Effective Business Intelligence

How Business Intelligence Can Help

Effective business intelligence starts with questions that will help your organization make better decisions.  For example, an Electrical Utility provider may be asking how to better understand its customer usage patterns for the next 1-6 months.  An insurance provider may be asking how to look for fraudulent medical claims.  Or an aviation company may want to know the impact of fuel prices on its flights for the next 3 months and how to maximize profit.  Regardless of the situation, BI is all about helping the business.  Anything less and all you are doing is providing reports.

Once you have your organizations BI questions the real work begins.  Does your organization have access to the right data to help answer the question?  Some of the data could be external to the organization and will need to be sourced and loaded into an environment that is accessible.  In many cases, this means taking data from multiple systems and merging it together.  How well do you trust the data that you are going to use for these questions?  For instance, using bad data to create a 6 month forecast model and then enacting business decisions based on that model will result in poor decisions.  For each data source and question you need to understand how clean the data must be.  Data quality at 100% is time consuming and expensive.  You need to understand the data you are leveraging to know the risks associated with using the data.

Now that you have the data sourced, you have to determine the best method to assemble the information to make if meaningful to the people asking the questions.  Are they looking for an interactive experience that enables them to slice and dice the data?  Would they prefer the data be used to create a forecast model and only need the results and how the model was built i.e. Linear Regression, Decision Trees, Market Basket Analysis?  Does having the data in a single large dataset that they can connect to with other tools, such as R or SAS and give them what they need to make decisions?  In today’s environment business users have many different skills and experience levels to analyze the data.  Before building the solution, you need to know exactly how to deliver the content.  Gone are the days of building just star schemas and OLAP cubes to deliver self-service BI to the masses.

Leave a Reply

Your email address will not be published. Required fields are marked *