I'd love to have some discussion about cutting ads when split testing age ranges.
In Part 8 of the tutorial Caurmen talks about treating each age range as it's own campaign. So in other words if everything else is the same except age you treat them as independent of each other.
The problem I'm running into is an image that has fantastic CTR for, lets say 35-39, is performing poorly in 40-44.
As I look between my split tests I have a hard time cutting an image. It may not do well in 2 of 4 age groups but does well in the other 2.
Is everyone really treating these totally separate or are you keeping images based on performance in different age ranges?
Any input is appreciated.
Thanks,
Evan
split age groups into different campaigns, cut the banners that aren't performing for a certain age group only, or even cut that age group altogether if it's not performing
Yeah cut images in age brackets independently - who cares if an image does well in one and not in another, the are separate campaigns.
Fair enough. Thanks guys. I guess I just have a hard time wrapping my mind around cutting an image that has a .2-.3 CTR in one age bracket and .05 in another that is separated by 5 years at most.
Don't get emotional about it heh, just compartmentalise campaigns by targeting, demographics etc. and treat each one like the others don't exist. Lander A gets 70% CTR in campaign X but 12% in campaign Y? Boo hoo the data is what it is!