In the last 20 years I have looked at dozens (maybe hundreds) of different bits of software designed to support the production of Management Information. The interesting thing is that they all demonstrate well. The formula is invariably a simple database with fairly small amounts of data, a little dragging, dropping and clicking and – hey presto! – a report produced by the salesperson in moments.

The real world is more challenging than that. Underlying data structures can be complex and even messy. Years ago I was doing some work for a large retail bank. Their core banking system had a view of a customer who could have a number of accounts. Their Credit Card system on the other hand had no concept of a Customer. It was based on card numbers – the “customer” was just a name associated with a card. The management wanted a holistic view of their customers. The customers wanted all their dealings with the bank on one sheet of paper. The bank spent millions on a specialised Management Information system and even then only partially solved the problem.

Depending upon working practices, the data available may be incomplete or inaccurate. Hospital Patient Administration Systems (PAS) may contain accurate administrative data (who was admitted when for example) but the diagnosis and treatment data on them are in general woefully and notoriously inaccurate. This is a result of the way that the data are recorded. In one case that I came across a “URTI” (Upper Respiratory Tract Infection) had been coded as a “URT” (Urinary Tract) infection.

Given these messy features of the real world, Management Information System software has to make a tradeoffs. Is it going to be like a…

• …magnifying glass – easy for anyone to pick up and use, but restricted in the kinds of problems it can handle, or like a…

• …specialised microscope – hugely powerful, but expensive to buy and needing constant input from highly trained operators

There isn’t always an easy answer. Software powerful enough to answer all the questions that need answering may not be cheap to buy or easy to use. Software that can be used by anyone may not be up to tackling the hard questions.

Please let us know how well you think we are doing with this balance in Talis Decisions. Do comment on this blog or e-mail me at jlgh@talis.com.