Stats Case Thoughts

[Daniel Zarovy]

Hey Guys,

I've been working on the Stats case and thought it might help us to share a few thoughts before our meeting tomorrow.  Feel free to share/comment if you've had time to look at it.  Otherwise, we'll obviously go over it all tomorrow.

Here's my thoughts so far...

• My regression results improved significantly when I used moving averages for the Consumer Packs (92% current month, 8% previous month) and Dealer Allowance (weighted over 3 months, w/ current month having most weight)
  
• There's no sales data for 1983, but some of the 1983 data can be used to create moving averages for the dealer allowances
• I converted the Dealer Allowance budget into cases using an average cost of $6/case.  I don't think it affected the regression, but easier to mentally see what's going on when using the same units as sales.
• Any ideas on how to make Month/Year significant?  (Month on it's own would be duplicating what the Index provides, but Time does play some factor...I attempted to combine year and month into one "Time" value.  Unfortunately, it doesn't seem to help the regression.  I also tried multiplying Time and Index...still no luck.)