| Date: | Wed, 4 Oct 2006 15:46:10 -0400 |
| Reply-To: | "Musial, Jerry" <jerry.musial@CINGULAR.COM> |
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
| From: | "Musial, Jerry" <jerry.musial@CINGULAR.COM> |
| Subject: | Stats: Market Basket Analysis?? |
| Content-Type: | text/plain; charset="US-ASCII" |
I am interested if anyone has completed or has an idea for type of
modeling technique(s) could be used to solve the following:
We want a 'Next best action' model to determine the which product an
existing customer would likely purchase next. For example, if they have
previously purchased the Tetris1 game or a Ludacris ringtone, we want to
send a message telling them about the new Tetris2 game or the new
Ludacris or Young Dro ringtone. We have been discussing using a Market
Basket Analysis, but to me that means 'If A occurs in the market basket,
then B also occurs in the market basket'. However, B does not yet
exist, as it will be released next week or next month. I can certainly
see using MBA for existing products, but we are interested in 'New'
products.
Thanks in advance
Jerry Musial
Advanced Analytics
Cingular Wireless
|