Three Steps to Disaster

<< Click to Display Table of Contents >>

Navigation:  Editorials >

Three Steps to Disaster

Peter Vogel            
 
I have this theory about the life cycle of a lot of computer systems. I call it Peter's Three Step Life Cycle (originality isn't my strong suit).
 
In the first step (Acceptance and Delight), some group of users is given a computer system. They're so excited to have this package (and so pleased that someone has finally spent some time and money on them) that they don't notice that the system doesn't actually match the way the business works.
 
Let me make a distinction here between "the way the business works" and "the way we do business." In the second situation, the client is talking about some mismatch between the company procedures and the way a computer system demands the work be done. The simplest solution is to change the company procedure. Fortunately, most people don't realize that the procedure can be changed and, instead, call in someone like me to build a new system for them. They're wonderful people who keep me employed.
 
However, there are certain realities that are beyond company procedures. I describe them as "the way the business works." Any application must match those needs in order to succeed. Your immediate reaction might be, "How could someone build a system that does meet the fundamental needs of the business?"
 
As an example, let me tell you about one of my clients. They had this wonderful system that allowed them to forecast how much product would be wanted by each customer in each month. The idea was that, at the start of each month, the production department could match inventory against forecast and determine how much product should be made. Unfortunately, this system didn't allow you to specify what kind of packaging the product was in. So, while the company might have a forecast for 500 tons of product X and actually have 500 tons of product X in stock, it didn't help them plan production. The customer would need the product in cardboard containers, and the stuff in inventory would be in metal containers, or on pallets, or in some inappropriate form. Effectively, the system forecasted products the company didn't stock.
 
It gets better. The system would tell these users that the stock was needed by, for instance, Customer 12. However, Customer 12 isn't one customer. Customer 12 has a plant on the East Coast and another on the West Coast. The plant on the East Coast wanted its material on pallets, and the plant on the West Coast wanted the material in boxes. Since the system lumped all of the demand from these two plants under the heading "Company 12," it effectively forecasted demand for a customer that didn't actually exist.
 
Since they were in Step 1 of Peter's Life Cycle, the users were so pleased to have any support at all that they actually started to use the system. They then entered Step 2 of Peter's Life Cycle: Workarounds and Disgruntlement. In this phase, the users start to develop manual systems to compensate for the deficiencies of the computerized system. In my example, this meant two people for two weeks at the start of each month, decoding the forecast system output. These two clerks would convert the demand for non-existent products to fictitious customers into demand for real product for actual customers. Since the results of their work were on paper instead of the computer, they would then match the results against inventory by hand.
 
This leads to the third phase, which I call Termination and Disgust. By the end the third phase, the company will have two systems: the official, computerized system and the unofficial, manual system. The work is actually done in the manual system and only recorded in the official system. In fact, it now takes longer to do the work than if the computerized system didn't exist. The users recognize this, and the system that they once welcomed is now despised.
 
This is where I came in and, at the department's request, put together a proposal for a simple Access application to handle forecasting of real product to real customers. The department's proposal was turned down by the information systems group.
 
After all, they said, we already have a forecasting system.