'Real Time Business Intelligence or Real Time decisioning is a big discussion item. Our recommendations are:
- Assume that real time Business Intelligence is not needed, unless proven: As real-time or online feed into a data warehouse is a complex subject, and one should have highly quantified business case to go ahead with it.
- Avoid non-operational needs for real-time BI: We don't think that a decision maker needs current real-time data for taking managerial or strategic decisions. Most of the operations related decisions do not require real time Business Intelligence.
- Focus on select operational needs: Some examples of real time BI are:
- financial reporting near period end close, whereby you need to keep track on almost hourly basis on the orders booked and invoices raised.
- Managing your infrastructure: For telecommunication companies, keeping a track of the network usage therefore fine-tuning the network band-width allocation to various channels.
- Risk Management: When you are authorizing a major spending on a credit card, the real time view of what customer has spent in the day helps. This is a typical case with stolen credit cards.
Tips to structure your real-time BI
- Use real-time BI for very special cases. Avoid as much as you can.
- Avoid using data federation tools as your enterprise BI solution
- Execution-MiH recommendation will be that in spite of a data set being loaded real-time, it should also be made part of the end of the day ETL routines for the traditional data warehouse. This ensures that nothing gets missed out.
|
