The night before, my co-leader and I had reviewed what we knew about
the target work process so we could draft a plan for gathering information
about it. The plan had to answer five questions: (1) how many observations
will the team make of the target work process; (2) what operations will
it observe; (3) which operators; (4) in what order; and (5) who will
perform each role (e.g., process observer, timekeeper, distance measurer)
in making the observations. We prepared a draft plan to speed our work
and ensure that the information we gathered was reliable.
Given the variability in the reports about the cycle time to complete
the blending work process and the manner in which it is done (i.e., interweaving
the work of different orders), it seemed to us that the team needed to
observe multiple cycles meaning the processing of multiple ordersso
it could average the times across these orders to understand the cycle
time for one unit of output (i.e., one order or two cylinders). Our plan
recommended observing the completion of three orders (six cylinders)
since that seemed to provide the team with enough observations to average
and still allow it to finish its evaluation by day's end. Our plan included
only one observation of the cylinder preparation work process, since
that seemed to have a consistent estimate of cycle time and was regularly
done in a batch of 12.
According to our plan, the team would make its observations on the day
shift. It would focus on the operations performed by workers as opposed
to machines since we lacked sufficient time to do both. Nathan would
be the operator for preparing cylinders and Reggie for blending gases
and filling the cylinders. Our plan had Vincent doing the process observation
since he was familiar with both jobs. As process observer, Vincent's
job was to record and describe each work activity executed by the operator
using the Process Observations Data Sheet. I took the utility role, which
includes collecting examples of documents that the workers receive or
produce as a part of their jobs. My co-leader assumed the documentor
role and would enter the information collected by the team into our electronic
forms. We assigned Thomas to the timekeeper role, James to handle distance
measurement, and Clarice to do the spaghetti charting. The timekeeper
reports the end time for each work activity using a stopwatch so that
the cycle time of each work activity can be computed. The distance measurer
reports the distance an operator moves while executing his or her job
using a distance wheel. Vincent, as the process observer, would call
for the time and distance measures and record them along with the identity
of each work activity on the Process Observations Data Sheet. The spaghetti
charter charts the movements of the operator as he or she transports,
travels, and searches in the performance of the job.
When the team met, I shared our plan and got the team's feedback. All
agreed that the plan made sense. Before we started implementing it, I
gave each team member a description of the role he or she was to perform
and had the member read it; we then discussed the roles together as a
team. It is important that each person understands what he or she needs
to do. My co-leader distributed the materials each team member needed
to do his or her assignment and provided Clarice with 12 copies of the
workplace layout form I had previously produced. We reviewed safety guidelines
and made sure everyone put on their safety gear. We then moved into the
workplace to complete our information gathering.
Gathering information took us about five hours. The team made the observations
of the cylinder preparation process, then took a break for lunch. After
lunch, the team completed observations of the blending process. My co-leader
entered information about each process into the forms included in the
electronic Kaizen Tool Kit. I brought him the observation sheets as Vincent
completed them; the co-leader entered them into our Process Analysis
Sheet, a spreadsheet that automatically computes a summary and distribution
of value added and waste. This allows the team to move quickly to the
analysis activity once it finishes gathering information.
Analyze the Amounts and Sources of Waste
Given our automated spreadsheet, analyzing the amounts and sources of
waste requires only that the documentor complete the entry of the process
observations (activity, time, distance) and assign each work activity
to the category of value added or a category of waste. The spreadsheet
is designed to compute the value-added ratio and total time per waste
category and to chart the information depicting the distribution of time
by value added and waste categories. We built separate spreadsheets for
the cylinder preparation process and the blending process so that we
could attack each on its own basis.
It took 141 activities to complete the cylinder preparation work process.
The process was made up of three types of work: paperwork (getting, reviewing,
filling out, filing, organizing, and discarding forms, tags, and orders);
readying the cylinders for preparation (locating and getting cylinders,
checking the cylinders for damage, hooking cylinders up to the manifold);
and prepping the cylinders (venting, vacuuming, and purging their contents).
The vent, vacuum, and purge cycle occurs twice as per the work standard.
Interspersed through the process is travel, transport, and search for
paperwork and cylinders. The worker traverses some 1,255 feet each time
he or she does this work process.
The cycle time we observed for preparation of a batch of 12 cylinders
was 2 hours 19 minutes, and 30 seconds. None of the activities could
be classified as value adding because none materially produce the product
the customer seeksthat is, gas
of a specific blend. The process takes a used cylinder and refurbishes
it for reuse by the company. Cylinder reuse is the standard approach
in the industry to providing vessels for transporting the customer's
gas and controlling cost. The entire process, however, exists by the
choice of the producers. The customer has not specified that he or she
wants the blended gas in a used cylinder, for example. We recognized
that the customer had indicated that price was an issue and that the company
had decided that reusing cylinders would help it address that concern.
That calculation is the producer's judgment and satisfies the producer's
assessment of what it feels it can do. The fact remains that the customer
has not requested this solution. If a customer did specify
that the gas he or she receives be delivered in recycled cylinders, then
at least some portion of this work process would be value adding.
10 presents the distribution of waste observed in the cylinder preparation
work process. The analysis of the cylinder preparation work process showed
that 65% of the work time was consumed in waiting for machine operations
to finish. These machine-driven operations were vacuuming the cylinders
and purging any remnants of old products from them. The team observed
that if it could speed up these machine operations, it could greatly
reduce waste in the work process. The remainder of the time was used
in setup activities (20%), unnecessary processing (8%), travel/transport
(6%), and search (1%). Almost all of the unnecessary processing involved
transferring information that was already recorded on the order sheet
to other documents. This added documentation was not required by the
work standard but had evolved in response to the preferences of different
individuals, some of whom were no longer at the plant.
The blending work process was made up of 336 activities to complete
three orders (six cylinders). These activities fell into five categories:
paperwork (getting, reviewing, filling out, filing forms, etc.); machine
setup (testing for leaks, checking the scale, etc.); cylinder preparation
(scraping and replacing labels, doing touch-up painting, etc.); blending
gases and filling the cylinders; and packaging (labeling, netting, moving
cylinders, etc.). The worker traverses some 3,602 feet to complete the
process for three orders. The total time to produce the required outputs
was 2 hours, 25 minutes, and 14 seconds, resulting in an average cycle
time of 48 minutes and 25 seconds per unit of output (one batch of two
cylinders)assuming that the blending operator does not need to
do the cylinder preparation work process.
11 presents the distribution of waste observed in the blending work process.
The analysis showed that 8% of cycle time was used in performing value-added
work. This was the percentage of cycle time during which the blending
and filling occurs. The remainder of the time was waste, including 28%
setup, 22% travel/ transport, 19% wait, and 18% unnecessary processing.
Most of the unnecessary processing involved paperwork: for example, the
operator recorded the results from testing the scale used in the blending
booth. Discussion with the team revealed that no one ever uses that information
and no one could recall the reason for its being collected. Nathan observed: "That
scale is checked and maintained by its manufacturer on a monthly basis.
I have been here 15 years. The scale has never failed a test. Even if
it did fail, I would need to fix it immediately. I could not continue
the blending process. I have never referenced that log book, nor has
anyone else." "Shooting the cylinder" was also cited as unnecessary.
This involves reading the bar code on the cylinder and associating it
with an order. It seems important, but, as it turned out, the information
is not used. Vincent, a supervisor, offered: "The cylinder has a label
on it that says what is in it. There is an order form that goes with
the cylinder. It has an order number and specifies the blend that should
be in the cylinder. No one uses the bar code. Everyone uses the label
and the order."
Summarize the Results of the Evaluation
As my co-leader completed the analyses of the work processes, Clarice
prepared a visual display of operator movements during cylinder preparation
and blending. She selected up to five sheets of the spaghetti chart of
operator movement for each work process, copying them onto overhead transparencies.
She then produced a single overlaid visual display of movements for each
work process. This gave the team a good way to judge the orderliness
of movement in each process.
The team reviewed the printed reports of the process observations including
the pie charts depicting the distribution of work by value added or category
of waste. It viewed the spaghetti charting for each work process and
pooled its observations of safety issues. The team then prepared a summary
of findings for each work process (see Exhibits 12 and 13).