Stop Throwing Software at Business Intelligence

Michael Vizard

Any time IT organizations are confronted with a new challenge, the conditioned response is to throw software and hardware at the problem. In most cases that usually helps, but when it comes to business intelligence it more often than not only serves to exacerbate fundamental management problems.

That's why Ambuj Goyal, IBM's General Manager for Business Analytics and Optimization, is pushing customers to engage with IBM's 4,000 BI consultants on how to create an effective methodology for managing information before they deploy BI software. All too often, we see BI projects fail because the data being analyzed isn't complete enough to be relevant. The end result is that users lose confidence in the data and fall back on intuition and the information that is readily accessible to them in a spreadsheet.

Speaking at an event in New York City, Goyal says that at this stage business intelligence should be less about IT and more about creating an intelligent business. That means first documenting and understanding all the factors that make up a business process before deploying any BI software. Once a company figures out how to overcome its internal divisions in a way that brings all known relevant information together, then the organization can start to derive some real value from BI software.

IBM has released a "Breaking Away with Business Analytics and Optimization" survey of 400 business executives that finds that top-performing companies are 15 times more likely to apply analytics to strategic decisions than their under-performing peers. You can debate the definition of top and under-performing all day long, but clearly applying BI software randomly to flawed business processes isn't going to help anybody.

IBM is also trying to create greater appreciation for the value of BI software by partnering with academic institutions such as Fordham University in New York City to create a business analytic curriculum. That's a much-needed step because not only do we need more IT people trained in BI, we need more business analysts trained in how to use the software. In fact, a separate CIO study conducted by IBM found that 83 percent of the CIOs surveyed identified business analytics as their current top priority. Right now, however, there is a shortage of both business professionals schooled in analytics and IT people skilled enough to deploy it in a meaningful way.

Of course, analytics are never going to replace the human factor when it comes to making business decisions. But most of our intuitive guesswork about business decisions today is based on incomplete data. With advances in using software to create a more intelligent business, we should be able to considerably narrow the gap between what is known versus unknown about the business in order to make more informed decisions.

Add Comment      Leave a comment on this blog post
Dec 10, 2009 4:00 AM Cathy Cathy  says:

Truer words were never spoken!

You've got to understand the business before you engage in any meaningful analysis...this is preaching to the choir.

Against that we have short-term, need-it-yesterday, next-quarter-results time frame for ROI.

No wonder nothing of substance gets done.  Too much upfront planning and understanding required.

Oct 27, 2010 4:53 AM Geoff Bazira Geoff Bazira  says:

It is important to understand your business vision, strategy, processes, and roadmap before implementing any type of meaningful reporting. There is always a temptation to report on data that exists (already being collected somehow) without a clear understanding of what type of business decisions need to be made with it and how it truly benefits your customers. It is the path of least resistance.  This approach leads to a lot of waste (development resources, operational innefficiency, and possibly competitive advantage).

The flip side to this is that no reporting happens (or too much time is wasted in planning) because you haven't yet figured out how to measure what you really care about. Same result as above, this results operational innefficiency and possibly competitive advantage.

Not everything that can be counted counts and not everything that counts can be counted - Albert Einstein.

I believe that the agile and pragmatic approach would require the following:

1/ Acknowledge the critical challenge of having "timely and meaningful" data. There is no quick fix - and so a flexible and balanced approach is needed.

2/ Without wasting time, start by identifying very simple and easily collected or estimated KPIs that meaningful decisions can be based on.  Some considerations may include:

a) Find KPIs that maintain a line of sight to your customer needs (as this avoids siloed development).

b) Find KPIs that reward "problem surfacing" and not "penalties" for employees.

c) Consider a mix of leading and lagging indicators for a balanced view.

d) Identify the degree of accuracy and risk associated with the data source and try to improve these overtime and be transparent about it.

3/ Plan to allow your KPIs to evolve as needed (as your data sources get better and as you learn more about what needs to be measured).

4/ Plan a BI framework that supports continuous improvement of KPIs. i.e. not too cluncky that minor changes to what is reported on requires months to change. This will dictate your choice of tools and architecture.

5/ Build a culture around "problem surfacing" instead of objectives that "penalize" failure to meet targets.  When penalties are associated with KPIs (e.g. system availability targets), employees (perhaps subconsiously) tend to highlight the best results, at the expense of collecting meaningful results that help their business to improve.


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.