Nuance is a challenge for journalists. It's tough to offer a nuanced view of dense topics -- and technology topics are among the worst -- without boring the reader to tears. In a recent story, I tried to bring out the (nuanced!) concept of companies starting small with business intelligence initiatives before applying BI more broadly across their business. Several of my sources, however, felt I was pitting one approach against the other.

Analyst Nigel Pendse, author of the annual BI Survey, feels it appears in the story as if he and Information Builders Chief Marketing Officer Michael Corcoran were "in violent disagreement" over whether BI should be widely deployed throughout an organization or just to key users. As Pendse points out in an e-mail, he and Corcoran "agree on most things." While Information Builders applications are sometimes widely deployed throughout organizations, writes Pendse, end users don't necessarily see them as BI and generally aren't directly involved in implementing the apps. The two examples of Information Builders clients I cite in the story illustrate this well, I think.

Information Builders' appproach differs from other vendors that focus on workers who use BI applications in a more ad hoc manner and are "well aware that they're using a BI tool." As Pendse told me in an earlier interview, many of those vendors are simply trying to sell more BI licenses.

Another of my sources, independent BI consultant Peter Thomas, takes issue with Pendse's suggestion for undertaking small, tactical BI intiiatives rather than enterprisewide ones. Beginning with small projects is the most logical approach, says Pendse, in light of the fact that the median age of BI applications at companies he surveys is less than two-and-a-half years. In a follow-up post on his BI blog, Thomas zeroes in on that figure, thinking it seems inadequate for a BI project. He writes:

For a BI project a lifetime of two-and-a-half years seems extraordinarily short, given the time and effort that needs to be devoted to delivering good BI.

Pendse's survey may not differentiate between small, tactical BI initiatives, failed larger ones and successful enterprise BI implementations, in which case the numbers may work out right, writes Thomas. Tactical projects are intended to have a short life and bad projects tend to get killed quickly, so those categories could exert downward pressure on the median.

He doesn't advocate "big" BI over "small" BI. Rather, "for me what counts is what works." He writes:

Tactical BI initiatives can be very beneficial in their own right, as well as being indispensible to the successful conduct of larger BI projects. However, it is my firm belief that tactical BI works best when it is part of a strategic framework.

Again, I'm not trying to create controversy where there isn't any. Thomas obviously feels tactical BI projects have merit. I suspect Pendse will concede that enterprisewide BI projects can create tremendous value, even though they'll inevitably be more challenging and cost more to boot.

One area on which I know the men (and my other sources) agree is the importance of involving users in BI iniatives. They all mentioned it, though those remarks failed to make the final cut in the interest of keeping the article reasonably concise. I'm now feeling some editorial remorse, since I think it's such an important best practice, one that I've featured a number of times in this blog, most recently in December. I like this quote from John Colbert, vice president of research and analytics for BI advisory company BPM Partners:

Don’t underestimate end users in the process, since they will drive benefits. Users will keep doing spreadsheets if they don’t see enough benefit (from BI). They should be pulling the project of your hands, versus you throwing it over the wall to them.