Something that tends not to be well publicised in the media is the fact that Business Analytics (or Business Intelligence) projects can fail. In most cases the solution goes live on time, within budget, but ends up being completely ignored by the user community.
The reason is often simple: people have just been too ambitious with their objectives. They have attempted to implement too many solutions too fast, and the business users were not prepared to absorb all the novelty. Often the business users were the very people who pushed for the solution, but somehow the organisation was not prepared to do the jump so quickly and, at the end, they implemented something that does not perform, does not meet the requirements, or is just not used, even though it is exactly what has been specified.
Unlike OLTP (Online Transaction Processing) solutions, which the users must use to do their job, Business Analytics solutions are decision support tools, that the users choose whether to use or not.
The risk of failure tends to be higher with newer or less frequently implemented solutions. Standard Business Content‐based solutions are more likely to succeed, whilst Business Objects® , Mobile BI®, HANA® and even the not‐so‐new Integrated Planning® and BPC® will carry higher risk.
An area that must be approached with extreme caution is the integration between BW and BusinessObjects®. SAP has made great advances in integrating the tools in BI version 4.0, but there can still be bugs, areas that the consultants do not understand, and several cases of people who have bought the software and have not implemented it.
This is the bad news: failures can happen, and can happen with any well‐intentioned IT department. The good news is that there is a way to increase the chances of success.
First and foremost, it is important to understand that business Analytics is not a one‐off project, with a defined start and end. It is a journey, a programme, which starts with the implementation of the first solution, and continues with each new project, each building on the success of the previous one.
The best way to succeed is to implement discrete solutions, work hard on removing all potential bugs, get them well‐established and accepted by the user community, and only then expand the reach of the solution from there. It may appear slow at first, but considering the alternative (which could be complete apathy from the user community), the "small parcels approach" builds trust with the business, and makes it easier to get funding for each new project.
Another good practice in achieving success of Business Analytics solutions is to treat them as separate from the ECC implementation. It is very common for the BW/BOBJ project to be a module of a major implementation or upgrade project, with centrally defined time‐line and phases. This approach puts unnecessary pressure in the Business Analytics project, which has different timing, development and testing cycles. Sometimes, it only makes sense to test some analytics solutions once the base system is already live.
In a nutshell: treat Business Analytics as a separate work programme, make the objective of each project simple and achievable, and, if possible, a separate project from any major ECC project. One more piece of advice: focus on the user, the user experience, make them part of the project (more so than you would in a typical OLTP project) and your Business Analytics programme will live long and deliver real benefits to the business.