The promotion dimension is potentially the most interesting dimension in theretail sales schema. The promotion dimension describes the promotion conditionsunder which a product is sold. Promotion conditions include temporaryprice reductions, end aisle displays, newspaper ads, and coupons. This dimensionis often called a causal dimension because it describes factors thought to cause achange in product sales.Business analysts at both headquarters and the stores are interested in determiningwhether a promotion is effective. Promotions are judged on one or more of thefollowing factors:■ Whether the products under promotion experienced a gain in sales, calledlift, during the promotional period. The lift can be measured only if the storecan agree on what the baseline sales of the promoted products would have been without the promotion. Baseline values can be estimated from prior saleshistory and, in some cases, with the help of sophisticated models.■ Whether the products under promotion showed a drop in sales just prior toor after the promotion, canceling the gain in sales during the promotion (timeshifting). In other words, did you transfer sales from regularly priced productsto temporarily reduced priced products?■ Whether the products under promotion showed a gain in sales but otherproducts nearby on the shelf showed a corresponding sales decrease(cannibalization).■ Whether all the products in the promoted category of products experienced anet overall gain in sales taking into account the time periods before, during,and after the promotion (market growth).■ Whether the promotion was profi table. Usually the profit of a promotion istaken to be the incremental gain in profi t of the promoted category over thebaseline sales taking into account time shifting and cannibalization, as wellas the costs of the promotion.The causal conditions potentially affecting a sale are not necessarily trackeddirectly by the POS system. The transaction system keeps track of price reductionsand markdowns. The presence of coupons also typically is captured withthe transaction because the customer either presents coupons at the time of saleor does not. Ads and in-store display conditions may need to be linked from othersources.The various possible causal conditions are highly correlated. A temporary pricereduction usually is associated with an ad and perhaps an end aisle display. Forthis reason, it makes sense to create one row in the promotion dimension for eachcombination of promotion conditions that occurs. Over the course of a year, theremay be 1,000 ads, 5,000 temporary price reductions, and 1,000 end aisle displays,but there may be only 10,000 combinations of these three conditions aff ecting anyparticular product. For example, in a given promotion, most of the stores would runall three promotion mechanisms simultaneously, but a few of the stores may notdeploy the end aisle displays. In this case, two separate promotion condition rows
would be needed, one for the normal price reduction plus ad plus display and one
for the price reduction plus ad only. A recommended promotion dimension table
is shown in Figure 3-11.
From a purely logical point of view, you could record similar information about
the promotions by separating the four causal mechanisms (price reductions, ads,
displays, and coupons) into separate dimensions rather than combining them into
one dimension. Ultimately, this choice is the designer’s prerogative. The trade-off s
in favor of keeping the four dimensions together include the following:
■ If the four causal mechanisms are highly correlated, the combined single
dimension is not much larger than any one of the separated dimensions
would be.
■ The combined single dimension can be browsed effi ciently to see how the various
price reductions, ads, displays, and coupons are used together. However,
this browsing only shows the possible promotion combinations. Browsing in
the dimension table does not reveal which stores or products were aff ected
by the promotion; this information is found in the fact table.
The trade-off s in favor of separating the causal mechanisms into four distinct
dimension tables include the following:
■ The separated dimensions may be more understandable to the business community
if users think of these mechanisms separately. This would be revealed
during the business requirement interviews.
■ Administration of the separate dimensions may be more straightforward than
administering a combined dimension.
Keep in mind there is no diff erence in the content between these two choices.
NOTE The inclusion of promotion cost attribute in the promotion dimension
should be done with careful thought. This attribute can be used for constraining
and grouping. However, this cost should not appear in the POS transaction fact
table representing individual product sales because it is at the wrong grain; this
cost would have to reside in a fact table whose grain is the overall promotion.
đang được dịch, vui lòng đợi..