Why Testing Doesn't Always Work...
I love testing... but I've found in my own tests that it's not always 100% accurate... and treating it as such can be a killer.
I just read in "How to Write a Good Advertisement" by Schwab that most advertisers (at least when the book was written... which was I think in the 60's) ignored a split-test result if there wasn't a 15% difference in response rates... because even the EXACT SAME AD could have a 15% response rate difference.
That doesn't mean a 5% and 20% CR BTW... that means say a 1% and 1.15% CR... just so we're clear.
I would suspect that these days tests are probably a little more accurate... but it's still an interesting theory... and one I personally agree with.
Anyone care to share their own experiences on this?
-Dan
P.S. Be careful though... I think Vin sleeps with a copy of this thing under his pillow, so if you are going to say something bad about the book... do it from behind some riot gear
Over $30 Million In Marketing Data And A Decade Of Consistently Generating Breakthrough Results - Ask How My Unique Approach To Copy Typically Outsells Traditional Ads By Up To 29x Or More...