The problem with Business Intelligence (BI) Systems

BI systems have changed dramatically over the last few years and there are now two types. The old fashioned system that uses an OLAP database that extracts data from the transactional database using an ETL tool. Then the newer, modern systems that run in-memory and use a column store index that can run directly off the transactional database and run consolidations across dimensions very quickly.

Pros and cons of using them for operational reporting.

Strengths of BI Systems:

  • Good at consolidating disparate ERP systems or other databases
  • Building portals and dashboards
  • Giving high-level historical summary information
  • Looking at historical trends
  • Providing high-level executive reporting
  • Controlled by IT

Problems with BI Systems:

  • Not suitable for ad hoc queries
  • Not Real-time (Unless heavy investment in data mirroring hardware)
  • Not an end user tool
  • Uses a different database
  • Needs a ETL tool (Extraction Transformation and Load) to populate the BI database
  • New security needs to be implemented
  • Expensive to implement and people dependant
  • Unable to respond quickly to new requirements
  • Additional hardware required
  • Controlled by IT

Due to the innate complexity of BI systems, a new reporting requirement will typically require 2-3 man-days to design and implement.  Depending on the level of resource available, it can take many more days or weeks to schedule in this time. This then forces operational management to take things into their own hands and start to build data shadow systems in Excel.If we now look to the new in-memory BI systems you have similar problems to traditional BI systems, but the key issues are;

  • Aimed at C level and executives that want visualisation rather than raw numbers.
  • Still complex to set up and heavily dependant upon IT
  • Not designed for operational reporting and analysis