Business intelligence
This is the second in a series of articles on how managing information and data can help your business. This article looks at data warehousing and data mining.
In simple terms, a data warehouse is a large database containing consolidated, historical and topical operational data that has been re-modelled for analytical purposes. It may be directly accessible by business users and have a business user context but not necessarily.
It is often the creation of a business user interface that provides the translation between the data warehouse database design and the common business terms. The key element of the user interface is referred to as meta-data (information about data) or a semantic layer and it provides the translation of the consolidated database design into common business terms.
Meta-data can pre-define common business measures and formulae and wrap them up as simple re-usable definitions, ensuring there is a single corporate version of the truth or what is commonly understood across the organisation.
To see where the use of data warehousing may be of benefit, consider a bank acquiring new credit card accounts. The bank employee who signed up the customer will be keen to register the account form the day the customer signed up, as their performance may well be measured against the amount of new business they generate. In corporate terms however, this may be too simplistic.
Customers who open an account but do not use their card, or cancel their agreement within 3 months of signing up, may actually cost the bank money. In such cases, the bank will want to differentiate between new sign ups and active accounts. Sign-ups may be counted at the time the customer signs, but active accounts may be considered to be those that have transactions equal to or above a certain monetary amount within the first 3 months after sign up.
Defining sign-ups and active accounts differently within the semantic layer by incorporating the necessary qualifying factors, ensures at a corporate level, that there is no ambiguity between the different performance measures. Anyone using these pre-defined measures in reporting analysis will always be using the same consistent definition. It also provides a starting point for managers to analyse the relationship between sign-ups and account activity, thus promoting a better understanding of the types of customers and activity that generates revenue rather than costs.
The semantic layer that sits on top of the data warehouse is the enabler for those capabilities termed Management Information (MI) and Business Intelligence (BI). These terms are closely aligned but BI typically offers a more hollistic benefit. Management information has historically been used to refer to the ability to report upon Key Performance Indicators (KPIs) for the business. The MI viewpoint tends to provide the benefit of consolidated reporting, but in a pre-determined, scheduled fashion such as daily sales reports that show progress against targets.
Business Intelligence takes this a step further by not only reporting out the current position of the business, but allowing detailed analysis (often referred to as data-mining) of current and historical information.
Key performance measures such as new customer accounts or product sales are combined with dimensional context information such as the store where a sale was made, demographics relating to the customer and time dependant elements. This holistic view of business information allows users not only to see how well the business is performing now, but gives insight into historical trends, customer profiles and their spending habits, the effects of changes to products and pricing structures and a wealth of other analytical information.
The next and final article will look at how effective use of Business Intelligence can be a powerful planning tool for your business. www.bis.gi <http://www.bis.gi/>

Comments(0)