While scrap is not desirable, our production process will always produce a little. The scrap consists of primarily "drop" pieces left over from cutting stainless steel tube to length. We also lose a little during our press operations. We discovered that we had three different methods of reporting scrap, depending on the circumstances. This left our production crews trying to figure out how scrap was to be reported in any particular case. They guessed wrong more often than not and inventor)' was not being relieved properly. We reduced this to one way to report scrap. This becomes important during our assembly and shipping process.
Our studies found that our shop employees spent on average six percent of their day reporting labor. The accuracy of the labor we collected was suspect. We abandoned traditional labor reporting for a method that reports work at standard and collected actual labor by department. Based on what foremen reported was packed off from their departments, actual labor is spread based on product lines.
PRODUCTION AND SHIPPING
In addition to the problems we encountered with scrap and labor collection, we found that any resemblance between what a work order was issued for and what it finally closed for was purely coincidental. Whether or not the quantity reported was correct was open to debate. Cycle counts revealed that most adjustments could be offset by adjustments in the opposite direction at either the parent or child levels. Likewise, based on customer complaints, we learned that what we said we shipped didn't necessarily match what the customer actually received.
THE TEAM'S GOALS
Based on our study, the action team established the following goals:
1. Shop floor data collection had to be built into the
production process and entered directly in to the ERP
2. There had to be only one method of data entry in
any particular work center.
3. Since we where going to automate data collection,
producing and tracking paperwork was a duplication
of effort and should, therefore, be eliminated if hu
4. Final assembly must result in a product that could
have its identity verified at shipping.
5. Bill of material structures needed to be flattened to
reduce the amount of reporting required.
6. The shipping process must be changed so that prod
uct verification is built in.
We also realized that we needed a system of measures to ensure that we were meeting our goals. The following measures were determined not only to meet our immediate goals, but were also useful for the company as a whole.
1. Increase turns.
2. Decrease customer backlog (thus increasing sales).
3. Reduce product returns.
LEADING OFF WITH A WIN-FLATTENING BILL OF MATERIALS
In order to set the tone for our project, we felt we needed a quick and easy win to build confidence in our vision. Too many rimes in the past, improvement projects were launched that failed, either due to resistance out in the plant or apathy by managers. The attitude was, "Oh, it's just the change of the month. Don't worry, it'll go away." For our initial change we chose goal #5, reducing the number of levels in our bill of material structure. It was felt that this was primarily a clerical change and would increase turns in a hurry.
At the time, the average BOM had six levels. A bill for any particular end item generally had an hourglass shape. Many components made a few mid-level subcomponents (housings), which in turn could be used to make many finished goods. This led us to immediately decide that we only need three levels in our BOM—raw material, housings, and finished goods (a few OEM products went to two levels, raw material and finished goods).
We decided to take advantage of our ERP systems "phantom" process to make the initial change and future changes easier. By simply flagging the subcomponent levels that we were going to eliminate as a phantom, we could cause our work order process to skip these levels, showing the phantom's components instead. It also made future engineering changes easier because we could change a couple of phantom BOMs, rather than many stocking BOMs.
Within three weeks we had flagged our phantoms and used up the phantoms' inventory that was on hand. Using measure #1, turns increased almost 40 percent. Home run!
THE BIG PUSH-GOALS 1 TO 4
It became apparent that goals 1 to 4 were dependent on one another and needed to be accomplished simultaneously. Not wanting to reinvent the wheel, we benchmarked techniques used at two plants near us that were known for their advanced production techniques. Based on this research, it was decided that a modified version of the traditional kanban process would be used. We felt that by using these methods we could do away with all paperwork on the shop floor, with the exception of the daily schedule. We could have gotten rid of a printed schedule if we put more data terminals on the shop floor, but we haven't had the budget to do that yet.
While our ERP system did have shop floor reporting, if was found to be more detailed than necessary for our production process. It was also prone to user errors. For example, a work order for 10 units, with the 3 steps on the routing, could report IS units at step 1, 7 at step 2, and 11 at step 3. Obviously, not all of these can be correct, if any of them are. Some software modifications were needed to simplify and fail-safe the process.
We also realized that our production mix would involve two distinct reporting issues. The first involved serialized valves, which would require us to save, test, and serial number data. Everything else would comprise a second category, that we called "batch."
To Be Continued