My first Publication Agile-Data-Warehouse-Design-eBook | Page 282

262
Chapter 9
Why and How
Why Dimensions
Why details become
causal dimensions
that help to explain
why facts occur in
the way they do
Direct causal factors
have a recorded
influence on the
facts
Indirect causal
factors may or may
not have influenced
the facts
Causal factors can
be internal: under
the control of the
organization, or
external: beyond
its control
The why details of an event become causal dimensions, such as promotion,
weather, or just reason. Causal dimensions explain why business events occur
when they do, in the way that they do. They describe what stakeholders believe are
the influential factors for a business event; for example, price discounts driving up
sales transactions, or storms triggering home insurance claims. Causal factors fall
into two categories: direct and indirect.
Promotional discounts are examples of causal factors that are directly related to the
facts. You know with absolute certainty when they are or are not related to a sale
because the promotional code (or discounted product code) and the discounted
price are recorded or not recorded as part of the sale transaction.
Other causal factors—such as weather conditions, sporting events, or advertise-
ment campaigns—are only indirectly related to facts. Stakeholders may know that
these took place at the same time, in the same location as the facts they want to
measure, but they can only speculate that they had an effect on them.
Causal factors can also be described as external or internal. Weather and sporting
events are examples of external causes that an organization has no control over
(unless it is sponsoring the sporting event). Whereas, price discounts and advertis-
ing are examples of internal causal factors which the organization does control.
Some internal causes—like seminars, sales calls and advertising—can be significant
business events in their own right, and warrant dedicated fact tables to analyze
their associated costs and activities by who, what, where, and when. In these cases
causal dimensions may be conformed across multiple cause and effect star schemas
that typically represent process sequences.
Internal Why Dimensions
PROMOTION is an
internal why
dimension that can
contain a mixture of
direct and indirect
causal attributes
Figure 9-1 shows a simple PROMOTION dimension. This is an internal why
dimension that would typically contains a combination of discount, display and
advertising descriptions. These are a mixture of direct and indirect causal factors.
DISCOUNT TYPE is a direct causal factor captured on every transaction along
with a DISCOUNT amount fact ($0.00 when there is no discount). Advertising
attributes such as CHANNEL are indirect causal factors if there is no way to know
for sure that customers saw the adverts. However, if the DISCOUNT TYPE is
“Coupon” or “Discount Code” and an advert contains the information that the
customer must supply at the point of sale, then it becomes a direct causal factor.
The special “No Promotion” record (PROMOTION KEY zero) will be the most used
record in the PROMOTION dimension, if most products are not on promotion
every day.