Comments on: On Facebook, racial profiling and the mob http://responsiblemarketing.com/blog/2010/03/03/on-facebook-racial-profiling-and-the-mob?utm_source=rss&utm_medium=rss&utm_campaign=on-facebook-racial-profiling-and-the-mob where commerce and conscience come together Fri, 18 May 2012 23:03:00 +0000 hourly 1 http://wordpress.org/?v=3.3.2 By: Rob Wolf http://responsiblemarketing.com/blog/2010/03/03/on-facebook-racial-profiling-and-the-mob/comment-page-1#comment-6434 Rob Wolf Thu, 04 Mar 2010 07:32:35 +0000 http://responsiblemarketing.com/blog/?p=1805#comment-6434 While I agree that seeing some of those juxtapositions of "you are a fan of X, become a fan of Y" can be a little jarring, they're usually done by the advertiser, not Facebook. A real world example: I'm running a campaign right now in which I was able to very specifically choose my targeting criteria. My product is a DVD and CD series made for Jewish babies. While I can't target Jewish moms explicitly, I can get an awfully close proxy by doing the following: Targeting married women age 30-44 with college degrees who FRIENDS of the current fans of my product. I could have also targeted fans (or the friends of fans) of similar products. (Ethnic marketing tip: all Jews know each other. We were social networking since long before there was a name for it!) So, back to Michael Jackson. Maybe the geniuses running his fan page decided that some combination of demographics plus being a fan of Barack Obama made you a likely fan of MJ. I don't think there's anything nefarious going on. Just a new form of targeting that can yield some occasionally head-scratching results. While I agree that seeing some of those juxtapositions of “you are a fan of X, become a fan of Y” can be a little jarring, they’re usually done by the advertiser, not Facebook.

A real world example:
I’m running a campaign right now in which I was able to very specifically choose my targeting criteria. My product is a DVD and CD series made for Jewish babies. While I can’t target Jewish moms explicitly, I can get an awfully close proxy by doing the following: Targeting married women age 30-44 with college degrees who FRIENDS of the current fans of my product. I could have also targeted fans (or the friends of fans) of similar products. (Ethnic marketing tip: all Jews know each other. We were social networking since long before there was a name for it!)

So, back to Michael Jackson. Maybe the geniuses running his fan page decided that some combination of demographics plus being a fan of Barack Obama made you a likely fan of MJ.

I don’t think there’s anything nefarious going on. Just a new form of targeting that can yield some occasionally head-scratching results.

]]> By: Patrick Byers http://responsiblemarketing.com/blog/2010/03/03/on-facebook-racial-profiling-and-the-mob/comment-page-1#comment-6433 Patrick Byers Thu, 04 Mar 2010 07:07:44 +0000 http://responsiblemarketing.com/blog/?p=1805#comment-6433 Good stuff, Jeremy. Thanks for sharing. Good stuff, Jeremy. Thanks for sharing.

]]> By: Jeremy Tanner http://responsiblemarketing.com/blog/2010/03/03/on-facebook-racial-profiling-and-the-mob/comment-page-1#comment-6432 Jeremy Tanner Thu, 04 Mar 2010 07:01:51 +0000 http://responsiblemarketing.com/blog/?p=1805#comment-6432 There's a much simpler explanation. Facebook happens to be sitting on one of the worlds largest datasets of it's kind. Data on Obamas 8 million fans says, "An Obama fan is more likely than non-fans to also be a fan of Michael Jackson, or at least become a fan when presented the option" That's not racial profiling, racial profiling is "People who ARE [whatever race] will like [thing or person]" Collaborative filtering is "People who LIKE [thing or person] will also like [thing or person]" Facebook's algorithm isn't operating off of racial stereotypes, but from the statistically significant decisions and likes gleaned from half a billion users. Wikipedia's page on Recommender systems goes into further detail. Ok, maybe not a simpler explanation, but a much more likely one. Jeremy There’s a much simpler explanation. Facebook happens to be sitting on one of the worlds largest datasets of it’s kind. Data on Obamas 8 million fans says, “An Obama fan is more likely than non-fans to also be a fan of Michael Jackson, or at least become a fan when presented the option” That’s not racial profiling, racial profiling is “People who ARE [whatever race] will like [thing or person]” Collaborative filtering is “People who LIKE [thing or person] will also like [thing or person]” Facebook’s algorithm isn’t operating off of racial stereotypes, but from the statistically significant decisions and likes gleaned from half a billion users. Wikipedia’s page on Recommender systems goes into further detail. Ok, maybe not a simpler explanation, but a much more likely one.

Jeremy

]]>