Monday 15 September 2014

Facebook wants to know why you hate specific adverts

Facebook has announced yet more slants to its constant ad fiddling.
Namely, it says it's going to do two things to try to avoid users clicking that dreaded "hide" option on the ads in our newsfeeds.
First: after we tell Facebook that we don't want to see an ad, it will ask us to explain ourselves.
"Help us understand the problem," it will ask in its best Freudian accent as we stretch out on the couch and it grabs a notepad.
The choices for why we don't want to see ads will include:
  • It's not relevant to me
  • I keep seeing this
  • It's offensive or inappropriate
  • It's spam
  • Something else
Facebook ads
When testing the feature, Facebook says that it used the results of this particular update to stop showing bad ads that were offensive or inappropriate and thus saw a "significant decrease" in the number of ads people reported as offensive or inappropriate.
That's nicer for all, Facebook said:
This means we were able to take signals from a small number of people on a small number of particularly bad ads to improve the ads everyone sees on Facebook.
The second ad move Facebook's making is to pay more attention to feedback from quiet people: i.e., those who don’t often hide ads.
Because they don't hide ads much, when they do hide them, those ads must be ultra bad, the thinking goes.
If someone hides things very rarely, we’ll consider that when we choose what to show them. If we think there is even a small chance they might hide an ad, we won’t show it to them. This affects the type of ads we show everyone, but has a bigger impact for people who don’t often hide ads.
When Facebook tested the "listen closely to quiet people" update, it saw that people who rarely hide ads ended up hiding 30 percent fewer ads with the change.
Facebook took that to mean that by listening carefully, it can show better ads - even to those who aren’t very vocal.
But who knows? Maybe the quiet ones have slipped further still into advertising-induced comas.

No comments:

Post a Comment