The two prerequisites for any BI project: It has to be right. It has to be pretty.

Business Intelligence is an agent of change. It’s a tool to examine existing processes, highlight which aspects of them work, which do not, and point the way towards improvement. It is disruptive technology and can have major impact on individuals’ day-to-day work as well on the entire organization of course. And yet, with all this power and promise, the biggest challenge any BI project faces is around adoption. Why is that?

Well, for one, using a report, a dashboard or an application to examine information is not necessarily a step people HAVE to get through to complete their work. Over time, everyone establishes ways for gathering the information they need to perform their work. They learn to vet the information, trust it, and work around the effort it takes to produce it. A common response to the latest BI marvel just rolled out is “So what if the new report takes 1 minute to run, and it takes me 4 hours to gather the data I need otherwise? It’s time well spent, and at the end of the day, I trust my manual process more than the work of those folks in IT who have no idea what I’m doing anyway, since they never really talk to me.”

Change is hard, and efficiencies are not always desirable by everyone in the organization. In fact, many employees in large organizations rely on inefficiencies for their livelihood.

When you roll out a new financial software to manage your invoices and revenues, everyone in accounting starts using it immediately. They have no choice. When you roll out a new reporting capability, it’s not as simple. Everyone can continue doing their work without it. They just don’t know they can do it better.

What can you do about that? There is no easy or simple answer to the question of BI adoption. It is certainly one of the most difficult feats you will face. There are, however, two basic things you MUST do to assure adoption is possible. Consider the following two as the fundamental pre-requisites for your roll out plan to even get off the ground. You simply cannot succeed without both of these elements:

   1. It has to be right. 100% right. There is no room for error on the information you provide. If you roll out wrong information, and the people who are using the information perceive it as such, they will never trust it again, and will look for other ways to get the information they seek. If you’re not sure the information is trust worthy, delay your rollout until you get it right.


This table looks reasonable and is easy to read and understand. The use of colors, fonts, formats, sizes and alignment have all been applied to focus attention on information. However, looking at the Sales number for 2004 Q3, it is obvious a gross mistake was made. You do not need to be an accountant to understand that if the average sales per month is in the $2M-$4M range, a $134M number for the same quantity of product as other quarter just can’t be right. While this report may look nice, it is doomed, shunned as a silly mistake by technologists who don’t understand the first thing about the business.

   2. It has to look good. You have to employ intuitive navigation, controls, look and feel in the display of information. Someone who has never seen the report, but is familiar with the subject matter, should be able to take one brief look at it and immediately understand the story it is telling. The information has to be presented in an appealing and thoughtful way that will be easy and simple to understand.

The data in this report is correct and has been verified beyond any doubt. However, it is so unappealing that no one would actually be able to use it. The colors are distracting, the formatting is poor and misleading, and it radiates lack of thoughtfulness for the folks who should actually be using it.

The first prerequisite is obvious. There’s really no need to explain it much and I don’t think anyone would argue otherwise. The second prerequisite is less trivial, and in fact, many technologist ignore it all together. They focus all their efforts around the technology, and the accuracy, and at the end produce a very weak presentation of their many months of very hard work on things that no one would ever be able to see, touch or understand.

It’s also not a tradeoff between the two. They are both equally important. There is simply no way to succeed in a BI project without both the data accuracy and the information design to be considered together as the two fundamentals that must be in place before go-live.

This entry was posted in BI At Large and tagged , , . Bookmark the permalink.