Google-sponsored research helps ad-slingers find fanbois on social networks

Re: Watch and bag

I read it as a.i. learning where and when to advertise, rather than specific individuals per se.

Re: Impulse buyers?

Well, that's a bit of a narrow assessment. There may be various reasons for responding positively to an advert:

* It's for something you never knew even existed;

* It's for something you never realised you needed/wanted until the advert convinced you as much;

* It's for something that you instantly think looks desirable and will elevate your social status;

* It presents a value-proposition such as an unprecedentedly-high discount on an item you wanted but previously considered too expensive.

As for "ignoring" the adverts, may I submit that I bet you don't, not really or at least not all of them. Much as I dislike everything they stand for, every so often one does catch your attention, whether it then successfully converts that (through one of the above qualifications) into a desire to buy or not.

All that said, I still believe the business of targeting adverts is pitifully primitive, and I can't see this latest wrinkle making much difference. I mean, if the wonks at Lacoste *need* the help of these clever algorithms to know that most of their apparel's most ardent fans are in the [image/context metadata] "PURE BANGIN MENTAL CASTLEMILK PO$$E FAK YUZE", then (a) they must be terribly sheltered and (b) what exactly are they going to do about it anyway?

I see a lot of scope for positive feedback loops in this system too. How do they know it's not the latest viral of V-Hudge falling out of [or in] Nobu, or some teehee macro featuring Michelangelo's David doing the rounds, that they've just hoovered up 187,000 times because it featured a Prada bag? If one had a mind to, I suspect you could game it remorselessly.