1. Submitted on: April 24th
, 2015
Submitted to: Dr. Jeffrey Liker
2015
IOE 591 – Final Report
ZINGERMAN’S MAIL ORDER – CHEESE WASTE KATA PROJECT
ARJUN BHALLA, ERIC LAM, SHIVANI PALEKAR, TING-YI LI

2. Index
Executive Summary
Background………………………………………………………………………………………………… 1
The Challenge……………………………………………………………………………………………. 1
Current State Analysis………………………………………………………………………………… 2
Beginning the Kata Project…………………………………………………………………………. 6
PDCA Exercises
PDCA Exercise 1: Gouda Cheese Waste………………………………………… 7
PDCA Exercise 2: Mountain Wheel Cheese Waste………………………… 10
PDCA Exercise 3: English Clothbound Cheese Waste…………………….. 13
Results…………………………………………………………………………………………………………… 16
Storyboard…………………………………………………………………………………………………….. 17
Future Work: Continuing Kata to strive for 50% reduction……………………………. 18
Kata Learnings………………………………………………………………………………………………. 18

3. Executive Summary
The team from the IOE 591 course from the University of Michigan initiated and successfully completed
an Improvement Kata project for the Cheese Room at Zingerman’s Mail Order (ZMO). In this final report,
the team details the progress of the project over 8 weeks.
The team was invited by ZMO to participate in a Kata project to reduce waste in the Cheese Room. The
focus of the project was reducing waste due to miscuts. After research based on data analysis and
interviews with the operators in the cheese room, several key areas affecting the cheese cutting process
were identified, such as variation in operator skill, limitations and inconsistency of cheese cutting methods
and limitations in current waste reporting procedures.
The studies revealed that 14 cheeses were causing 80% of the total waste caused by miscuts. After
identifying the cheeses with highest waste as a percentage of volume, to start the PDCA exercises, the
team selected cheeses with the greatest potential for improvement, aiming to maximize improvement on
cheese yield in the limited amount of time. The team selected 3 cheese families to focus its effort: Gouda,
Mountain Cheese, and English Clothbound Cheese.
The 3 PDCA exercises had multiple iterations and resulted in significant improvements in waste reduction;
up to 70% and 38% reduction in waste for the cheeses in Gouda and Mountain cheese families
respectively. The PDCA exercise for the English Clothbound cheese was in its early stages of
implementation and exact data for the final PDCA result could not be gathered. However, the team and
the operators are optimistic about the results and estimated the projected waste reduction to be at least
30%, based on the interviews and early results of the pilot. The main improvements implemented during
the 3 PDCA exercises were new cutting techniques and procedures for each cheese family, standardizing
cutting process for operators and a spreadsheet tool for analyzing waste data and creating run charts to
monitor progress.
Overall, the data shows a decrease in waste volume by 17% and a cost reduction of 14% in the 8 weeks of
the Kata project. The PDCA exercises will continue after the departure of the University of Michigan team
with the goal to achieve 50% reduction of waste by October, 2015 by continuing the current PDCA
exercises, developing a structured training program for all operators in the cheese room, conducting PDCA
exercises for other cheeses and benchmarking cutting techniques against large-scale competitors.
Lastly, the team had also gained many insights from experiencing a Toyota Kata project firsthand. From
the challenges of practical implementation of tools and lessons from the Toyota Way, to the importance
of a positive culture in the workplace. These lessons will no doubt be a solid foundation for success for all
participants in this Kata project.

4. Background
The Zingerman's Community of Businesses (ZCoB) is a family of small food-
related companies and entrepreneurial ventures. Each is located in the Ann
Arbor area. They are operated by one or more partners who share ownership
and manage the business.
The online shop for Zingerman’s food, gifts and more. Zingerman's Mail Order
sends extraordinary, traditionally made foods anywhere in America.
Featuring hearth baked breads, handmade cheeses, varietal coffee, estate
bottled olive oils and customized professional presents.
The Zingerman’s Mail Order facility is located at 620 Phoenix Drive, Ann Arbor. It receives orders online,
assembles the products, packs them and dispatches them. Our team worked in the Cheese Room (i.e. the
cheese cutting and packing section) of the facility with its main manager Lisa Roberts and key operators
Tara Stow and Diego Aliste.
The Challenge
The challenge is to achieve a 50% reduction in the quantity of cheese wasted due to miscuts.
• It is considered a challenge because miscuts occur frequently, especially during the holiday
season, and result in monetary losses as the cheese which is undercut (in terms of weight) cannot
be sold.
• The purpose of this improvement kata project, is therefore to increase the amount of saleable
cheese obtained from each wheel of cheese.
• The current data indicates that during the last holiday season (i.e. the fiscal month of December
2014), the cheese waste was at a record high, resulting in a loss of 339.15 lbs., equaling lost sales
of $3452.47.
• ZMO currently sells cheese in the denominations of 0.5 lb., 1 lb. and 2 lbs., with a few customized
orders of 1/3rd lb. as well.
• This issue is compounded by the fact that miscut cheese is left on the shelf till it can potentially
be used for a lower denomination, and this sometimes results in the cheese turning blue or
molding. The waste is then classified as obsolescence instead of a miscut, and capturing the
accurate quantity of cheese wasted due to miscuts is another challenge that the kata team faces.
• The cheese waste during the last holiday season was pegged at 2.8%.
1 | P a g e

5. Current State Analysis
After conducting the initial site visits to the “Cheese Room” at the ZMO facility, the team mapped out the
process flow with high level processes and metrics to be studied closely.
The team found that cheese waste is at a record high of 339.15 lbs. due to miscuts during holiday season,
resulting in lost sales of $3452.47. The team obtained relevant data – cheese wastage reports collected
over the 2013 and 2014 holiday seasons from the coach and the process coordinator as well as the
standard operating procedures for cutting, weighing and packing cheese. The data for the year 2013 was
not usable, as the coach explained that it was not captured accurately, and that some of the miscut cheese
had been wrongly classified as expired/obsolete cheese. So, the cheese waste percentage was actually
higher than what was recorded.
Figure 1: Process Flow in the Cheese Room
2 | P a g e

6. The team found that these are the following key obstacles to cutting cheese efficiently and accurately:
• Optimize Rind-to-Paste ratio to ensure customer satisfaction. The rind is the outside layer that
forms on a cheese during the cheese-making and aging process while the paste is the soft interior
of the cheese.
• Minimize cuts to maintain moisture of cheese and prevent spoilage.
• Manage weight constraints to ensure the customers get what they pay for, while preventing
excess cheese from being shipped.
• Customize cheese the cutting techniques for each type of cheese, specifically:
o Density of cheese
o Shape and size of wheel
o Hardness of rind
o Texture of cheese
• Variation in operator skill due to seasonal and new employees
Using the data from the 2014 holiday season, the team conducted a preliminary analysis and the findings
have been discussed below.
Figure 2: 15 Cheeses are responsible for 80% of total volume of sales as per Quantity (lbs.) of cheese
ordered in December 2014.
Quantity 516 507 495 374 372 287 281 237816901038 960 708 677 669 662 614
Percent 4 4 4 3 3 2 2 1914 8 8 6 6 5 5 5
Cum % 62 66 70 73 76 78 81 10014 22 30 36 41 47 52 57
Cheese Type
O
ther
C-M
A
N
C-TA
L
C-CCC
C-EM
M
C-FM
T
C-EFC
C-TCH
C-PR
L
C-GRU
C-PRR
C-PIA
C-GLC
C
-M
K
E
C-CO
M
C-PAR
12000
10000
8000
6000
4000
2000
0
100
80
60
40
20
0
Quantity
Percent
Pareto Chart of Cheese Type
3 | P a g e

7. Figure 3: Top 14 cheese of waste by quantity that make 80% of total waste. Pareto analysis shows the
total waste per cheese.
The team identified the types of cheese that were subject to the highest miscuts by volume percentage
and identified that 14 of the cheese types were responsible for 80% of the waste by volume. The team
then identified the top ten cheeses that produced the most waste as waste by weight.
Figure 4: Top 10 cheeses by % waste.
The team then determined the cheeses that should be focused on, based on the opportunity for creating
the most amount to waste reduction. Several factors were considered to evaluate the possible
opportunities for improvements.
Waste (lb) 12.8210.569.55 7.37 6.56 6.4766.1647.9830.2430.1027.6125.1620.1519.8118.61
Percent 4 3 3 2 2 2 2014 9 9 8 7 6 6 5
Cum % 69 72 74 77 79 80 10014 23 32 40 47 53 59 65
Cheese Type
O
ther
C-TAL
C-CCC
C-G
LC
C-LAM
C-FM
T
C-PIA
C-EFC
C-EM
M
C-PRR
C-PR
L
C-GRU
C-M
K
E
C-CO
M
C-PAR
350
300
250
200
150
100
50
0
100
80
60
40
20
0
Waste(lb)
Percent
Pareto Chart of Cheese Type
0%
1%
1%
2%
2%
3%
3%
4%
4%
5%
C-GRU C-PRL C-EFC C-MKE C-PRR C-COM C-PAR C-PIA C-GLC C-TCH
Waste/Weight
4 | P a g e

8. There are 5 types of cheese wheel shapes:
• Mountain Wheel: It is a particularly large cheese, typically 80 lbs., and requires special care to
avoid spoilage and bruising.
• Parmesan wheel: It has high volume of sales and has the hardest rind. It is bulky and has a shape,
requiring special tools to cut it accurately.
• Square wheels: Due to the distinct square shape, it is cut diagonally into halves and into odd
weights like 7 lbs.
• Cake shape wheel: It typically weighs 7-12 lbs., but the cheese has different densities and textures,
requiring unique cutting techniques for each cheese.
• Gouda Cheese: One of the highest volumes of cheese sold at ZMO. It has a round wheel shape
and has a hard rind.
After understanding the types of cheese, the team discussed the characteristics of the various types of
wheel shapes with the coach. For example, it was found that Parmesan and Gouda cheese wheels were
the toughest to cut because of factors such as hard rinds, high density and size and shape of the wheel.
The team identified the cheeses sold most frequently and cheeses causing greatest difficulty in cutting,
and the cheese with the highest waste. The team then suggested the approach of improving the cutting
process for specific wheel types that produced the most waste, creating the opportunity to reduce waste
for all cheeses in those wheel types.
The team then further analyzed the data to identify the monetary equivalent of cheese wasted to identify
which types of cheese miscuts were leading to the most monetary losses for ZMO. While the team is
currently focusing on reducing cheese waste by volume, the approach of reducing monetary losses could
be pursued at a later stage.
From the current state analysis and discussions with the coach, the team had the following findings:
• The cheese team needs to make the minimum possible number of cuts in the cheese wheel to
preserve the moisture in the cheese as loss of moisture leads to reduction in quantity.
• In certain wheel types, it was observed that although the weight and texture was similar, certain types
of cheese had double or even triple the waste from miscuts as others. The team concluded that this
was because they had a higher density which required smaller cuts for the same ordered amount. This
left the team with a smaller margin of error to operate in, and led to an increased percentage of waste.
• The team then decided to focus on segmenting cheese by wheel shape and targeting specific shape
to fix cutting of the complete family of cheese.
5 | P a g e

9. Beginning the Kata Project
During the course of the kata project, the team had weekly meetings with the Zingerman’s Cheese Room
team, to discuss obstacles that we encountered and countermeasures for these obstacles, progress in the
PDCA exercises and strategies to ensure continuous
improvement. A tool that the Kata team found extremely useful
and adhered to throughout the process was the “5 coaching Kata
questions”, which is a pattern and routine that’s central to the
Improvement Kata. At the start of each meeting, we went
through these 5 questions in order to understand our position in
the continuous improvement cycle, to evaluate out progress and
to build a future state map. In our absence, the Cheese Room
team used this tool as a common platform for communicating
the status of the ongoing PDCA cycles to each other.
Due to the challenges with cutting processes for multiple types of cheese, we decided to develop 3 PDCA
exercises with multiple cycles within each to ensure simultaneous progress was being made. The pilot
program for each PDCA cycle would run for 2-3 weeks, after which analyses are conducted and further
improvements are made to the cutting process. The figure below is a snapshot of the current state for the
cheeses in the 3 PDCA exercises.
Figure 5: Focus on 3 most wasteful types of cheeses to reduce waste
6 | P a g e

10. PDCA Exercises
PDCA Exercise 1: Gouda Cheese Waste
Our first PDCA cycle focused on the two types of Gouda cheeses (MKE
& PRL). We chose Gouda because not only is it one of the most
frequently ordered cheeses, thereby giving us enough volume to study
and improve upon, but both varieties rank in the top five cheeses by
waste percentage. Additionally, the team at the Cheese Room has
already developed a new cutting technique which they wanted to try
out to see if it would lead to a reduction in miscuts.
Obstacles: In addition to the obstacles discussed previously, the team encountered a few more while
conducting the PDCA exercise, which formed the basis of future efforts.
• Cheese Texture: Due to the hard rind, but relatively soft texture, an 8lbs wheel is cut into 1/8ths
of thin slices. However, the large amount of cutting makes the margin of error very large. Our
current experiment focused on investigating whether reducing the number of rind cuts by cutting
0.5lb pieces across the face would reduce waste due to miscuts.
• Operator Skill: On the graph showing the PRL miscuts, the spikes at 11.57% and 10.14% on the
12th
and the 19th
of February are attributed to the fact that Tara Stow, the expeditor, captain and
the most experienced operator in the “cheese room” was not present, and so the entire operation
was run by two novice cheese-cutters. These values have been classified as outliers in the adjusted
waste graph and the identified obstacle of lack of training and variation in skills was then
incorporated as the basis of the 2nd
PDCA cycle (for reduction of waste from mountain wheels),
during which the kata team will work with the operators in the “cheese room” to develop
standardized cutting procedures for mountain wheels, using jigs and measurement markings on
the wheels to reduce dependency on visual judgement.
Original method New Method
Figure 6: Change in Gouda Cheese cutting technique
Change cutting
7 | P a g e

11. Table 1: Gouda PDCA exercise snapshot
Gouda Cheese: MKE & PRL
Target Condition: Reduce by 30% from 3.60% to 2.40%
Date Started: 2/4/2015 Date Ended: 3/5/2015
PDCA #
(Status)
What we did? Expected Impact What happened? What we learned?
1 Monitored cheese
cutting with respect
to operators
Difference in operators
skill results in waste;
high skill operators have
low waste and others
have high waste
High waste levels
were found
irrespective of
skill level
Errors in cheese
cutting was caused
due to difficult and
error-prone cutting
process
2 Developed new
standardized
cutting process and
cut cheese shape
Reduced waste in
cutting cheese
Reduction in
cheese waste
was found for
few operators
Variation due to
difference in skill and
training
3
Completed
Trained all
operators to follow
new standard
cutting process
Reduced waste
observed for all
operators
Waste in Gouda
is reduced to
1.24%.
Training is key to
sustaining low waste
using new cutting
process
Next Steps: Sustain decreased waste of Gouda Cheese by monitoring new cutting process and continuing
training
There are two main types of Gouda cheese (MKE & PRL). The results from the individual pilots are shown
below. The baseline data is collected from the “Holiday season 2014”, specifically October to December
of 2014, which is considered to be representative of peak cheese volumes. The waste during the pilot is
recorded on specific days. The data is cumulative of the waste and volumes generated from the prior date
to that specific day. The spikes and outliers in the data are caused due to a combination of several factors,
such as lower skill level of new personal, high volumes of cutting and nature of orders. The primary cause
for the outliers was found to be new cutting personnel, who had less practice of cutting cheese as
compared to the regular personal. In order to account for these outliers, we estimated a 50% lower waste
rate for the new operators. These data were compiled as adjusted waste.
MKE
Date
Baseline
Waste
Pilot
Waste
Adjusted
Waste
2/4/2015 3.10% 4.09% 4.09%
2/10/2015 3.10% 5.03% 5.03%
2/19/2015 3.10% 1.70% 1.70%
2/24/2015 3.10% 2.04% 1.77%
3/5/2015 3.10% 0.70% 0.35%
Figure 7: Gouda MKE waste table and run chart showing waste decrease
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
2/4/2015
2/6/2015
2/8/2015
2/10/2015
2/12/2015
2/14/2015
2/16/2015
2/18/2015
2/20/2015
2/22/2015
2/24/2015
2/26/2015
2/28/2015
3/2/2015
3/4/2015
Gouda - MKE
Baseline Waste Adjusted Waste
8 | P a g e

12. PRL
Date
Baseline
Waste
Pilot
Waste
Adjusted
Waste
2/9/2015 4.10% 0.40% 0.40%
2/10/2015 4.10% 5.13% 2.56%
2/12/2015 4.10% 11.57% 5.79%
2/17/2015 4.10% 1.07% 0.54%
2/19/2015 4.10% 10.14% 8.14%
2/24/2015 4.10% 2.07% 1.03%
3/2/2015 4.10% 0.95% 0.95%
3/4/2015 4.10% 1.38% 0.69%
Figure 8: Gouda PRL waste table and run chart showing waste decrease
On the graph showing the PRL miscuts, the spikes at 11.57% and 10.14% on the 12th
and the 19th
of
February are attributed to the fact that Tara Stow, the expeditor, captain and the most experienced
operator in the “cheese room” was not present, and so the entire operation was run by two novice cheese-
cutters. These values have been classified as outliers in the adjusted waste graph and the identified
obstacle of lack of training and variation in skills was then incorporated as the basis of the 2nd
PDCA cycle
(for reduction of waste from mountain wheels), during which the kata team will work with the operators
in the “cheese room” to develop standardized cutting procedures for mountain wheels, using jigs and
measurement markings on the wheels to reduce dependency on visual judgement. Therefore, over the
next few weeks of the final PDCA cycle, the team focused on enhancing training of the operators to reduce
the waste generation.
After training had been completed, the team found a significant reduction in waste generated, as shown
in Figure 9. Therefore, training of operators became a significant part of the future PDCA exercises for
other cheeses.
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
14.00%
2/9/2015
2/11/2015
2/13/2015
2/15/2015
2/17/2015
2/19/2015
2/21/2015
2/23/2015
2/25/2015
2/27/2015
3/1/2015
3/3/2015
Gouda - PRL
Baseline Waste Adjusted Waste
Training operators
9 | P a g e

13. Figure 9: Run chart showing Gouda MKE sustained significant waste reduction
after operators completed training
PDCA Exercise 2: Mountain Wheel Cheese Waste
In our second PDCA cycle, we targeted the mountain cheeses (COM,
GRU and LET). Our experiment focused on developing a new
standardized process to cut this type of cheese wheel. Specifically, the
process involved first cut at the blue label lines to get a piece weighing
in around 1 lb. and the second cut at about 5 inches from the rind which
yielded two large pieces. Further, we also tested whether we could
establish a standardized first cut ratio, based on the size of the cheese
wheel, rather than simply using the 5 inch mark.
Obstacles: In addition to the obstacles discussed previously, the team
encountered a few more while conducting the PDCA exercise, which
formed the basis of future efforts.
• Wheel Size: Due to the large wheel size and weight (typically 70-80 lbs.), it is difficult to cut the
cheese into pieces of accurate weight.
• Cutting Process: There is no current standard for cutting this wheel. So, getting a precise cut
involves a lot of visual judgment on the operator’s part, which is counterproductive in the case of
inexperienced operators. We wanted to introduce a technique which would involve more
conventional measurement and which would be able to be implemented by operators of all skill
levels and experience.
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
2015/2/2 -
2015/2/8
2015/2/9 -
2015/2/15
2015/2/16 -
2015/2/22
2015/2/23 -
2015/3/1
2015/3/2 -
2015/3/8
2015/3/9 -
2015/3/15
2015/3/16 -
2015/3/22
2015/3/23 -
2015/3/29
2015/3/30 -
2015/4/5
C-MKE
C-MKE C-MKE
10 | P a g e

14. Figure 10: Change in Mountain Wheel Cheese cutting technique.
Source:www.youtube.com/watch?v=Qu9Tg7tw6rE
Table 2: Mountain Wheel PDCA exercise snapshot
Mountain Wheel Cheese – C-COM, C-GRU, C-LET
Target Condition: Reduce by 30% from 3.35% to 2.35%
Date Started: 3/8/2015 Date Ended: 4/5/2015
PDCA #
(Status)
What we did? Expected Impact What happened? What we learned?
1 Used the sticker
markings on the
cheese wheels to
cut them into 1/8th
of the size before
making the initial
cut.
Reduced dependency on
visual judgment to
break the wheel down
into workable pieces.
Reduction in waste due
to more precise cuts at
the initial stage.
Steady reduction
in cheese waste
(~15% on an
average for the 3
varieties)
Increased confidence
and ease in cutting for
the operators as the
marking system was
considered to be a
very reliable
alternative.
2 Developed new
measurements for
making the 2nd
cut
in the wheel eights.
Reduced cheese waste
due to a standardized
cutting technique
Dramatic
reduction in
cheese waste
(~51% on an
average for the 3
varieties)
The standardized 1st
and 2nd
cut resulted in
a much lower
variation in the weight
of cheese resulting
from each cut, which
in turned reduced the
waste due to miscuts
3
Completed
Trained all
operators to follow
new standard
cutting process
Gradual further
reduction in cheese
waste for all operators,
and ability to sustain the
improvement.
The cheese waste
was reduced
further.
(~52% on an
average for the 3
varieties, as
observed on
4/5/2015)
The final waste
due to miscuts
was 1.7%
Reducing visual
dependency during
cutting lessens stress
on the operators.
Additionally, training
is key to sustaining
low waste using new
cutting process
Next Steps: Sustain decreased waste of mountain wheel cheese by monitoring new cutting process and
continuing training
Change cutting
Standardized
second cut
11 | P a g e

15. COM
Date
Baseline
Waste
Current
Waste
2/2/2015 2.91% 6.16%
2/9/2015 2.91% 3.90%
2/16/2015 2.91% 3.56%
2/23/2015 2.91% 1.77%
3/2/2015 2.91% 2.62%
3/9/2015 2.91% 2.15%
3/16/2015 2.91% 1.96%
3/23/2015 2.91% 4.47%
3/30/2015 2.91% 1.41%
Figure 11: Mountain Cheese COM waste table and run chart showing waste decrease
GRU
Date
Baseline
Waste
Current
Waste
2/2/2015 4.17% 3.44%
2/9/2015 4.17% 4.63%
2/16/2015 4.17% 2.41%
2/23/2015 4.17% 0.00%
3/2/2015 4.17% 2.71%
3/9/2015 4.17% 4.93%
3/16/2015 4.17% 1.16%
3/23/2015 4.17% 0.00%
3/30/2015 4.17% 1.00%
Figure 12: Mountain Cheese GRU waste table and run chart showing waste decrease
LET
Date
Baseline
Waste
Current
Waste
2/2/2015 3.43% 2.15%
2/9/2015 3.43% 2.56%
2/16/2015 3.43% 8.25%
2/23/2015 3.43% 0.00%
3/2/2015 3.43% 2.71%
3/9/2015 3.43% 4.93%
3/16/2015 3.43% 1.16%
3/23/2015 3.43% 0.00%
3/30/2015 3.43% 1.00%
Figure 13: Mountain Cheese LET waste table and run chart showing waste decrease
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
Mountain Cheese - COM
C-COM Current C-COM Baseline
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
Mountain Cheese - GRU
C-GRU Current C-GRU Baseline
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
Mountain Cheese - LET
C-LET Current C-LET Baseline
12 | P a g e

16. PDCA Exercise 3: English Clothbound Cheese Waste
We have identified a new cutting methodology for this bulky and heavy cheese. The original cutting
method involves cutting the cylinder into half and then cutting the required pieces from the resultant two
halves. This new methodology is based on a traditional way of cutting English clothbound cheese, and
involves cutting a smaller piece in the middle of the cheese cylinder, which results in three pieces of
cheese, and then cutting the resulting pieces into the required size. We expect the miscut waste for the
English clothbound cheese to decrease.
Obstacles: In addition to the obstacles discussed previously, the team encountered a few more while
conducting the PDCA exercise, which formed the basis of future efforts.
• Since the EFC is ordered less frequently than the previously covered cheeses, it was difficult to
observe waste generated on a regular basis, and to implement the new cutting techniques on a
widespread basis, as cheese is cut only when an order for it is processed.
• EFC is an aged cheese with a tougher, wider and bitter tasting rind. So, getting an optimized
cheese to rind ratio is even more critical and difficult to obtain.
• Due to the lack of data, the only valid quantitative method of evaluating this new method is to
continue to wait and collect more data. However operators do report that the new method feels
that it will be more accurate for them in time, but a more standard length of cut for the middle
piece was needed. The kata team will work with the operators of the “cheese room” to
standardize this cut based on the experience of the operators for the new method.
English Clothbound Cheese
Original Method New Method
Figure 14: Change in English Clothbound Cheese cutting technique
Change
cutting Standardize
middle cut
13 | P a g e

17. Table 3: Snapshot of EFC PDCA exercise
English Clothbound Cheese: EFC
Target Condition: Reduce by 30% from 4.3% to 2.7%
Date Started: 4/2/2015 Date Ended: Ongoing
PDCA #
(Status)
What will we do? Expected Impact What happened? What we learned?
1 Observe the current
cheese cutting
method for EFC
Identify the major
weakness causing
inconsistency in the
current cutting method
Varying levels of
waste was found
even for
experienced
operators
The current method of
cutting is unreliable it
required cutting very
thin slices, which is
difficult due to the
large size
2 Developed new
cutting process
where the initial
cheese block is cut
into thirds
Easier cutting process
from the thirds will
result in reduced cheese
waste due to miscuts
There was no
notable changes
in waste levels
when compared
to the original
method
We learnt that it is
important to find an
appropriate length of
middle cut for each of
the (slightly) different
sizes of cheeses
3
Ongoing
Experiment with
size of middle cut to
find in optimal sizes
of thirds
Optimal cuts will result
in thirds that make for
easy cutting, and so will
further reduce waste
due to miscuts
Next Steps: Develop and experiment with additional cutting techniques and continue monitoring the waste
to determine which technique is effective. Once the effective technique is determined, it will be standardized
and all operators will be educated and trained accordingly.
The baseline data is collected from the “Holiday season 2014”, specifically October to December of 2014,
which is considered to be representative of peak cheese volumes. The data is cumulative of the waste and
volumes generated from the prior date to that specific day. The waste during the pilot is recorded as the
data of the first 2 weeks of the exercise. At the time of this report the new method of cutting was only
implemented for one week. There was little data for the English Clothbound cheese because of a
significantly low volume of cheese during these weeks as the spring sale had just ended. During this time,
there was relatively less change in the waste levels of EFC from the previous weeks when the original
cutting method was used.
However, the operators report that the new method seems to be more accurate for them with a standard
length of cut for the middle piece was needed. The operators of the cheese room will continue to
standardize this cut based on their experience with the new method.
14 | P a g e

18. EFC
Date
Baseline
Waste
Current
Waste
2/2/2015 4.30% 0.00%
2/9/2015 4.30% 8.83%
2/16/2015 4.30% 7.75%
2/23/2015 4.30% 2.96%
3/2/2015 4.30% 1.32%
3/9/2015 4.30% 0.18%
3/16/2015 4.30% 9.70%
3/23/2015 4.30% 2.51%
3/30/2015 4.30% 2.69%
Figure 15: English Clothbound EFC run chart with projected improvement
The peaks of waste in the graph on the weeks ending 2/15/2015 and 3/22/2015 are due to particularly
low volumes of cheese demanded during those weeks, which causes waste percentage to seem especially
large even though the actual volume of waste is relatively small when compared to other weeks’ waste.
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
English Clothbound - EFC
C-EFC Current C-EFC Baseline
Projected
15 | P a g e

19. Results
Our team achieved several accomplishments over the 8 weeks period with Zingerman’s Mails Order. We
identified the three main types of cheeses with improvement opportunities, and focused our PDCA
exercises on these areas that resulted in significant waste reduction, as shown in Figure 16.
Cheese Type Improved %
MKE 70%
PRL 58%
GRU 38%
COM 9%
LET 32%
EFC >30%
(Projected)
Figure 16: Decreased waste volume by 17% and reduced cost by 14
Our team decreased the overall cheese waste volume by 17% through our 3 PDCA exercises, which
translates to 14% cost reduction in cheese waste. In addition to cheese waste reduction, our team also
developed a standardized waste reporting procedures to monitor the current state of cheeses. The model
will take inputs from the team, calculate the important output metrics, and generate run charts for the
team. Throughout the process, our team also facilitated cheese team communication to ensure that all
available resources and data are transparent to everyone on the team. We wanted to ensure that all
updates are communicated thoroughly within the team, so the PDCA exercise can be performed efficiently
and effectively.
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
MKE PRL GRU COM LET EFC
Improvement in Waste %
16 | P a g e

20. Storyboard
Once the coaching cycle commenced, we created the initial storyboard and both the coaches and the
learners determined their responsibilities as well as how the Five Coaching Kata questions method would
be used throughout the process. In addition to the question cards, the story board was updated after each
PDCA cycle, and the team then reflected upon it to determine where we currently stood and if the
experiment had improved or worsened the situation. The storyboard served as a common platform for
communication about the Kata updates and the Cheese Room team used it to post ideas for further
improvements as well.
Figure 17: Initial Storyboard
Figure 18: Storyboard towards the end of 3 PDCA exercises
17 | P a g e

21. Future Work: Continuing Kata to strive for 50% reduction
The Improvement Kata project succeeded in reducing the waste due to miscuts by developing
standardized procedures to cut three different types of cheese wheels, improving the waste reporting
process and facilitating sharing of best practices amongst the team in the Cheese Room. However, the
inherent nature of the Kata and of lean thinking is to strive for continuous improvement. Keeping this in
mind, the Kata team provided the team at Zingerman’s with appropriate tools for training the operators
in the Cheese Room and to use the PDCA cycles conducted so far as a stepping stone to further progress.
The goal of the Kata is “to bring about a 50% reduction in the cheese waste due to miscuts by October
2015”, and the team identified areas of further improvement and strategies that will aid in accomplishing
the aforementioned challenge.
• Continue the PDCA exercises for the three types of cheese identified, especially for Gouda and
Mountain Wheels, which are among the most frequently ordered varieties at Zingerman’s.
• Create a structured training exercise for all operators in the cheese room, especially prior to sales and
the holiday season because a lot of part time operators are employed to respond to the increased
volume of orders, and the variation in the skills of these new operators increases the cheese waste
due to miscuts.
• Replicate PDCA cycles discussed above for all 14 types of cheese that were identified as the highest
contributors to waste by volume. As in the PDCA exercises, optimized cutting techniques and
standardized implementation of these techniques would be the key to reducing waste.
• Benchmark cutting techniques against other vendors who deal with large volumes of cheese, such as
Whole Foods as well as Zingerman’s Mail Order cheese suppliers.
Kata Learnings
While working with the team at Zingerman’s on the improvement kata project, we learnt about a variety
of topics, ranging from origin, texture and age of cheese to how critical organizational culture is in order
to bring about a lean transformation. The Kata was a great learning experience for the entire team, and
we hope to transfer some of the knowledge we acquired in the process to future any future projects we
may undertake.
• An existing process may seem to be perfectly satisfactory until you decide to change it. During our
initial briefing of the kata project by the Zingerman’s team, we were faced with the lofty challenge of
halving the waste in what already seemed to be a stable, lean operation to us. (The initial waste due
to miscuts was 2.8%) But, the Toyota Kata provides a platform to not only initiate this change, but
also the framework to work towards the desired results.
• The Kata is a building block to a culture of continuous improvement, and is greatly benefited by a
learning organization with a strong culture which is highly receptive to new ideas or improvements,
and is flexible and open to experimentation. The entire team at the Cheese Room – the learners,
coaches and the managers were extremely encouraging and cooperative with our Kata efforts. After
18 | P a g e

22. every PDCA trial, the operators willingly gave their inputs without worrying about the experiment’s
outcome. This spoke volumes about the openness of Zingerman’s culture and commitment to
learning.
• It is essential to be patient and to not be discouraged by roadblocks while progressing through the
Kata, as every roadblock is a potential opportunity for success. During our PDCA exercise concerning
the Gouda cheese, we observed sudden spikes in the waste volumes after implementing the cutting
technique, despite the new cutting technique being proven to reduce waste. After a discussion with
the Captain at the Cheese Room, it was revealed that the dates corresponding to spikes in the waste
volumes were days when the Cheese Room was run entirely by new operators with little or no
previous training of the new cutting technique. This emphasized to us the importance of creating a
structured training system for new operators and documenting the new cutting techniques so as to
standardize them, and we followed this system for the ensuing PDCA cycles which gave us drastically
reduced waste due to miscuts.
• Finally, the Kata taught the team to be open to all opportunities of learning - an attitude that was
greatly complemented by the Zingerman’s culture. There was a great deal of brainstorming involved
in developing optimum cutting techniques for the PDCA exercises and all avenues were explored,
ranging from mathematically determining the length of the initial cut by incorporating the density and
radius of the wheel into the calculations to watching YouTube videos to study how expert cheese
cutters deal with tough rinds. Benchmarking played an important role in the success of our PDCA
exercises and the Improvement Kata.
19 | P a g e