08.17.12
How Facebook knows what you really ‘like’
It’s vaunted advertising algorithms scan your profile and status updates for simple keywords. If you post something like ‘alcohol’ or ‘drink,’ you get beer ads linked to the ‘likes’ of your ‘friends’, then offers for rehab. If you live in southern California, you get anything local. If you talk about playing guitar, you get ads and ‘giveaways’ run by every music store in the country. If you play guitar really well and post evidence on your wall, Facebook will give you links for lessons from Esteban or some local jazz blues and cocktail bar-playing dad. If someone in your ‘friend’ list is from Canada, you will get a pitch to vacation in Alberta.
And if you are in your fifties you get blandishments for treating Alzheimer’s before it strikes.
It’s not at all like this bullshit written by Cade Metz and printed in Wired:
Facebook is different from Apple or Google or Amazon or Microsoft, says Mark Zuckerberg, because it doesn’t build products. It seeks to improve the products built by everyone else …
But Zuckerberg wants more. As the Facebook Platform enters its sixth year, the company is expanding its mission through something it calls the Open Graph. This isn’t a visual graph. It’s not a line graph or a bar graph or a pie graph. In this case, graph is a mathematical term. It’s a way of representing connections between pieces of data …
But Open Graph is more than just way of moving song names from one place to another. It’s at least a small step towards what has long been called the semantic web — a web where information is structured in a way that it be more easily analyzed and refined and reused by outside services. Facebook’s more than 900 million users generate so much data on the social network — and beyond — the company can’t just shuttle all this information into your Newsfeed. Open Graph provides application and website developers with a way of structuring their data, so that Facebook machines can readily use it and restructure it and reuse it as need be.
“We could have just done text analysis,??? says Vernal. “But we decided that if we could create a framework where developers can tell us the structure of this information, we could build much more interesting and much more compelling visualizations of this data both in Newsfeed and on Timeline.???
In short, Facebook is striving to organize and use data generated by other companies in much the same way it has always organized and used data on its own site. Facebook beat out the likes of MySpace because its data was structured in a way that gave it some context …
If you’re in a tech firm you can tell a Wired reporter you piss iced tea and they’ll publish it without blinking.
Then there’s the real world where companies are getting the idea that Facebook’s advertising engine isn’t that great, that like Google Ads, it’s more suited to Internet bottom-feeders and chumming for suckers.
Today, on my Facebook wall:
If you ridiculed Chik-fil-A or ‘liked’ a straight-to-video movie called “Ticked Off Trannies with Knives” as whimsy you will certainly want to go to a bar in WeHo this weekend. And the ‘semantic web’ of your language also tells us that you might wanna run criminal background checks on your ‘friends’ and family.
311 people ‘like’ this. Facebook advertising value: Effin’ priceless.
And if they keep going great guns there’s a “free credit report” in your future, I hear.
Throw your own little share of sand in the gears. Either unlike all the stupid shit that points to businesses, movies, tv, books and bands that your ‘friends’ aren’t really interested in, anyway. Or “like” a lot things you know to be total crap or which you actively hate. Or only “like” stuff that is defunct and/or decades old. All these strategies work, reducing your value while pumping corrupt or un-useful data into the system.
As part of the proposed settlement, Facebook had agreed to give users more control over whether they became unpaid endorsers in ads aimed at their Facebook friends. Facebook also agreed to pay $10 million in legal fees and $10 million to nonprofit organizations specializing in privacy including the Electronic Frontier Foundation.
Facebook has said that Sponsored Stories is one of its most effective forms of advertising on computers and mobile devices. The ads feature the name and photo of a Facebook friend who has clicked that he or she “likes??? the advertiser.
It has begun testing expanding the ad feature to let marketers drop more messages into users’ News Feeds. Those messages would target users on their mobile devices and computers even if they have not “liked??? the advertiser.
So, if for example you follow DD blog advice and ‘like’ Dos Equis beer, even though you think it’s urine, Facebook will want to use your name on a ‘friends’ feed to sell it to them.
Also, from the Times, on poor, poor, pitiful Facebook employees and their dwindling value:
For many staffers, the precipitous drop means their Facebook stock is not going to yield the returns they hoped, at least not right away. They have had to defer or downsize their dreams of buying a home or a new car.
“People made life plans and calculations,” said a Silicon Valley chief executive who spoke on the condition of anonymity to preserve his relationships inside Facebook. “This is very, very painful.”
And the noisy public criticism of Facebook has become nearly impossible to shrug off, hurting employee morale.
“These are people who like to win. Now there’s this external measure of winning which is difficult to ignore,” said one former Facebook employee who also requested anonymity to preserve relationships at the company. “It doesn’t feel good.”