If You Liked This Post, Perhaps You'd Like To Look At The History Of Failed Recommendation Systems
from the we-recommend-you-read-this-article dept
For years, we've been hearing about the holy grail of the perfect recommendation system -- and yet, every time these systems fail to live up to the hype. Early on in the web days, there was FireFly and its amazingly overhyped collaborative filtering system that was supposed to revolutionize the space. Others have come and gone, and while the systems have improved marginally, they're still far from useful on a regular basis. They're either exceptionally limited, or they tend to provide mostly obvious solutions. That is, of course, if they can provide decent recommendations at all. There was the famous article a few years ago in the Wall Street Journal about how people were noticing when their TiVo's recommendation system thought they were gay after they recorded a "gay" show. It's gotten so bad lately that Netflix has resorted to throwing up its arms and asking programmers everywhere for help.That's why it's a little surprising to see Fortune Magazine claim that these types of personal recommendation systems are somehow the next big thing. It's true that they've always been something of a holy grail for marketing types, who would love to be able to provide more accurate and relevant choices to you at all times. However, we're still a long way away from that. While some of the hype over personalization in the ad space got swept aside when Google figured out how to do better keyword matching and contextual advertising, many now believe that this kind of "discovery" advertising -- recommendations based on what you're like, rather than what you like or what you just searched for -- is the future. The problem is that it's not an easy problem to solve, and the long history that's littered with failures in the space would suggest that just claiming you've got the best solution (or performing simple parlor tricks, as one group does in the article) isn't actually enough to signal a revolution.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Reader Comments
Subscribe: RSS
View by: Time | Thread
I loved this article!
all kidding aside, doesn't it boil down to constantly changing option sets? as long as variables keep changing, it's hard to predict how groups and subgroups and individuals will respond to new choices as they become available. a truly difficult problem!
[ link to this | view in chronology ]
Amazon
[ link to this | view in chronology ]
Re: Amazon
[ link to this | view in chronology ]
[ link to this | view in chronology ]
I really don't want any recommendations
[ link to this | view in chronology ]
Wall street only has one goal - to make you buy and sell and repeat that cycle often. Therefore, they have no credibilty.
Any profits or losses you experience by following their recommendations are of no interest to them.
Ooooh - it must be the next big thing - Wall Street says so. NOT!
[ link to this | view in chronology ]
failed recommedation systems
TiVo and TV on Demand are going to change all this just as soon as they figure out how to get cameras into their equipment (that digital box over top of your tv) and sell us on the idea that being constantly watched is a good thing.
Allegedly, they're going to watch (and guage) our reactions to commercials and then decide what commercials to send, based on our reactions. Did we get up and walk away from that truck commercial? Are we watching, in rapt attention, to that sexy lingerie commercial?
And what are our children doing?
What does it take to keep our attention? (And whose attention do they want to keep?)
You don't need a person to decide this, the technology is available for computers to do this.
Conspiracy theory or just a matter of time?
[ link to this | view in chronology ]
Recommenderz
As computers get faster and there are more datapoints to collect these systems will become better.
[ link to this | view in chronology ]
False Negatives, False Positives
Why not? Because I hate Mel Gibson, because he's a racist, sexist, homphobic bigot and holocaust denier.
How is a recommendation system supposed to figure that out? A vector in n-dimensional space with a dimension just for Mel?
[ link to this | view in chronology ]
Last.fm logs the music you are listening to and uses this to build a personal history/profile. This is then compared to thousands of other profiles to see which people have the most similar listening habits to your own (your 'neighbours'). Music that these people like that you have not yet listened to is then recommended to you (and it works very well!).
[ link to this | view in chronology ]
brand new nokia n95i just................$500usd
get back to us contact:moremobileinc1@yahoo.com
larry_dish@yahoo.com
lanmobilestoreinc1@hotmail.com
joymobilestore@inmail24.com
address:12/F, 248 Queen's Road East,
Wanchai,
Hong Kong.
[ link to this | view in chronology ]