Kanban for Commercial Competitive

Advantage
Dr David J. Newlands
IESEG
2016

The Kanban System

The Kanban system uses simple cards to
strictly control production
The basic idea is that no station is
permitted to produce more than is
immediately required by the succeeding
station
This simple idea prevents the buildup of
inventory
No computer is required!
2

Kanban Production Control

System

Kanban card indicates standard quantity of

production
Derived from two-bin inventory system
Kanban maintains discipline of pull
production
Production kanban authorizes production
Withdrawal kanban authorizes movement
of goods

A Sample Kanban
Part no.:
7412
Description: Slip rings

From :
Machining
M-2

Box capacity

25

Box Type

Issue No.

3/5

To:
Assembly
A-4

The Origin Of Kanban

a. Two-bin inventory system
Bin 1
Reorder
Card

Bin 2

Q-R
R
Q = order quantity
R = reorder point
= demand during lead time

b. Kanban Inventory System

Kanban

Dual Kanbans
P

X
X
X

W Container with withdrawal kanban

Flow of work

P Container with production kanban

Flow of kanban

Kanban Squares
X

Flow of work
Flow of information

10

 http://rides.webshots.com/photo/1516449
426065300684ibyTlR

Kanban Bins
11

 http://rides.webshots.com/photo/1516449
828065300684LOFBBR

12

13

 Volvo fixed position layout

14

Types Of Kanbans
Kanban Square
marked area designed to hold items
Signal Kanban
triangular kanban used to signal production at the previous
workstation
Material Kanban
used to order material in advance of a process
Supplier Kanbans
rotate between the factory and suppliers

15

Determining Number Of Kanbans

average demand during lead time + safety stock
No. of kanbans =
container size
dL S
N
C

where
N = number of kanbans or containers
d = average demand over some time period
L = lead time to produce parts
S = safety stock
C = container size

16

Kanban Calculation Example

Problem statement:
d = 150 bottles per hour L = 30 minutes = 0.5 hours
dL = (150)(0.5) = 75
S = 10% dL = 10% x 75 = 7.5
C = 25 bottles

Solution:

dL S (150 x0.5) 7.5

C
25
75 7.5

3.3 kanbans or containers

25

Round up to 4 (allow some slack) or down to 3 (force improvement)

17

When is the AVERAGE EVER the same as

ACTUAL

Lead time
Dell lead time is much shorter than HP
A train has much higher capacity than a single box

Determining Number Of Kanbans

Actual orders during lead time
No. of kanbans =
S
container size
oL
N
S
C

where
N = number of kanbans or containers
o = actual order quantity required by the customer during a
time period
L = lead time to produce parts
S = safety stock an integer, not a percentage. This is reduced
periodically by 1. When problems are recognised, they are
resolved.
C = variable utilisation of container from 1 to maximum
20
capacity

Kanban Calculation Example

Problem statement:
d = 150 bottles per hour
dL = (150)(0.5) = 75
late delivery
C = 25 bottles
Solution:

dL
(150x0.5)
N
S
C
25
75

3 kanbans or containers
25

21

No rounding required

L = 30 minutes = 0.5 hours

S = 10 pre-delivered as quality and
buffer

Capacity

5 5

56

135

150

165

 If actual demand fluctuates, hold available idle kanban dunage and

stillage
If demand drops below full capacity of a kanban bin, reduce active
capacity.
Start off with a given amount of safety stock.
As peformance reliability (quality and delivery) are demonstrated, and
confidence rises, reduce the number in safety stock by 1 unit per day
until problems arise. Work on those problems to ensure they do not
occur again, then resume reducing safety stock.
Ensure safety stock is produced in special kanban bins that have very
limited capacity and are clearly distinguishable from regular stock.
Enter safety stock on a different cost line than regular orders. Provide
profit & loss data to employees to motivate them.

23

Kanban Calculation Variables to Manage

Nobody uses economic order quantity to calculate order sizes.
You can change the safety stock, redesign the supply chain so that
suppliers respond faster (because they are closer or because they have a
store nearby),

Safety stock as a fixed quantity, not as a percentage

Move closer to customers. Move suppliers or create them
closer to you.
Reduce deliver lead-time
Speed up transport mode
Demand can be split between various suppliers, reduced because of multisite manufacture, split into multiple product lines at the same site.
Size vehicles according to the size of the goods to be transported.

24

DHL: express and standard transport

This type of company provide vehicle size flexibility and optimisation

Choose the type (size and25number) of vehicles according to the volume

(quantity and size requirements of goods) defined by the kanban calculation

The Peel 50, the world's smallest road-legal car, sits

parked between two vans in London's Piccadilly Circus.

http://news.bbc.co.uk/2/hi/in_pictures/78
72510.stm

26

Table comparing JIT (kanban) with MRP2

How Toyota's Kanban philosophy differs from a typical Western company

Factors

Toyota's Kanban

Western Philosophy

Inventory

A liability. Every effort

must be extended to do away
with it.

Lot sizes

Immediate needs only. A minimum

replenishment quantity
is desired for both manufactured
and purchased parts.

Set ups

Make them insignificant. This

requires either extremely rapid
changeover to minimise the impact
on production, or the availability of
extra machines already set up. Fast
changeover permits small lot sizes
to be practical, and allows a wide
variety of parts
27 to be made
frequently.

An asset. It protects against

forecast errors, machine
problems, late vendor
deliveries. More inventory is
"safer".
Formulas. We're always revising
the optimum lot size with some
formula based on the trade-off
between the cost of inventories
and the cost of set up.
Low priority. Maximum
output is the usual goal.
Rarely does similar thought
and effort go into achieving
quick changeover.

Table comparing JIT (kanban) with MRP2

Factors
Queues

Vendors

Toyota's Kanban

Western Philosophy

Eliminate them. When problems

occur, identify the causes and
correct them. The correction
process is aided when queues are
small. If the queues are small, it
surfaces the need to identify and
fix the cause.

Co-workers. They're part of the

team. Multiple deliveries for all
active items are expected daily.
The vendor takes care of the
needs of the customer, and the
customer treats the vendor as an
extension of his factory.
28

Necessary investment. Queues

permit succeeding operations to
continue in the event of a
problem with the feeding
operation. Also, by providing a
selection of jobs, the factory
management has greater
opportunity to match up varying
operator skills and machine
capabilities, combine set ups and
thus contribute to the efficiency
of the operation.
Adversaries. Multiple sources are
the rule, and it's typical to play
them off against each other.

Factors

Quality

Toyota's Kanban

Western Philosophy

Zero defects. If quality is not

100%, production is in
jeopardy.

Tolerate some scrap. We usually

track what the actual scrap has been
and develop formulas for predicting it

Equipment
maintenance

Constant and effective.

Machine breakdowns must be
minimal.

As required. But not

critical because we have
queues available.

Lead times

Keep them short. This

simplifies the job of
marketing, purchasing, and
manufacturing as it reduces
the need for expediting.

The longer the better. Most

foremen and purchasing agents
want more lead time, not less.

Workers

Management by consensus.
Changes are not made until
consensus is reached, whether
or not a bit of arm twisting is
involved. The vital ingredient of
"ownership" is achieved.

Management by edict. New systems

are installed in spite of the workers,
not thanks to the workers. Then we
concentrate on measurements to
doing it.

Source: Goddard, W., Kanban versus MRP2 - which is best for you?, Modern Materials Handling, Nov. 5, 1982
29

Mixed Model Production

JIT permits the manufacture or assembly of
a range of products simultaneously, using
the same production equipment
This is known as mixed model production
The result is repetitive flow production
versus traditional batch production
Mixed model production requires small
batches and short setups

30

Balanced Scheduling - Heijunka

Demand

Time

Production
Rate

1
3
4 Week
2
Stabilise orders within production cycle time
Production rate = Sales rate
Not Load Hours
Not Machine Rate
31

Not Constant Period to Period

Not Ability to Produce

Heijunka = Leveling
Non-Level

Leveled
Level
Volume &
Model Mix

Distribute work smoothly over a period of time.

Even distribution of work includes:
Total Volume
Model Mix
Benefits are uniform resource requirements:
People
Material
Machinery
Even workload demand on previous process

32

Example of mixed model

schedulingDaily average

Product Monthly demand

output

A
B
C
D
E
F

1200
400
1600
400
600
600

60
20
80
20
30
30

Totals

4800

240

33

Traditional Batch Schedule

Product
A
B
C
D
E
F

Week 1
1200

Total

1200

Week 2
400
800

Week 3

Week 4

800
400
600
600

34

1200

1200

1200

JIT Schedule
Product
A
B
C
D
E
F

Week 1
300
100
400
100
150
150

Week 2
300
100
400
100
150
150

Week 3
300
100
400
100
150
150

Week 4
300
100
400
100
150
150

Total

1200

1200

1200

1200

35

JIT Production Schedule

Product

Number per lot

A
B
C
D
E
F

6
2
8
2
3
3
24 every 48 minutes

Mixed model production of six different products repeats

on a 48 minute cycle.

36

Smoothing
There are 3 similar products being manufactured.

They require approximately the same amount and type of

resources.
Item R (the red model) accounts for 62.5% of the total
demand; item W (the white model) 12.5% and item B
(blue) 25.0%.
Monthly requirements are 4,400 units.
Currently, the monthly requirements for each item is
produced in a single lot.

37

Smoothing Example
Days per week worked = 5
Item R Item W Item B Total
Product Mix
62,50% 12,50% 25,00%
Units / month2750

550

Current capacity
4400
22
= 200

1100

4400

days/month

2750 200
Production Leadtime
= 13,75
=
2,75
=
3
Available to DeliverLT
=
18

days
weeks
weeks rounded
Chronological Days
(Rounded)
Item R Item W Item B Total
Units per week625
125
250
1000 625 = 2750 monthly demand / 22days month * 5 days week=
Units per day 125

25

50

200

Some of each item becomes available every day.

The ideal would be to set up a repetitive cycle with each item's frequency based on product mix.
RWRBRRBR
Each model is produced every few minutes!

38

Effect on Components (Monthly Lots)

Red
A

White
B

Weeks

Blue

10

15

20

Time
Days
Item R

2750 units/lot
550 units/lot

Item W

1000 units/lot

Item B
Common
Component A
Unique
Components

A
A

D
C

39

Effect of Smoothing on Lower

Level Components

Benefits from smoothing parent item

production (finished products, assemblies)
may cascade to lower level components.
All three share a common component, item
A
Each has a unique component, items B, C
and D.

40

Levelling

Parent item R will run from day 1 into day 13

Item W will start on day 13 and finish on day 16
Item B will start on day 16 and end on day 22

Component requirements for

Day 1: 2750 of A, 2750 of B
Day 13: 550 of A, 550 of C
Day 16: 1100 of A, 1100 of D

With smoothing
Weekly lots: smaller, more frequent component
requirements
Daily production: regular, daily component quantities
41

Effect on Components (Weekly Lots)

Weeks

10

15

20

Time
Days
Item R
Item W
Item B
Common
Component A
Unique
Components

42

Weeks
Time
Days

Effect on Components
1
3
(Smoothed
into2 Daily Lots)
5

10

15

Item R
Item W
Item B
Common
Component A
Unique
Components

Daily smoothing assures a steady supply of the parent items and

eliminates lumpy component demand.

43

4
20

Advantages of Level Scheduling

Batch production becomes repetitive

Low inventories
Flexible and responsive
Simple to control - computers are not
needed!

44

Smoothing - Conclusions
Some of each item becomes available every
day.
The ideal would be to set up a repetitive
cycle with each item's frequency based on
product mix.
RWRBRRBR
Each model is produced every few minutes!

45