Is your organization experiencing more hype than value from your business intelligence (BI) initiatives? BI is receiving a lot of space in business magazines, air time in vendor webinars and prominence in research monographs. If you’re disappointed by what BI has delivered so far at your organization, then consider if the following issues stand between you and the BI value you’ve been promised.
Lack of adequate data quality
No BI software, no matter how sophisticated, can gloss over glaring gaps in the data quality of the databases you’re trying to integrate for your BI application.
For example, your customer identifiers in the billing system may not line up well with customer identifiers in the customer relations system. Similarly, simple spelling problems in addresses or incompatible reference codes can be enough to create problems.
Often a good approach to improving data quality is to first use the BI software to scope this data problem and help to resolve it.
Excessively complex goals
If your BI project charter contains lofty Star Wars goals that your software, data, systems and business processes can’t support, you will be disappointed.
Examples of Star Wars business goals include unrealistic goals for sales increases, customer retention improvements or mean service time decreases. On a technical level, examples of Star Wars goals include overly complex integration and expecting to process very large databases in very short timeframes.
A more realistic approach is to first tackle a business problem that can be materially reduced with the involvement of only a small number of databases and their responsible departments. If you’re not sure about what aspect of BI might produce the most results quickly, explore this list of topics within the broader BI domain.
Gaps in the application portfolio
Many organizations have gaps in their application portfolio that will undermine BI success. These gaps are typically bridged with a significant number of large spreadsheets that cannot be smoothly and reliably integrated into a BI application.
For example, sales analysis may be performed by merging invoice and inventory data in a giant workbook with multiple worksheets. Similarly, product returns may be tracked in a large worksheet.
In this situation, implementing a new system to fill the gap and deferring the BI project is often the most realistic approach.
Insufficient expertise with data and tools
BI software vendors often tout ease-of-use, superior development productivity and sophistication of their tools. While there’s a lot of truth in these claims, these advantages can’t eliminate the need to train development staff in the use of the tools, the details of your databases and at least a cursory knowledge of your business.
For example, without detailed knowledge of the structure and content of your databases, development staff will invest huge amounts of effort on many prototypes before meaningful BI functionality will emerge.
Address this problem by budgeting for training and coaching. Also expect to assign an expert in your business to the BI project team to contribute detailed requirements and to review designs.
For more detail on best and worst practices for BI project success, read this Forrester blog and download this Data Warehouse report.
What do you think? What are your thoughts on the most important considerations to ensure the success of a BI project?
