|
To test targeting power we randomly split a file
of consumers into two parts. We then profile one half and use the results
to predict the consumer behaviour of the other half.
For example one can test the 10%
point of a gains curve. Imagine a mailing:
 If
you randomly mail 10% of the population you would expect to
reach 10% of your target market - (black
line)
|
If
you used ACORN for targeting you would reach more, perhaps 25% of
your target market -
(red line)
|
With
the new ACORN the targeting will be better -
(blue line)
|
|
 |
By repeating the testing process frequently
one determines standard deviation and can measure the statistical significance
of the improved targeting.
legal - privacy
All information presented is copyright CACI Limited 2003
|
Converting
from the old ACORN