MRO Inventory
When you get your start with MRO inventory on an underway submarine, you realize quickly how important it is to know exactly what it is that you have.
Later, when Minute Maid installed its first CMMS (computerized maintenance management software), I had a chance to directly improve the storeroom operation. We made the accuracy of the inventory paramount. This enabled us to activate the automatic ordering of stock items. It was still a manual operation compared to what we can do now, but to run a report overnight, parse it into separate pages by vendor, and fax each order behind a prepared cover sheet was a huge improvement over the incessant telephone calls previously required.
I made a game of it when vendors or others visited. We would look up some part they would name at random. Reading the location and quantity off the screen, we would walk to the designated bin and count the items in it. The storeroom folks were praised every time it all matched up!
Our maintenance group always scrambled to bail out production units that ran out of cleaning supplies, hair nets, bump caps or whatever. So my boss offered to set up a "centralized storeroom" for these items. I cobbled together an off-the-shelf accounting system with our CMMS to create a charge-card kind of system for our various production units. They could draw whatever they needed at any time and we could accumulate their charges for the end of the month. And using the automatic reorder, we never ran out!
At Ocean Spray Cranberries I continued these kinds of improvements. By regularly bidding out our commodity groups (at least annually) we were able to reduce our purchase costs by nearly 20%, even when we stayed with our original supplier. Because we could give prospective vendors accurate usage information, they could bid with a sharp pencil.
We also scrutinized usage, compared to quantity on hand. A huge number of MRO items can be stocked with only 1 or 2 items, so long as your counts are kept accurate and you reorder religiously.
To improve accuracy, I wrote interface routines from a barcode scanner to our CMMS, and instituted daily cycle counts of a few hundred items in our inventory. Every item was counted on a 6 month, 12 month or 2 year cycle, depending on its criticality.
Inventory accuracy rose to over 98% (based on value). Inventory value declined by nearly 20%, even while stocking a larger number of SKUs.
With Nestle I learned more sophisticated formulae to calculate Economical Order Quantity (EOQ). But I also learned that formula is better suited for regular purchases such as production items, and that for irregular maintenance purchases, you can still stock a single unit of may SKUs (unless you use the item in pairs).
We were also able to improve greatly on the automatic reordering. With modern enterprise systems and e-commerce, we were able to integrate our CMMS with both e-commerce and financial systems. Approved orders are beamed automatically to our Electronic Data Interchange (EDI) vendors. For those not EDI-enabled, they still get a fax, but this fax is automatically generated and sent from our system.
Nestle also began a common parts nomenclature project (but did not finish it). We could easily aggregate spend from multiple plants by commodity group. Even with all our data residing together, however, we could not aggregate data from multiple plants for a given part without a Standard Modifier Dictionary. I think the argument is continuing to this day as to just how much resolution is beneficial.
In the end, you still want what we wanted on the submarine: to know what you need and to know what you have. If our mechanics will "charge out" the parts they use, and our storeroom quickly replenishes it, then the next time the mechanic needs a part, it will be there. No excess inventory. No delay in repair.
