I was the IT person for a manufacturing company during the Y2K switch. One of the business managers' favorite mantras was that everything had to ultimately be "flexible," and any rule could be broken if the situation demanded it. Where inventory was concerned, they were constantly robbing Peter to ship to Paul, and then wondering why their shipment to Peter had so many back-ordered items. This became my problem when I was tasked with creating inventory tracking software.
We had just chosen an ERP to replace some non-Y2k-compliant systems. The ERP implementation was slated to take two years. However, the inventory problems were serious enough that management wanted me to replace our existing workflow (paper-based data that was keyed into our legacy manufacturing system) with an electronic system right away. They wanted inventory movement handled by RF barcode scanners, and I was told to come up with a solution. I went with the ERP's preferred data collection partner.
[ Have you encountered the technology pro's six greatest enemies? Send your memorable tale to email@example.com. If we publish it, we'll send you a $50 American Express gift cheque. | Get a new tech tale delivered to your inbox every week in InfoWorld's Off the Record newsletter. ]
I designed the system so the interface on the RF scanners would be the same regardless of the back end. Until the ERP arrived, I had to interface the inventory movement to the legacy manufacturing system, hosted on a nonstandard obsolete platform that had been deprecated by the manufacturer.
The system worked well enough: Our inventory was the most accurate it had ever been (which still isn't saying much). We gave the warehouse staff a thorough tutorial on how to use the RF scanner and track the inventory. However, we were still plagued with inventory losses.
When the new data collection system had been in place for a little over a year, quarterly financials showed that we were on track to lose over $8 million. Focus quickly turned to inventory problems and my inventory movement system. To combat the ever-increasing scrutiny, I designed a test of the system to show once and for all that the inventory system worked.
The plan was that we would run a program that would "move" every item of inventory to a dummy location called "Foo." Then, on a Sunday when the factory was down, an army of people using the inventory guns would manually count and scan everything from location Foo back to their true location.
The first benefit was that our inventory would then be correct -- all errors would be in Foo. The second benefit was that we would be running a huge amount of data through the system, proving its stability or exposing any weaknesses.