Many BI initiatives have failed because the tools weren't accessible enough to end users, so the BI capabilities were never really applied. Users often relied on the BI system to pull data from the data warehouse and import it into Excel, bypassing the system's analytic capabilities that, to them, seemed overly complex or not very useful.
Fortunately, as BI evolves from nicety to necessity, new technologies are blurring the lines between desktop applications and the back-office systems that hold the data. This is allowing end users to access BI information as part of their day-to-day application and desktop environment in a context that makes sense to them, which removes the most significant barrier to BI success. Organizations can use the lessons learned from past efforts to uncover the guidelines and best practices that help ensure success in future BI initiatives.
Lessons Learned
BI technologies were a response to the unique and constantly evolving information needs of companies. These needs were rarely fully met by the front- and back-office applications used by employees in day-to-day operations. Reporting was usually an afterthought for most of these systems.
BI was supposed to solve that by making objective information widely and easily available throughout the organization. So what happened? Most often it was a classic case of technology for technology's sake. Most failed BI initiatives were simply a result of not meeting the end-users' needs effectively.
For example, end users were often not fully included in the development of the data warehouse, which is typically the foundation of the BI system's capabilities. The standard reports that were pulled from the data in the data warehouse were not necessarily useful, and end users often found the reporting interface of the BI system that was used to create ad hoc reports to be counterintuitive.
Not having been involved up front in the definition of the reporting capabilities left users feeling that there just wasn't enough value to justify the BI system's use. Besides, the tools they already had on their desktops seemed to serve them just fine. After all was said and done, the perception was the BI system provided reporting capabilities that weren't much more useful than what users had before.
Users turned to spreadsheets such as Excel to give them the information they need. The problem is that getting the data to enter into the spreadsheets is an inefficient and error-prone process. Users must typically gather information from multiple systems, often with duplicated information fields and some with more current information than others.
This produces numerous potentially conflicting analyses that are also not easily shared. Objective business decisions based on these spreadsheet analyses become difficult. The most politically savvy employee with the most persuasive arguments - and not necessarily the employee with the most accurate information or most insightful analysis - is deemed the holder of the spreadsheet of truth.
So, it is clear that spreadsheet-based BI is not only insufficient, it is potentially damaging to an organization. Line executives confidently, but mistakenly, make business decisions based on the information from spreadsheet analysis that can lead separate functions in diametrically opposed operational directions. With companies creating more data of all types and changing that data more frequently, it is a bigger challenge than ever to organize and make sense of this data. An enterprise-wide BI system that fosters consistency and alignment of the decisions across functions is a mandate going forward. The viability of the organization depends directly on the information that these systems provide.
Thankfully, BI vendors have responded with products that help unify databases, application environments and desktop tools - and system integrators can provide the expertise to fully integrate the storage infrastructure with the database, application and desktop environments. Now developers, administrators and end users are able to create complete BI solutions that make high quality information available widely and consistently.
Involve End Users Early and Often
Organizations that focused on infrastructure issues while trying to make BI work were not wrong to do so. The data warehouse is the foundation of any BI initiative. Get this wrong, and the effort is sure to fail.
Today, the data warehouse is no less important. Data warehouses are large, all-encompassing repositories that must bring together many different types of data, including structured transaction data and unstructured documents, which can be active or archived. The data warehouses must be able to fulfill a high volume of read requests for very large data sets with good performance. Database structure and storage layout are critical. Expertise in data warehouse technology and information management is important to the success of the project. Organizations should take advantage of the best practices that have been developed from past efforts.
But the end users don't use the data warehouse directly. They use applications and tools that are based on the software on their desktops. Any reporting capabilities that are part of the BI system and draw from the data warehouse must be integrated with the users' desktop environments. The success of the BI program depends greatly on not just how well information and the data warehouse are managed, but also on how easy it is for end users to take advantage of that information and how intuitive they find the tools.
The end-user perspective must be an integral part of the effort from the very beginning. By allowing the end users to shape the solution, the true business requirements are met more effectively. This results in higher adoption rates and subsequent broader use of the tools.
Be the first to comment on this post using the section below.