A global telecommunications company had developed a sophisticated network and deployed statistical monitoring tools that were recording more than 1 million data points per day. However, these tools were collecting data that was of interest to engineers, not business managers. As a result, significant analysis of the data was required before business decisions could be made. A few of the negative consequences included:
1. Failure to disconnect un-used circuits in the network
2. Incorrect decisions when purchasing circuits for redundancy
3. Inability to manage issues by exception
4. Labor-intensive analysis process, repeated every week
5. Inability to add new statistics without increasing workload
After reviewing the current statistical tools and analysis procedures, an automated, scalable business intelligence engine was established, with a web front-end and automatic report distribution by email. This system included three main components:
1. Modular data collection engine
2. Optimized report-writing capabilities
3. Business intelligence reports
Modular Data Collection Engine. More than 1 million data points were being collected daily from the network, through a variety of tools. Some data was accessible through SQL, while other data was stored in home-built data sources. A modular, XML-based API was developed to absorb data from all current data sources. With a pluggable architecture, this API will automatically absorb new data sources and types without additional integration.
Optimized Report-Writing Capabilities. Although the data was aligned with the network, business management required information that was aligned with the internal structures of the company. To meet this requirement, an optimized report-writer was developed, allowing users to cooperatively develop and share reporting components. Shared components could be incorporated as-is, or another layer of business rules could be applied to meet each individual business owner’s needs.
Business Intelligence Reports. The report writer produced a reporting object that could be easily manipulated to provide specific kinds of business intelligence. In addition to traditional aggregate statistical measures, a full range of exception reports and graphs could be activated without increasing the amount of time required to generate a report. Users can view specific reports in real-time via a web interface, or schedule reports to run and send results via email. Full integration with Excel was also provided, allowing business users to manipulate report data with the tool they are most comfortable using.
Within two weeks of concluding this project, the customer reclaimed two full-time analysts who had been dedicated to the prior labor-intensive process. This cost recovery alone provided an ROI many times the one-time investment required to install and activate the solution.