Saturday, August 11, 2007

The BI/EPM Revolution (..needed, please!)

Frank Buytendijk makes a good case for why generic BI/EPM frameworks are most probably pretty useless in practice.

What puzzles me deeply is that so much of the focus of BI remains concerned with the technology infrastructure and dashboard bells and whistles. This is seen most clearly when a new enterprise application or business process is being introduced, and you hear the almost dismissive assertion made that "it will integrate with you DWH", or that it will "plug in to your management dashboard". And "we'll scope for, let's say, 20 reports to be built".

I think we are all missing the point.

I'm going to suggest that on the whole, business is still struggling with two symptoms of IT growing pains ...
  • The IT profession has lost sight of the fact that reports, dashbords and the like are not end products in their own right. They typically have no intrinsic value. The value lies purely in the decisions and actions that may be taken as a consequence (the exception perhaps being for reports that are required by an external party such as shareholders, government or regulatory authority).

  • Equally, I think business management is still adapting to a world where information is plentiful. Even when fact-based decision making is practiced, it is all too common to find it be based on spreadsheets (which may or may not have had their original source data pulled from an enterprise system).

Every time I hear the ERP vendors boasting about enterprise adoption, I cynically imagine a Microsoft advertisement that could (probably truthfully) claim
"99.9% of Fortune 500 companies run their business on Excel!!"

I think its easy to understand how we got to this position. I remember working for a reinforcing mesh ("fabric") manufacturing company in the late 80's. We had factories filled with fabric machines like the one in the picture. They weren't networked or anything. The only information that the production manager had to work with were the production sheets filled out by the operator each shift, and the maintenance manager had to send his technicians around to check out the machines periodically.

We were working in an environment where information was scarce, yet the business imperitives were very much the same as today: production manager was concerned with productivity and profitability; and the maintenance manager cared about the maintenance costs and machine performance (netting to the "TCO" of the plant). We managed through a combination of MWA (Management by Walking Around, aka gemba / genchi genbutsu), MGF (Management by Gut Feel! .. some would say experience;-), and also selected special projects to capture data related to problem hypotheses [what I was most involved in].

But the world has turned in the last 20 years. Today, most organisations have more information sitting in various databases and enterprise systems than they know what to do with. I believe the challenge these days is firstly to make that information accessible, and then - most critically - to ensure that our management practices can make effective use of that information.


We actually have great technology these days for cracking that first challenge - making information accessible and actionable. I like to think in terms of three tiers (management, operational, infrastructure), as in the diagram.

Unfortunately, that's often as far as our thinking (and implementations) go.

However as Frank points out in his post, we need to address the strategic framework for BI. How BI impacts, changes or even transforms the business.

For me, the implications are clear for all decision-makers (which may or may not just mean "management")...
  • In the past we have excelled at managing in an information-poor environment. Now we need to go back to school, unlearn some of our assumptions and practices, and start exploiting the information that (should be) at our fingertips.

  • This means a renewed relevance of management science, particularly quality management (e.g. 6 Sigma) and key concepts such as kaizen/continuous improvement.

  • In true "Flat Earth" style - if you don't, someone else will and then you are history, mate.

And for IT, I think a few things to consider...
  • First, get real and get connected to your business. You are not delivering 20 reports, you are delivering information that will hopefully help your business users may really good decisions, impact the bottom line and keep your job safe.

  • Avoid the temptation to dumb-down and de-scope BI and reporting. It could be the most business-critical aspect of the whole project ... Promote monitoring and management to first-order use cases.

  • Consider all tiers of monitoring and management: from corporate performance, to operational management, to infrastructure.

So maybe this is an aspect of what Enterprise 2.0 is really all about?

2 comments:

Tim said...

Hey Paul
Great post - I think in your comment earlier on in the post - "What puzzles me deeply is that so much of the focus of BI remains concerned with the technology infrastructure and dashboard bells and whistles" ties in nicely with what you say later, "more information sitting in various databases and enterprise systems than they know what to do with"
Thats the big hurdle for many companies, how do I get all that info into a single 'source of truth' for my users. Hence the obsession with infrastructure. I think once that hurdle is past then IT can start to look at what the end users need but its a big stumbling block in my experience.
As for not dumbing down - Iam whole heartedly with you on that one.
Tim

Paul said...

Thanks for chipping in Tim!

I do agree. Wrestling a decent data bank into useful shape is a major issue in practice.

From the infrastructure perspective, we can push that forward.

I think I was campaigning to have this complemented by a stronger 'pull' factor from the decision-makers and business users. Those who really know what they need to know to do their jobs, and are not being shy about demanding the necessary infrastructure and support!