We have a digital campaign running, for Brand A, which we decided to survey. What we do is gather a Control group (before the campaign start) within the footprint of the campaign and ask them which brands they think are “Cool”
We then survey people that saw one of the creative formats, either Flash banners, Video or Rich Media.
Turns out, from the looks of the results (see above), that the banners are great at generating the “Cool” messaging, but terrible at branding since it helps increase the coolness level of all competitive brands.
Hmmm… something doesn’t look right…
I saw the results and went “Wow!!!”…
Then I remembered the tip I give to my colleagues… “when a number creates an emotion in you… call me.”
So I’ve consulted myself (Multiple personalities) and thought a bit harder
(Spoiler below, don’t read yet if you want to try and guess what’s fishy here)
Turns out the targeting is different for Flash banners (way more targeted to “Cool” people) compared to Video & Rich Media.
So my Control group represents the average targeting of the campaign, so does my Rich Media and Video. My banners, on the other hand, ask “Cool Kids” which brands are “Cool”. Obviously, the Cool Kids have a different opinion about the surveyed brands, therefore highly skewing the results.
Now to fix it!
So I’ve divided the Control group in 3, per type of targeting/creative and compared the lift ratio instead of pure lift %. (Eg. If Banners Control Group: 15% and Banners Exposed: 20%, then it’s a pure 5% lift, but it’s 33% higher than the initial 15%, which speaks louder here.)
Turns out Strong Messaging/Weak Branding still has something to do with it, but it is minor.
Banners showed to be incredibly more targeted than Video & RM. But Video is much better at increasing Brand A’s Cool-itude level
Now… to figure out a way to retarget banner viewers with Video for this campaign…