We do BI projects because of the high-quality deliverables end-users derive from them, writes Erwin Bisschops, senior consultant at Harvey Jones. Of course, we as BI professionals like to do them as well because of the intellectual challenges they bring.
But without successful outcomes, no BI projects.
Unfortunately, one of the disadvantages of BI environments is that, when push comes to shove, they're not as important as the transactional systems from which they derive their data. So, an important question is: “How can we increase the popularity of BI environments?” The answer is as straight forward as the question: by choosing the right deliverables.
Many organisations don't primarily use BI tools and concepts for their analytical capabilities, but for the following reasons:
* BI is user-friendly;
* BI is fast (well, it should be);
* There is no need to collate information from various sources, because with a bit of luck that is already done in the underlying data warehouse; and
* It produces posh reports on demand in a format the end-user desires at that moment, without having to ask someone else to do it for them.
In other words, BI fills a big gap because it can deliver what operational systems can't. The number of Excel spreadsheets that float around in organisations is an indication of so-called “shadow IT”. This term is used because the reports are created by the end-user community and are not maintained by the formal IT function.
If many of these nifty Excel sheets float around, the formal, transactional systems in the organisation are perceived as ineffective.
The use of proper BI tools can also be an indication of the shadow IT phenomenon. I've seen SAP Business Warehouse, not one of the most impressive and cost-effective BI environments, being used purely as an operational reporting engine because people can't easily print using SAP R/3, the transactional environment. Is this a good thing or a bad thing?
The answer to the question “what are the right deliverables?” is operational BI. Just listen to what end-users really want: something simple that simplifies their day-to-day activities and saves them time that can be spent in a better way. Operational BI has the following characteristics:
* There is no focus on long-term trend analysis or typical BI reporting such as monthly YTD figures, combined with previous year monthly YTD figures and growth percentages between the two; and
* The focus is more on daily information, detailed reporting used for day-to-day business monitoring and reporting.
The big advantage of proper BI (apart from all the so-called quick-click tools that process everything in memory) is the integration of various source systems. In other words, yes, it is possible to compare productivity stats of a team (sourced from the production system) to the sick leave percentages of the same team (source from the HR system).
Why? Because the employee numbers in both systems have been properly mapped to each other so that the systems are truly integrated. This delivers much needed information to the organisation about capacity problems and its effects on the workforce! So, any kind of reporting that traditionally happens in Excel, where various data sources need to be cleaned up and integrated, benefits enormously from operational BI.
Pushing the analytical side of BI into the organisation will only be successful when the end-user community is ready for it. Even though BI has been around for a while, it is still difficult for the average end-user to break away from standard reporting and be able to sit behind the steering wheel, instead of just consuming information other people prepared for him. The group of people that will really go for analytics is not more than 10% of the organisation. The other 90% wants standard reports.
So it's time to loosen up a bit when it comes to doing “traditional” BI projects and just accept that sexy BI tools delivering operational BI can pave the way to a better embedding of the BI environment in day-to-day business activities. And by delivering what the end-users really want, the way is paved to reap the benefits in one or two years of proper analytics, because people will be more used to the potential of a BI environment.