Information Systems
by
Arthur M. Schneiderman
I placed IS at the center of my QIP symbol to stress the role of data as its very foundation. QIP represents the application of the scientific methodology to process management and improvement; and data is the foundation of science and engineering (applied science) as we all know.
But
there's a lot of data out there and to make it useful, we need to first
convert it to information. Information tells us about patterns in the
data and allows us to make deductions or inferences about their causes. They in turn provide the basis for our new knowledge which
gives us the confidence to take action. The Knowledge Pyramid captures
the essential objective of information systems.
Nowadays, we tend to equate IS with automation using computers. Data warehouses, data mining, and data bases have become synonymous with information systems.
All too often, information is generated for its own sake rather than for its role in improving the effectiveness and/or efficiency of actions. In my view, IS should be based on tops-down use of the Knowledge Pyramid : What action do we need to take (e.g. improve a given process)? What knowledge would improve the quality of that action (root causes of defects)? What information is needed (Pareto diagram of defects). What data is required (record of individual defects)?
But today, this IS process is frequently run in reverse: The computer has tons of data (so much that it's earned the name "data warehouse") so lets analyze it and publish a report of lots of different views of the data. Someone will find it useful, I'm sure. And, by the way, I just saw a neat new program that will do this all for us. In practice, this expensive approach (referred to as product-out vs. market-in) rarely provides action agents with useful knowledge.
The value of IS in process improvement is maximized when backward-chaining (action to data) rather than forward-chaining (data to action) is employed. Backward chaining eliminates the wasteful generation of information which is not useful for the issue at hand.