Predicting Daily Store Sales
We have observed beer sales on a daily basis for a major retailer for one of their outlets for the last 2 years. Our objective is to model/forecast beer sales for a particular SKU for a particular store and to make projections for the next 140 days although in practice a much shorter and much more accurate horizon would be used (e.g. 5 days).
Here is some background information about Beer Sales:
The USA Today breakdown of Holiday Sales vs. the Holiday effect for the SKU at the one
Outlet that we studied
Holiday % of the Year % of the 7 Holidays % of the 7 Holidays
July 4th |
7.5% |
17.4% |
12.7% |
Memorial Day |
6.8% |
15.8% |
8.7% |
Labor Day |
6.6% |
15.3% |
11.1% |
Christmas Day |
6.0% |
13.9% |
42.0% |
Thanksgiving Day |
5.6% |
13.0% |
9.7% |
Easter |
5.3% |
12.3% |
0% |
Super Bowl |
5.2% |
12.1% |
15.9% |
Total 43%
Here is a word on the importance of modeling in order to forecast……
Bacon, writing in Novum Organum about 400 years ago said: "Errors of Nature, Sports and Monsters correct the understanding in regard to ordinary things, and reveal general forms. For whoever knows the ways of Nature will more easily notice her deviations; and, on the other hand, whoever knows her deviations will more accurately describe her ways." [ II 29]