I'll bet my house that you, the marketing professional, have no idea what the folks down in the data-warehousing department do. If you're a database manager, odds are you don't know what the suits in marketing do either. Yet I'm sure you both know that, somehow, you depend on each other every day.
Ironic, isn't it?
Hopefully, this article will help fix that. And with this knowledge, you'll knock the chocolate out of the competition.
WHAT IS DATA WAREHOUSING?
A data warehouse is a big database that uses large stores of transactional data to analyze the business and support decisions. A company can use a data warehouse, for instance, to help segment the market, discover latent customer needs, make distribution decisions, help guide production numbers, and make a host of other business decisions.
A data warehouse is not a transactional database. A transactional database uses data to help operate the business. For instance, a hospital may use a transactional database to call up a patient's medical history; an insurance company may use its transactional database to help match a customer with the proper insurance rates. Transactional databases use current data - say 60-90 days old. They are built for speed and efficiency to keep the company's day-to-day operations moving fast.
Unlike a transactional database, a data warehouse uses data to help analyze the business and make decisions. A hospital's data warehouse, for example, would look at all of its patients' information to decide whether or not there is a need for additional oncologists. The insurance company may use its data warehouse to set auto insurance rates by analyzing driving records and habits. A data warehouse is built to analyze vast amounts of data that spans many years. Unlike a transactional database, the data warehouse often runs very slow.
HOW IS A DATA WAREHOUSE USED?
So a data warehouse stores a lot of data and analyzes it. Big deal.
Its usefulness comes from the ability to extract marketing information not offered by the raw data alone. Extracting such information is called data mining. Managers mine data to find hidden patterns and relationships among groups of data. These hidden patterns help companies make an endless variety of smart business decisions and forecasts.
A catalogue company may mine its data warehouse to find those customers who are most likely to buy something from a mass mailing. This helps the company minimize the number of mailings it makes to customers who simply won't buy anything no matter what the company sells. A bank may mine its data warehouse to predict which loan candidates are most likely to default or go bankrupt.
A data warehouse's design is what enables such insightful analysis. It does this by organizing factual data around "dimensions." For instance, a dimensionally designed data warehouse might store fact tables like sales quantities, inventory levels, stockouts, and revenue in a table. Related to those tables may be tables with dimensions such as time, location, customer, product and store.
From this structure the firm can conduct some pretty robust analysis. It can look at sales quantities by region and season and time of day. It can analyze stockouts by store and region. I can even look at sales quantities by time of day, customer gender and product type. From this analysis the manager can make a vast array of intelligent marketing decisions.
DEFINITIONS
That's the long and short of it for data warehousing. As you learn more about it, you may grow confused by all the acronyms and definitions. Below is a short list of some of the most common data warehousing terms. You can refer back to this when the data warehousing folks start using those funny nonsensical words.
OLAP - On Line Analytical Processing. Describes the tools used to analyze data warehouse information.
OLTP - On Line Transactional Processing. Describes the tools used to make transactions with data.
Metadata - Data about data. Describes the data you have.
Data Mart: Separate, smaller warehouses typically defined along organization's departmental needs. This selectivity of information results in greater query performance and manageability of data. A collection of data marts (functional warehouses) for each of the organizations business functions can be considered as an enterprise warehousing solution.
SQL (Structured Query Language): The computer language used to ask a database to perform a function.
Decision Support Systems: Another way of describing data warehouses and other technologies that support business decisions.
Executive Information Systems: These are business intelligence tools that are aimed at less sophisticated users, who want to look at complex information without the need to have complete manipulative control of the information presented.
Scrubbing Data: The processes of altering data from its original form into a format suitable for business analysis by non-technical staff.
