Earlier this week, I had the privilege of delivering a guest lecture on Business Intelligence for Professor Malone’s graduate “IT Essentials” class at MIT Sloan. The experience brought back fond memories from when I taught back in the early 1980s.
Most in the class already had some real-world business experience. As a result, there were a number of excellent and insightful questions.
Here are a few of the questions (paraphrased), with my responses:
Q: How do you create and sustain an “information democracy” in a startup venture?
A: First, you need to plan as though your “startup” will someday become a large enterprise. Decisions made today will either help or haunt you well into the future. So, with that in mind, at the top of the list are transparency and the open sharing of knowledge and insight. By nature humans are hoarders and prefer to withhold information for their own benefit. That’s why it’s nearly impossible to change the culture of a large organization where “in-transparency” has been the norm. Of course, it’s easy to put transparency in the mission statement and ignore it. However, to be truly successful, one must “live” transparency. This means encouraging and incenting the sharing of information – especially if it exposes problems!
It’s also useful to look at the dysfunction surrounding BI implementations in large organizations today: disparate applications and data, conflicting business semantics, multiple master files, and a multitude of user tools (including spreadsheets!) – have created pockets of BI automation and what I call information “myopia”. These problems can be minimized if you start out with the notion of an enterprise data model and build (or implement) enterprise applications that are fully integrated and which are equally well designed to analyze the data which they process and store. Although data warehousing will still be necessary, its creation and maintenance will be substantially easier.
Q; How do you balance data quality shortcomings and urgent user demands for access to information?
A: Data quality and integration are areas where getting it “right” is critical. It’s expensive, takes lots of time and energy (and skill) - but is hard to readily demonstrate value to business management. In contrast, users are easily excited by fancy BI tools with cool visualizations. And, with innumerable tools available for purchase and download over the internet – users are buying them and loading them with report extracts, spreadsheet data, etc. and making decisions based on incomplete or erroneous information.
The solution is to balance urgent end user requests with needed data quality programs. This means doing some things that are expedient: delivering user applications quickly without perfect data quality. As a part of this, users must understand the limitations and that the reliability and utility of these applications will eventually improve as a result of strategic data quality programs. One of the best ways to achieve this is through the creation of a BI competency center –whose charter it is to document and implement best practices for BI.
Q: Can organizations use the tools they have or do they have to buy new ones to succeed with BI?
A: It turns that that success with BI has much less to do with the tools than the people using them. I’ve seen some BI successes that employed very modest technology and some colossal failures that had all the technological bells and whistles. That’s why much of my book focuses upon the human and organizational issues that determine the success of BI initiatives. These include management vision and commitment, organizational alignment, culture, and skills. Although buying “yet-another-tool” is easier than solving these problems, it ultimately makes things worse.
Check out my website for details on where I'll be speaking, presentation abstracts, articles, my book and more!