Page 299 - How To Implement Lean Manufacturing

276 Cha pte r Se v e ntee n

Even though these data had huge variation, we
were assured that these production data were typical.
Point of Clarity Prior to
In this six-week period:
implementing a Lean initia-
tive, processes must show sta- • Six daily shipments had been missed.
tistical stability. This process • Production was well below the 36-unit target.
does not!
• The production rate had a huge day-to-day
variation.
• Weekend work (both Saturday and Sunday) was required to replenish inventory.

We Pareto-ize the Days of Low Production, We Find Two Problems
We reviewed production data to find the reasons for low line production. In this 30-day
period, there were nine days where actual production was ten or more units below
demand. (Each tray had 25 production items, with four trays per box, so a box con-
tained 100 items. This was their production jargon: boxes, which they called “produc-
tion units” or just “units” for short). In all nine cases, either the automatic welding
machine failed or the sensor, a high-cost component, had a stock out, or both occurred.
At any rate, their large problems were not quality problems, but availability problems
caused by these two items.

We Investigate Further
With this information in hand, we wanted to have another discussion with Miguel to
gain more insight. First, the welding machine availability was number one on our list of
items to correct. Unfortunately, the home office facilitator had characterized the prob-
lem as inadequate maintenance: specifically, the need to train a replacement for Jorge,
the welding technician, who had retired four months earlier. The plant was not allowed
to replace Jorge because they had been asked to reduce manpower. Miguel told us the
problem had nothing to do with Jorge, it was a capacity problem with the welder. It
would simply overheat at the new rate and the electrode and holder would fail, requir-
ing a shutdown to replace the parts and several hours to complete the setup, which
included alignment and testing. Work was under-
way to implement Single Minute Exchange of Dies
“Let’s not work hard to (SMED), quick changeover technology on the machine
get good at something which startup, but a practical solution was months away;
should not be done at all. ” however, Miguel did not believe that starting up
J. Keating faster was the issue. We did some more investigating
and found out he was right. The problem was poor
reliability.
As for the sensor stock outs, this sensor was a high-cost component comprising
44 percent of the total raw material cost of the product. The home office had imple-
mented strict inventory guidelines and aggressively reduced the inventory levels of all
components, including the sensor. After the ramp-up, the supplier could meet demand,
but only by working overtime. Since these issues were managed by the central purchas-
ing group, Miguel thought it would be futile to attempt to increase inventory levels,
although that would surely solve the problem of stock outs.
We inquired about the line capacity and the recent increase from 30 to 36 units per
day. The nameplate bottleneck on the line would limit the line to 40 units in a 24-hour