NW Business Intelligence

For the past half dozen years I have spent the majority of my time helping companies get down that last mile or two of a BI project. While not always, it was often the case that I was there because IT had found the burden of delivering a BI solution was more than they bargained for. It may be they underestimated the skills required; maybe they were oversold by the vendor on the ease of implementation; often they were just slammed with other vital work. One advantage I always had in these situations was being allowed to work mostly outside the boundaries of the typical IT processes. Ah yes, the so called “cowboys” had arrived is what I’m sure the IT people thought. However, they had no choice because their customers demanded a solution be delivered now! And there in lies my point. We all exist to serve those business users, our real customers. In BI that means delivering solutions in a way that is often very different than other applications IT typically delivers.

I just read a Forrester BI trend report that confirms what I have been seeing and thinking for a while now. This trend towards centralizing BI into IT is not paying off. I understand why it has happened. We needed the buy in from upper management, the sponsors, data governance, master data management and all that. No doubt about it. That is all vital but it is a Fuastian deal. At least as far as our real customers are concerned because along with becoming a part of the centralized IT structure inevitably comes bureaucracy, layers of process ill suited to BI and we end up lost and bogged down unable to deliver on the promise. I’m sure somewhere a company has managed to make it work but overall it appears to be a growing problem. The Forrester report suggests we let IT hang on to data preparation and turn the rest back over to the business. Interesting. We need to do something different, that is for certain.

I don’t want to return to the “cowboy” days and throw all the hard earned lessons and best practices out the window but we need to realize that our customers move at the speed of their business. They are mad at us because we can’t keep pace. We need to be brutally honest with ourselves and answer the question “Is everything I’m doing adding value for my customer and am I doing it in a way that meets ALL their needs?”. That means getting them solutions quickly and standing side by side with them so we know what it feels like when we fail to deliver.

A couple of years ago I had figured Microsoft was killing their BI product efforts. Word on the street was that key people and products were being slayed left and right. Well that may have been true but it seems to have been part of a plan to attack the market from a different angle. I’m not a sales guy or even a marketing whiz but I am good with technology and from what I have seen in 2008 R2, Microsoft BI is back! OK, maybe I’m late to the party but I finally arrived and it looks pretty good. I’ve been reading a bunch of “what’s new in R2″ articles and blogs and finding them interesting so I ordered up my copy of 2008 R2 a few weeks ago.

The upgrade was pretty painless. I had some issues with my laptop installation and setup of SSRS but thanks to some help from a very smart friend and the removal of AVG Free antivirus software things are humming along nicely. Report Builder is looking really good and they have done a lot in the area of ease of use based on my recollection of 2005. I know Cognos Report Studio very well and I have to say Cognos, get cracking on some new updates. This stuff looks quite good at a price point that a lot of companies are going to find hard to pass up. I’ve never been really thrilled with the charting features in Cognos so I want to see what I can do with the latest in Report Builder.

Next stop is PowerPivot then SSAS. SSAS being the bigger nut to crack but I want to see how it stacks up too. I’ve heard there are even more cool things on the horizon so it may be time to fully spool up on Microsoft BI. After all, they are at the top of the Gartner BI Quadrant this year. Amazing.

Does it feel like sometimes the BI team does nothing more than deliver reports, updates to reports and more variations on the same reports? I have a hunch that is the case more often than a lot of us would like to admit. The promise of BI is always big when making the initial investment in software and hardware. That juicy suite of tools and studios are demonstrated as a tightly integrated whole capable of pumping out solutions that will turn your business into a well oiled machine. Executive management is sold if even half of it turns out to be true. But, initially at least, efforts almost always focus on reports, operational in most cases.

It’s understandable. These are the low hanging fruit. The data is easily captured and delivered to a data mart and the reports themselves are well defined and heavily used. A nice quick win to leverage a big investment. Everyone wins. However, what happens next is key I think because this is where organizations have to fight to hang on to their original vision. That quick initial win feed on itself but not always in a good way. You are now going to get more and more requests to build operational reports. Your demand is going to grow exponentially. Along with it comes production support burdens that are not trivial. You will be supporting a growing user base that is dependent on a system that must be operational during peak business periods. Unless you have planned for this burden you will find your team now consumed and converted to an operational reporting development and support team! And that brings me back to my opening sentence. It’s now going to be very hard to go back to management and ask for more resources to deliver on the original big picture solution they were so excited about.

It’s the proverbial slippery slope. I think what has to happen is more BI leaders have to resist that urge to shine by delivering on operational reporting and push harder towards engaging on projects that deliver complete package solutions. That means finding projects that can use a wider range of components. One that has an analytical piece, a simple dashboard, some basic event notification based on business process and maybe some scheduled report delivery element. Obviously this will vary based on the tools you have to work with. The idea is to look for projects that will exercise the full suite and deliver a solution to a real business problem from end to end. That is where you will be contributing the most to the company’s success. Letting your BI team get stuck in the rut of operational reporting is a failure to deliver on the promise of BI.

A few months back I wrote about the aesthetics of information, or information visualization, and how it is a vital part of getting your reports (charts and presentations, etc.) to be more effective. I’m always on the look out for tools and articles on the topic so this posting is just a nod to a couple things I’ve seen and liked lately in the area of data visualization.

First, to Stephen Few at Perceptual Edge. I recieved an older white paper he had written for Cognos on the principals of viusal communication that led me to his web site, which of course is his business storefront. Of interest is a nice list of articles and other white papers here. He includes both articles of his own as well as from other authors on the topic. I like his style of writing and he seems to keep an open mind and avoid stating absolutes. I have not had a chance to read many of his articles but the few I have read lead me to think he is a real champion of data visualization.

My experience has been that most BI developers don’t understand the importance of visual design in the their work. The tendency is to brush it off as “eye wash” or leave it for the end users to bring up in user acceptance testing. I’m biased perhaps because my educational background in architecture included the visual as a key element in every “deliverable”. Communicating with the eye in that field is essential. I don’t see it any other way in BI when part of a deliverable includes something more than lists and cross tabs.

Lastly, a nod to a company getting some buzz lately is Lyzasoft. People are blogging about them and there are some lively discussions. I have not had a chance to check them out beyond getting a feel for where they sit in the broad spectrum of BI. Sounds like they sit somewhere between Tableau and QlikView. I’ll have to add them to my list of tools to try after my very positive experience with Tableau a few months back. From what I’ve read I don’t expect them to be that advanced in visualization though.