Re:Is Facebook a publisher? In public it says no, but in cou...
by EpicEng on Thursday July 05, 2018 @12:14PM
(#17464396)
This is fairly well settled science. If you have conflicting data then by all means, share, but you can't just write off decades of _real_ work in the field based upon your own musings and expect to be taken seriously.
[ Parent
| on HN ]
Re:Is Facebook a publisher? In public it says no, but in cou...
by hackinthebochs on Thursday July 05, 2018 @02:23PM
(#17465535)
I'm not sure what decades of research you're referring to, but after my own search the literature seems suspiciously sparse on the effects of endorser to perceived quality of the product. I agree that celebrity endorsements are effective generally, just not for this reason. I would be interested in seeing some of these studies you're referring to.
Re:Is Facebook a publisher? In public it says no, but in cou...
by mozumder on Thursday July 05, 2018 @02:40PM
(#17465693)
So, when advertising agencies run A/B tests, they don't publish the result in scientific journals. They just use the data internally along with art direction and relationship/political factors to make advertising decisions.
This isn't an open source field with a handy guide. This is a 'knowledge through experience' field. You learn from those people, not through academic research or Google search.
In fact, a good advertiser can instantly find holes in published research based on the many variables that a paper doesn't take into account, but an advertiser is already familiar with. "Brand value correlates with celebrity status... but inversely with the celebrity nationality.. with a correlation with the accent.. but also with season of the year.. except when using this font... and based on president's party" etc.. All these variables may have a nonlinear relationship.
To get a good introduction to what you're dealing with, just watch "Mad Men".