Garner, the famous technology research group have once stated that 70%-80% BI projects fail and ended up in a trashcan. Even in my personal experience, I have seen how some of the DW/BI projects ended up in nowhere, and the organizations ended up in wasting lots of money. Since 70%-80% fail rate is enormous, I thought of sharing my experiences which would help you to make your BI solution a success.
We can do it by our OWN… Can’t We?
Most of the IT departments of the organizations believe that they can implement a BI solution by their own even though they do not have required skill set and the competency in that area. They believe that to implement a DW solution, you need nothing but some knowledge about what’s facts and dimensions are, and having some idea about a BI tool is more than enough to implement a BI solution to their organization. Therefore, what they do is they send some or all their department employees, who are main from software development, DBA, or software maintenance background, to a DW/BI training or a workshop and then expect them to build a DW/BI solution for that organization. In my opinion, what you should understand is that implementing a DW/BI solution for an organization require good understanding about the core concepts of DW (Star Schema, Snowflake schema, Conformed dimensions, etc..) as well as deep knowledge in BI tools such as ETL tools, OLAP tools and reporting tools. If you are not 100% sure about your IT departments competency level in DW/BI, then best option would be to go for reputed DW/BI solution provider.
It’s not a magic wand, have realistic expectations
Another major issue in DW/BI project is that organizations expect all most all their reporting requirements, dashboard requirements as well as forecasting models from the newly started DW/BI project at once. Implementing a total reporting framework or analytical framework cannot be done within 3 months,6 Months or even in one year. On the other hand, it can be observed that the end users fail to distinguish between a report solution and an analytical solution, and therefore, try to incorporate any information which are currently available in a tabular format in to the BI solution. The real issue in such cases is that the new DW/BI solution have to focus more in whatever the reports already with the end user rather than providing valuable insight on the business. Consequently, at the end of the project, the organization might not find any value addition from the newly implemented DW/BI solution. Personally, I believe that the IT department of the organization must educate the end users about the difference between a reporting framework and an analytical solution before they give their requirements to a DW/BI solution so that they can obtain maximum out of the solution. In conclusion, I believe that it is necessary to have realistic milestones as well as meaningful requirements when implementing a DW/BI solution to your organization.
Eat piece by piece, if not you won’t be able to swallow it
Some organization make the mistake of trying to implementing the whole DW/BI solution in one go. Which means that they start the project and after finishing all the development only, they expect to see the outcome of the project. This method is called waterfall method in software engineering terminology. However, this approach has certain issues. The main issue is that until the whole development phase of the project is done, the end user has no way of knowing how the output of the project looks like. Therefore, they end user’s inputs are not considered until UAT period of the project. What I have seen in such scenarios is that, since there is no way to collect end user’s feedback until the last moment, the output of the project does not tally with the end user’s requirements. Therefore, the best approach for BI/DW solution is to follow agile methodology where you do the development phase by phase and give releases to end user regular base. In that way, you can collect end user’s feedbacks rather than surprising him at the very end of the project which might cause lots of problems at the very end of the project timeline.
New is good, But I like it old way
Another reason for DW/BI solutions to fail is end users of the organization are reluctant to use the newly implemented solution. This can be due to different reasons. The first reason I have seen is, although top level of the organization requires their employees to use new BI solution, those employees are already well trained and are in the comfort zone in their existing reporting/ analytical methods. From their point of view, new systems will always have a learning curve which they don’t want to experience. On the other hand, even though employees are open to new a new BI solution, the complexity of the new system or lack of proper user-training about the new system would make them to go back to their old system. Therefore, if you are going to implement a BI solution to your company, make sure it’s simple and your employees are well trained to use newly implemented BI system. Apart from that, the top management need to push their employee use the new system for all the possible analytic requirements.
Remember these are not the only reasons to a DW/BI solution to fail and there can be more reasons which are not mentioned in here. Therefore, if you can comment whatever the root causes for failed DW/BI projects, it will be helpful for someone else when implementing a BI solution to their organization.