|
||
|
|
||
|
 |
 |
 |
|
|||||||||||||||||||||||||||||||||||||||||||||||
|
Average = Auction
Stanley Bentley, President of dsi Most people in our
industry celebrated the end of 2001 because it was “the end of
2001!” While dsi did not
post stellar numbers, the fact that we were profitable (without the
help of Arthur Anderson, creative accounting, or off-shore affiliates)
is somewhat unique. We can all explain away Q4 of 2001, but what about
Q1, Q2, and Q3? We need to look at how we operated during those times
to better understand why we were affected by September 11. Virtually every
dsi customer pushed out orders
during Q4, and we, in turn, did the same thing. For our Printed Circuit
Board division, this was a fairly smooth process. For our Assembly division,
this was a much more difficult task. The reasons are simple. PCBs are
a process, and while different boards use different processes, the raw
materials for a given process are common to all boards. Therefore, the
task of rescheduling material is predictable because you are dealing
with bulk material and only a few part numbers. With assembly, the number
of parts in play is large, and many parts are unique to a given customer.
dsi has over 30,000 active part numbers in our system; over 15,000 of these actually reside in our stockroom. If all customers reschedule in the same time period (like Q4), then it is reasonable to assume that most of these 30,000 part numbers will require some attention. I can very confidently say, “systems saved our butts!” Many Manufacturing Requirements Planning (MRP) systems were designed to satisfy accounting purposes and tend to view manufacturing as a necessary evil. The extremely high-mix nature of our assembly operation (we built over 2,500 different part numbers in 2001) requires that systems, reports, and procedures be created in addition to the deliverables necessary to generate an income statement. Properly designed, the same systems that manage growth can also smooth out a decline. Always remember, the market will tolerate and forgive mistakes during growth, but it is brutal to miscues during a decline. I contend that if you examine the growth mistakes of Q1, Q2, and Q3, you will find the root cause to the problems of Q4.
One of the advantages
of a little gray hair is that people will sometimes solicit your advice
(they might even pay you). Consider the case of an assembler who was
losing money. The assembler felt there was not enough business to cover
cost and began taking additional low margin orders to improve “absorption
of overhead.” It took very little analysis to determine that the
problem was the opposite of this assumption. The assembler did not need
more business because, in fact, business was too good. This company
quoted using a “cost per placement” obtained from some industry
publication. Because the company
was a job shop, not a volume assembler, the assembler failed to understand
that survival depended upon converting to an activity based cost, not
an average. If you lose a little on everything you do, more volume hastens
the funeral. A simple example shows the problem: Assume a line costs
$1 million and is used for a single product. The product has ten placements
and all material is supplied by the customer. The line runs one shift
with three people per shift: a loader, unloader, and supervisor, and
can produce 1 million parts per year. The customer changes the design
to only one placement, which he still supplies. Does the new design
cost ten percent of the old design? Of course not; they actually cost
the same—the same capital, same people, and same volume. Remember, Average
= Auction. I have long contended that dsi is an early warning indicator. In a company meeting early this year, I told our employees that we believed the bottom had been reached in December and a recovery had begun. We could not predict the shape or slope of the recovery but we did anticipate that it would not be steep. For that reason, we continue investing in new technologies even though the customer demand is still anemic. Why? Simple—no guts, no glory! It takes over a year to train newly-hired CAM operators and at least six months to properly stabilize and characterize a newly purchased process line. If you want to take advantage of a recovery, start before it is in full swing. To quote our new national slogan, “Let's Roll!” |
