Four tutorial examples are supplied with WITNESS Miner illustrating several ways in which it can be used. The first is data analysis pure and simple - the others explain how the results can help complement simulation modeling and analysis with WITNESS, Lanner's leading simulation software. It is important to note however that the latter three tutorials are data mining studies in their own right, well worth exploring even if you are not involved in simulation. They each identify important trends and information that can be used to achieve enhanced business performance.
Customer Data Analysis - Increasing Mailshot Accuracy
This tutorial examines customer records from a software company. The data includes whether the customer responded to a recent company mailshot. WITNESS Miner is used to determine the most important factors that influenced the response. This enables better targeting with the associated lower costs for future mailshots.
Production Analysis - Identifying key factors affecting productivity
This analysis is based on a table of results generated from a computer simulation model although it could well be historical data too. It identifies the key factors that give the most profitable results (complementing the simulation analysis).
Call Center Modeling - Identifying hidden performance triggers
This analysis centers on a large database of call answering performance by different teams at a call center. Beyond the obvious analysis that factors such as the number of staff affect service levels WITNESS Miner is used to identify trends that are better hidden in the data obscured by the more obvious factors.
Insurance Selling - Idenfying key target customers for cross selling opportunities
This example shows a large data sample (30,000) records being used to identify the general rules affecting purchase of add-on products. (For WITNESS Users the results offer better data for the setting of timings for inputs into a business process model simulation). This tutorial is different from the others in that is uses macro level data - i.e. each record represents a group of customers.