Initiating A BI 2.0 System Project
Business intelligence uses key performance indicators [KPIs] to
assess the present state of business and to prescribe a course of
Historically, this type of performance data was manually imputed
or configured into weekly or monthly reports.
Today, this data must be available at shorter intervals, with reduced
delays. To assist in this, the KPI methodology has been further
expanded with other peformance methodology to incorporate KPIs and
root cause analysis into a single methodology.
The time requirement for performance data largely depends upon
the criticality of time in the provision of a product or service.
For instance, the banking industry faces higher operational/credit
risk loading, as do credit card companies and "wealth management"
services, Weekly data is insufficient. Daily will be more effective.
Hourly is mostly unnecessary. The understanding of these varying
needs is the first step in designing a BI strategy. Developing beyond
requirements is not only costly, it also delays projects, and as
such delays competitiveness.
The first step in any BI project, is the intiation phase. During
this time, the scope of the project is determined and the business
requirements identified. This can be achieved by following the underlying
Goal alignment helps the company identify where and why a BI system
will add value. It determines the short and medium-term purposes
of the programme. This must include:
- Which strategic corporate goals will the system address?
- What organizational mission/vision does it relate to?
A business case will be developed to outline how this initiative
will eventually improve results / performance [i.e. a strategy map].
This is a current analysis stage to determine:
- Current information-gathering competency
- Current capability of monitoring important sources of information
- Data the organization currently collects
- How current data is stored
- Statistical parameters of current data - e.g. how much random
variation does it contain? Does the organization measure this?
Cost and Risk Analysis
This analysis assesses the financial consequences of a new BI initiative.
- Assess the cost of the present operations
- Identify any increase in costs associated with the BI initiative
- Identify the risk that the initiative will fail
This risk assessment should be converted into a financial metric
and included in the planning.
Customer and Stakeholder Benefits Analysis
This analysis determines who pays and who benefits for the BI system.
It also includes all stakeholders that are impacted by the system
in ways other than direct benefits [indirect benefits] and those
that may be resistant to any change in current procedure:
- Who will benefit from the initiative
- Who will pay
- Those with a stake in the current procedure
- Customer and stakeholder groups that will benefit directly
- Customer and stakeholder groups that will benefit indirectly
- The quantitative / qualitative benefits to each group
- Alternatives to increase satisfaction for each group
- Way in which benefits will be monitored
- All others impacted by the intitiative - employees, shareholders,
distribution channel members etc.
Information requirements identified from above analysis must be
operationalized into clearly defined metrics. This phase of the
- What metrics to use for each piece of information being gathered.
- Why these metrics are the best to use
- Validation for using a particular metric
- The number of metrics to be tracked
- How they will be tracked - what kind of system can be used to
- Standardization of the metrics, so they can be benchmarked
against performance in other organizations?
- Industry standard metrics available
Although measurement is an extremely important part of not only
implementing a BI system, but in it's ongoing use, the gathering
and recording of measurement needs to be acceptable both in terms
of intrusiveness and cost.
The project plan must therefore establish a methodology or a procedure
to determine the most acceptable way of measuring the required metrics.
- Methods used
- Frequency of data collection
- Industry standards for collection method and frequency
- Validation as to why the suggested method is best
The BI program will be monitored to ensure that objectives are
being met. Like any complex, long term program, adjustments in the
programme may be necessary.
The programme should be tested for accuracy, reliability, and validity.
- How to demonstrate that the BI initiative (rather than other
factors) contributed to a change in results.
- How much of the change was probably random.
Business Intelligence Databases
The basic form of data or information used in BI is stored in
the databases. When the amount of data is significant (i.e. >
10 GB), it is transformed into a data warehouse.
A Data warehouse houses several databases, typically from several,
often disparate engines. For instance, one may be Oracle-specific,
another SQL Server-specific, or even half SQL Server and half DB2.
Each of these database configurations stores data in a different
format. A Data Warehouse stores the combined data and resolves these
differences in format and other variances.
When a database application is created and attached to data base,
certain properties need to be considered.
Access Permissions - within any given application,
different users are created and given certain permissions: Admin,
Director, Manager, SalesPerson, Temp. Some of them are made administrator
while others are made users.
Roles - whilst the application may be 'general'
its abilities can be refined to suit certain roles [Director, Manager,
Operational Supervisor etc.]
Tasks - the overall application handles different
tasks, depending on the role of the current user — certain
parts are exposed and enabled while others are not.
A Data Warehouse
Back To Top