The life blood of a business intelligence (BI) system is data - a conglomeration of structured, unstructured, quantitative, qualitative, time-related, transactional or snapshot numbers and information all jumbled together and then cleaned, transformed, and loaded into a form of data repository.
It is critical that external environmental data be incorporated in BI.
Yolanda Smit is senior BI business analyst for PBT.
This repository is then published through various technologies and impressive interfaces for the different audiences that want and need access to the data. The question is, however, is a BI system intended to enrich intelligence in order to provide decision-makers with insight that was not possible from the source system, or is BI merely a system through which the access to data is simplified?
One of the 'side effects' of a properly designed and developed BI system is to minimise the effort for business users to access data from source systems in order to compile reports, but it was never intended purely as a factory to produce automated reports.
The life blood that adds the 'intelligence' to BI is the data, and therefore careful attention needs to be paid to two key considerations as highlighted below:
1) What data must be included into the BI system?
2) How is the data to be integrated?
BI data content
Every knowledgeable BI owner has bought into the ideal that a BI system should comprise a comprehensive warehouse of data from all business areas within the organisation. They've also assimilated the balanced scorecard 'law' that their data should not be only financial, but the fact that the metrics or key performance indicators must be carefully chosen from all business areas and across the financial, customer, internal and learning layers within the organisation.
A key component of data that seems still to be omitted is the data from outside the organisation. During strategic planning sessions, a lot of money is invested in market research on the macro and micro environment levels.
Hours, days and months are spent on analysing this data in isolation in order to decide in what direction the organisation should go, after which the strategic direction is set and the environment data is forgotten. This practice can be related to the idea of plotting the course of a ship from Cape Town to Buenos Aires based on the wind speed and direction at that particular point of time when leaving the Cape Town harbour. Setting sail and putting the ship on autopilot based on existing weather parameters would be considered ignorant, as anyone knows that weather conditions may change unexpectedly, and if not monitored, the ship may well end up in Antarctica.
Therefore, it is critical that external environmental data be incorporated in BI to empower management to adjust the course to sudden changes in the environment. External data may include macro and micro economic indicators, as well as industry benchmarks and competitor data, etc, in order to not only ensure the user is successfully steering the course, but also to measure comparative performance in the race to success.
Data integrated
Integrating data across business areas seems to be a big challenge for even the mature BI solutions in South Africa. Although internal data is comprehensively collated into a data warehouse, the data seems to remain in isolated data marts with no effective way to do integrated analysis across business areas. Business still cannot gain insight or model the cause and effect relationships between different departments.
Strategy maps can conceptually be drawn, but there's no way to measure KPIs with the relationships modelled into the data. Thus, businesses cannot monitor or model scenarios related to questions such as the following:
What impact does human resources have on the productivity of the operations department?
What is the cost impact of service delivery from operations that do not meet the expectations created by the marketing department?
If the company was to pursue a new avenue of client acquisition, what would the workload impact be on the IT support teams?
The key to data integration lies in Ralph Kimball's core design principles of conformed dimensions, and each design must be founded in the business process it represents. If each data mart represents a business process that is either preceded or followed by another business process represented by another data mart, and there is a golden thread of conformed dimensions that are present in all of these data marts, then doing cross-functional analysis, root-cause analysis, what-if scenario planning, etc, is as easy as keeping one's heart beating.
The life blood of a BI system is data, and feeding the system the right data in the right format and with the right structural design will upgrade the quality of the system from basic health to strategic BI vitality.
Share