference.” Net Perceptions was the underwriter.

So now, when I read that Yahoo! is saying that “the future of the web is personalization” — well, frankly, I feel like I’m in a bit of a time warp.

And there’s also just *something* about the term that doesn’t feel quite right. Now, don’t get me wrong, “personalization” (ie, filtering the onslaught of information for the individual) is a great and absolutely needed thing, and I also think I get where the Yahoo! guy is going with this (that you won’t “search” for information, but rather the web will be “personalized” to deliver you what is appropriate). It really all relates to the “passive discovery” that I’ve spoken about in this space before.

Still, is it me, or does “personalization” feel a tad too narcissitic? I mean, the web is a *network* of *connections*, right? And maybe, just maybe, it shouldn’t be all about me, myself and I.

Am I wrong? One too many cups of coffee this morning?

4 comments for this entry:

P-Air
June 5th, 2007 on 7:19 am
The personalization that Net Perceptions referred to was collaborative filtering, which I believe is diff than what Yahoo! is up to. Actually, Amazon was an early customer of Net Perceptions though I believe they went to build their own platform after using this one for a while.

More analogous offerings to Net Perceptions today are those offered in the form of platforms by companies like Aggregate Knowledge, Loomia, Criteo, and Baynote to a certain extent. These aren’t as much about personalization and much more about cluster analysis (wisdom of crowds so to speak).

I’m not sold on the general idea of personalization, given that most of the offerings I’ve seen neglect the user’s context (who am I right now?…dad, kitesurfer, account executive…?). Heck, in households where more than one person shares a computer, personalization algorithms should have some fun

My biggest concern w/personalization however is that we do not yet have the necessary laws and protections against the use of this information for purposes outside of our control. The exhaust of our usage is becoming quite valuable and in some cases being abused (IMHO). The question I see frequently pondered is whether we’d be willing to give up some privacy in order to gain a better online experience. Unfortunately, too many people are simply answering yes to this w/o understanding its ramifications.
Eric Norlin
June 5th, 2007 on 7:51 am
P-Air: agree that the collaborative filtering of Net Perceptions is significantly different that the “personalization” going on today - but I still think they’re related….and I think the topic of “personalizing the web” is a big one.
Mark
June 5th, 2007 on 10:25 am
Eric: I think this time it’s different.

The way I like to think about it is not so much “personalized” because I agree, that is too narcissistic and limiting given what the web has to offer. I’d rather think about it as attraction. I want stuff that I’m interested in (and maybe interested in…think “slight mutations in my genetic interest profile”) to be attracted to me. I like Dave Matthews…Google knows, Apple knows, so why shouldn’t I get an offer for two tickets to see Dave when he’s near by without having to search for it. Obviously there are all kinds of issues regarding privacy and content sharing but that’s why you’re having the conference, Right?(looking forward to it.)
Chip Griffin
June 6th, 2007 on 9:24 am
I actually like the term “personalization.” I think it aptly describes what the benefit to the user is. It may not describe the process of how you get there, but I think it is incumbent upon us to simplify language and communicate with users in the terms they can identify with. Ultimately, it’s all about user benefits, isn’t it?