Best uses for a data warehouse

Data warehousing was proclaimed by some to be the end-all of data

discovery, but it has missed this goal by a long shot. The promise of

the data warehouse has been diluted mostly because of how the user

community uses the data warehouse.

Today, I believe that data warehouses are not being used to their true

potential. From what I can tell, most companies use data warehousing

for 1) validation, 2) tactical reporting, or 3) exploration.


Validation is the where the user community validates with data what they

already believe to be true. For example, Denver consumers buy products

differently than New York City consumers. New York folks tend to

purchase a candy bar on a whim (city population buying patterns), where

Denver folks are less likely to do so (rural population buying

patterns). This has been hypothesized for years, but empirical data

shows it to be true. Another example is the question "Who are my best

customers?" Once we get past the definition of "best", I bet that most

of the user community already knows the answer. The biggest part is

getting past the definition of "best". My estimate is about 45% of the

usage of the data warehouse is validation

Tactical Reporting

Tactical reporting is where the user community uses the data for a

tactical reason. For example, salesperson Daffy Duck of Acme Corp. is

going to visit customer Wylie T. Coyote and he wants to know what

customer Wylie T. Coyote bought during the last year. There is no

comparison of customer Wylie T. Coyote and customer Roadrunner to see if

there is anything that might suggest new products to sell. My estimate

is about 40% of the usage of the data warehouse is for tactical



Exploration is where you search for ideas or knowledge that you did not

know before. This is where data mining techniques (e.g., association,

classification, genetic algorithms) and applications (e.g., market

basket analysis, fraud detection) come into play. My estimate is about

15% (I hope that totals to 100%!) of the usage of the data warehouse is

for exploration.

The potential of the data warehouse comes from exploration. But I

understand that until the user community can spend less time doing their

tactical duties, they will not be able to explore the final frontier --

to boldly go where no man has gone before (oops, sorry I slipped into my

Star Trek voice!).

ITWorld DealPost: The best in tech deals and discounts.
Shop Tech Products at Amazon