Re: Acceptable Click Thru Rate

From: Christie Brown <clbrown_at_umich.edu>
Date: Fri 04 Feb 2000 07:00:42 -0500

CHRISTIE BROWN WROTE:
> In my class (yes, I am a card-carrying genuine MBA
> advertising professor), you'd better have a good reason
> for predicting higher than the average.

TO WHICH JOHN WHITESIDE REPLIED:
> And this ignores whether the mysterious average is even
> good. I've done direct mail campaigns where I got 15%
> response and was content -- because it was what I
> expected -- but not thrilled. I've been thrilled with
> 2% because knowing the list and offer and our goals, I
> was expecting and would have been happy with 1%.
>
> If you are going after the same audience repeatedly,
> the only way to make really good predictions is to test
> everything you do and build up some experience.

Yes, this is EXACTLY right - test and get better
information, then you will be able to make more
accurate predictions. But in the original question
the professor rejected the average as a valid
prediction because it "was too pessimistic...a good
manager should be able to blow the average away."
(Forgive the paraphrase.) In other words, the supposed
reason for raising the prediction was optimistic
self-confidence, not learning from careful testing or
experience. "I'm going to blow away the average
because I am a better manager" is a poor bet, but "I'm
going to blow away the average because my tests have
shown that this technique in this category yields
higher-than-average results," is a good reason for a
high prediction.

I didn't mean to imply that anyone should be placing
ads at random, only that in the absence of ANY other
information about an ad (the case for the professor),
the average is the sensible prediction. In John's
example he has other information, from testing and
experience.



Received on Fri Feb 04 2000 - 06:00:42 CST