Re: Acceptable Click Thru Rate

From: Christie Brown <clbrown_at_umich.edu>
Date: Sat 29 Jan 2000 11:07:06 -0500

DENNIS BALL WROTE:
> My company is considering banner advertising with one
> of the big networks. I'm seriously considering Double
> Click. Are they as good as they claim to be? Can anyone
> please tell me what click-thru rate I should expect
> initially? After a month's worth of banners on their
> network, I plan to limit the banners to the best
> performing sites. What click-thru rate should I expect
> then? Assume my company's services have a broad range
> of appeal. Any help would be greatly appreciated.

TO WHICH KIM BROOKS RESPONDED:
> I once presented at a WSA meeting that included a UW
> MBA professor of marketing, specializing in Direct
> Mail. She said she always asked her students on the
> final: what response rate should you expect from a
> direct mail campaign? If anyone answered "6%" (the
> average), she killed them. Why? Because expecting the
> average is pretty pessimistic, and you are bound to get
> below-average. If you are doing your job well, you
> will blow away averages. I always think of her
> comments in the context of the sad, low-response world
> of banner ads...

MY RESPONSE:

I do not mean to offend, but what this "professor" said
is so thoroughly wrong that it requires comment. She
is, of course, correct in saying that if you are doing
your job well, you will blow away averages. But you
are really foolish if you EXPECT or PREDICT other than
the average.

Look what happens if we "aggregate" this prediction:
Say everyone in the world takes your advice. So
everyone predicts a better-than-average response. But
by definition the average response of of all these
optimists will be the average click-thru rate. Thus,
if everyone does as you suggest, they will generally do
much worse than they predict. (And some of them will
do truly terribly.)

Stastically, the only reasonable expectation is to
expect the average (for that type of ad, using that
tecnique, in that category, on that site, etc.). Now a
lot of people scoff at statistics but I'd bet just
about any amount of money (not that I have any) that if
I used the average to predict a 100 click-thru rates
for randomly-chosen ads and YOU used your theory, my
predictions would be better than yours.

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.



Received on Sat Jan 29 2000 - 10:07:06 CST