Business Intelligence in Practice
Companies invest a great deal in database solutions (ERP or enteprise resource planning, typically covering GL, purchasing, sales, warehousing etc.) which span the enterprise, even more in implementing these solutions, and often do not have the opportunity to leverage the depth and potential insights the database offers.
Many of the technologies needed to unearth the intelligence in the heap of data, are purchased with the ERP. The skills to use these technologies and develop bespoke solutions using them is know as Business Intelligence (BI), or in some circles as Business Analytics. You'll find examples of business intelligence here to provide more insight into what it is.
Some areas also include some systems analysis influences, but they’re included as they still address how to practically address business needs using BI.
Provide executives with the information they most need to manage the business by seeing its heartbeat with a simple set of metrics shown over time.
Reduce stock overage and shortage, and reduce administrative costs of purchasing.
Optimize and automate creation of purchase orders by analyzing past sales trends to use in forecasting, average lead times of past purchases to forecast lead time, and analyze seasonal changes.
Understanding which lines of stock in which locations are being used inefficiently.
Develop reporting which analyses inventory positions against sales by location and item hierarchy over time.
Sales and Marketing
Detecting cashier behavior that indicates fraud.
Use historic information about known fraudulent users to predict users exhibiting similar patterns of behavior. Use data mining algorithm Memory Based Reasoning (MBR) to compare metrics of individual point of sale transactions, rolled up to each user.
Compare sales performance of promotions before and after promotions to determine if profits were maintained by promotions.
Control Customer Sales Cycles
The attrition of customers usually follows a pattern of behaviour. This is typically times from signing up from an offer, changes in buying patterns, changes in time to settle invoices, periods of delinquency, By using averages of these metrics and using data mining against known cases of attrition, the most likely customers at particular points in the sales cycle can be identified and contacted to intervene.
Customer clustering techniques through combinations of demographic information (internal to the company such as buying patterns, or external such as income and sex) customer type (such as price groups and sales channels) and buying patterns. These enable targeted promotions to the most responsive audience.
Market basket analysis of all past orders reveals which items are typically bought together and enables prediction of likelihood of one item when another is already placed in an order (basket).