Oct.

13, 2003

Six Sigma: The Critical Link

between Process Improvements and Business Results

by M. M. Kapadia, S. Hemanth, B. Sharda

Six Sigma as a strategy is different from other conventional improvement programs in its
focus on establishing relationships among business Y, customer Y, process Y, and inputs
(Xs). We strongly believe that these important linkages should form the backbone for all
Six Sigma projects.

We would like to present in this real-life case study how Tata Yazaki AutoComp (TYA),
an automobile wiring harness assembly shop in Pune, India, turned to Six Sigma and
achieved breakthrough results in delivery performance by focusing on the critical Six
Sigma linkages.

Metric Linkage Model

TYA recognized that their wiring harness shop’s delivery performance was the output of
internal processes. To improve this output required improving the processes that generate
the outputs. The Six Sigma team established a clear path: Improve work processes to
delight customers and, thereby, achieve business goals (fig. 1).

Fig. 1 Relationship between internal processes and business goals

In order to In order to
Improve work Improve customer Improve the
processes satisfaction business

• In the recognize and define phase, we identified which processes/products to

improve by considering business Y (measure of business goals) and customer Y
(measures of customer satisfaction).
• In the measure phase, we translated the customer Y into individual process Ys
(quality measures of the process output that describe the physical state of the output
and are clearly linked to the customer Ys).
• In the analyze, improve, and control phases, we dealt with Xs. What are the causes
of the undesirable condition of the process Ys? How can we counteract those causes?
How can we keep these solutions in place over time to maintain the gains?

www.sixsigmaforum.com
 2

Fig. 2 Six Sigma metric linkage model1

Business Y

Customer Y

Project Y

Lower Level Y 1 Lower Level Y 2

Key Input (X 1) Key Input (X 2) Key Input (X 3)

Control your
Control yourinputs
inputsAND
ANDOutputs
Outputswillwill
follow…
follow…

Figure 2 shows the different levels of metrics as they are linked in a Six Sigma roadmap.
By controlling the Xs, TYA maintained the improved performance in the process Ys,
which in turn improved business performance.

Our Improvement Journey Begins: Define and Measure Phases

Define Phase: Identify what’s important to the customer and scope the project

One of our customers was not satisfied with TYA’s delivery performance.2 This situation
not only had a major effect on business performance but also resulted in loss of customer
goodwill—a huge intangible. We embarked on our improvement journey with our Six
Sigma armory in August 2002, forming a problem-solving team to thoroughly investigate
and improve shortages in harness supplies.

Analysis of past performance portrayed a grim picture. We observed a wide gap between
customer demand and TYA’s delivery performance, causing loss of sales and customer
dissatisfaction.

For example, in August 2002, about US $62,000 was lost in assured sales revenue. Figure
3 shows that the company lagged behind % variance3 in FPDN (Finished Product
Dispatch Note) to the customer. In other words, even with having the capacity to
produce, TYA was producing far fewer harness sets than the customer was requesting,
resulting in significant losses to the company. This troublesome situation also forced our
customer to buy harnesses from another vendor.

www.sixsigmaforum.com
 3

Fig. 3 Customer Y4

% Variance FPDN
A huge loss
Lower the
of business
better opportunity

! " !
Max Tolerance Level

Target (0 %
May to Aug’02 Variance)
July 02

The team found that two wiring harness models constituted more than 80% of the overall
business for this customer. Hence, we decided to scope the project on these two models.
Each model is supplied to the customer as a set, with each set being made up of three
wiring harnesses–engine, main, and floor harnesses.

A Statement of Work (SOW) signed by the champion, process owner, team members,
Master Black Belt (MBB) and Black Belt (BB) documented the problem statement, the
project objective, financial benefits, project scope and boundaries, etc. The SOW also
included a detailed project time line along with a signed time commitment from all
involved in the project.

Measure Phase: Establish a measurement system and quantify current performance

The team realized that it would be difficult to address all of the issues (generic and
specific) that would come up during the course of the project. Clear demarcation of
starting and ending project boundaries helped us to maintain focus.

In the define phase we looked at the customer Y, i.e., the % monthly variance in FPDN.
In the measure phase, we attempted to quantify the current performance of the three
harnesses:

• Completed a high-level “as is” process map.

• Prepared “dashboards” to monitor daily production quantity against planned
production for each of the three harnesses (Process Ys) that went into making a set
for the two models.

www.sixsigmaforum.com
 4

• Initiated day-to-day monitoring for Process Ys.

• Began noting high-level causes (when actual production was less than planned
production) on the individual dashboards.

Figure 4 shows a sample dashboard for one of the Process Ys (engine harness % non-
compliance against production plan). These dashboards helped us to look critically at
daily performance and thus formed an essential platform for discussions.

Fig. 4 Sample dashboard for harnesses

% Inspection Non -Compliance Engine(Sept 02)

2 3 4 5 6 7 8 9 11 12 13 14 16 17 18 19 22 23 24 25 26 27 28 30
% Variance

Checker
Crimping problem. Connector
problem problem.

No of days

Analyze Phase: Discover and verify the causes (Xs) of the undesirable performance in
the Ys

As daily monitoring progressed, the team started forming insights into different causes.
We developed a comprehensive system to quantify the reasons for “sets lost” and
conducted a detailed Pareto analysis of the high-level Xs (fig. 5) to focus on specific
causes.

The team quickly realized that attacking each of the Xs in itself was a mini Six Sigma
project. A process owner and team members were identified for each of these
improvement projects. The MBB and BB started working with individual teams to lead
them through the improvement cycle. Based on "sets lost" data, five different subprojects,
as shown in figure 6, were initiated.

www.sixsigmaforum.com
 5

Fig. 5 Major areas of concern

(Based on Sets Lost for the Month of Sept and Oct)

100

80

Percent
X1 60
Sets Lost

X2
40
X3
X4 X5 20

0
tu p n
ge tar ow ut
ge utp
or ta or t
a nS a kd rO wn er s
Defect t er
ia l
Sh
a te
Sh
du
ctio
ine
bre
Ch
ec
ke
Un
k no
Oth
Ma Cr Pro ch
La
te Ma

Percent 38.0 30.3 12.0 8.3 5.7 3.9 1.8

Cum % 38.0 68.3 80.2 88.6 94.3 98.2 100.0

Fig. 6 Parent & subproject teams

Parent Project Team

Subproject 1 Subproject 2 Subproject 3 Subproject 4 Subproject 5

Checker Output Late Production Material Shortage Crate Shortage M/c & Applicator
(X5) Startup (X3) (X1) (X2) Breakdown (X4)

MAIC for Each Subproject (deeper analysis for each X)

Each subproject team started working on identifying subcauses for each project. For
example, during analysis, the team quantified the total number of cumulative sets lost
(recall figure 5) in the months of September and October due to crate shortage (X2). The
“Crate Shortage” team then brainstormed reasons and solutions for crate shortages. The
diagrams in figure 7 illustrate in detail the Crate Shortage subproject improvement
journey. By December 2002, the “sets lost” due to “crate shortages” were reduced to 0
sets.

www.sixsigmaforum.com
 6

Fig. 7 MAIC storyboard for Crate Shortage project

# $
(Based on Sets Lost for the Month of Sept and Oct) Crate Shortages
Red - Most probable
100 Bins not daily unloaded at Customer

80 Delayed product usage by Customer

Total sets lost because of

Percent
crate shortage
60 Transporter not bringing Crates on Time.
Sets Lost

40 Transporter not informed by Yazaki to bring in the bins daily

20
Transporter not doing his job, unaware of the impact on business

0 0 Green Bins used to store production/inspected wiring

p
ge rtu wn
rta ag
e Sta do tor
ak c wn
Defect teri
a lS
ho
te
Sh
ort
d uc
tion
ine
bre
w
Ins
pe
Un
kn
o
Oth
ers Impact of storing finished parts in green bins not known
Ma Cra Pro ch Ne
L ate Ma
Shortage of Trolleys /Wire Hangers.
Percent 38.0 30.3 12.0 8.3 5.7 3.9 1.8 Production
Cum % 38.0 68.3 80.2 88.6 94.3 98.2 100.0
Failures

Measure Phase Analyze Phase

Improve Phase

%
Sets Lost

0 0 0 0
Control Phase Before Nov Dec Jan Feb M ar
(Sep-Oct)

Improvement validation

Similarly, as other subproject teams implemented counteractions with appropriate

controls, the % FPDN variance (Y) started dropping from historic high values to closer to
target values. Figures 8 and 9 both present the before and after comparison.

www.sixsigmaforum.com
 7

Fig. 8 Before and after performance (improvement validation)

& ' ! (
) * +, - .
, ) * /0. ) 1* /2.
FPDN %

Day 0 10 20 30 40 50 60 70 80 90 100

34 ' %
Moving Range

' !

Fig. 9 Customer Y
Monthly % FPDN Non-Compliance
Lower the
Define Measure Analyse / Improve / Control Integrate into daily work
better
% Variance FPDN
July 02
May to

Nov
Sept
Aug

Oct

Mar

Apr

Sep
May

Jun

Jul

Aug
Feb
Dec

Jan 03

www.sixsigmaforum.com
 8

Figure 8 presents an I/MR chart showing daily % FPDN. As we can see, there was a
noticeable variation reduction in daily FPDN variance between “before” and “after”
figures. Also, the monthly % FPDN figures started showing dramatic improvements (see
figure 9). TYA went from negative FPDN variances (i.e., shipping less than plan) to
meeting the customer demand and shipping close to what the customer demanded. This
had a huge impact on TYA’s bottom line.

Control Phase: Put a plan in place to make sure problems stay fixed and sustain the
gains

Teams from each subproject wrote detailed control plans to hold on to their gains. For
example, for the Crate Shortage project (shown in figure 7) a new detailed standard
operating procedure (SOP) enforced a control mechanism for ensuring availability of the
required number of crates at any given point of time in TYA.

The importance of the control phase cannot be over-emphasized. Without proper controls
in place, any improved process can easily slide back to its previous performance levels.

The test of this came in May/June 2003 (the project was already three months into the
control phase) when we noticed a quality problem with a vendor-supplied part that led to
material shortages on the shop floor. As a result, % FPDN exceeded target levels for the
months of May and June 2003. The team tackled the problem at the grassroots level with
the help of the vendor, putting the project Y back on track from July 2003 onwards (fig.
9).

An important lesson was learned: “Even after closing the Six Sigma project (with a
traditional three-month monitoring period), the process owner and champion have to
keep monitoring the dashboard and follow the control and corrective action plan
meticulously to tackle any unexpected issues.”

March of TYA from Dependence to Independence

This eight-month-long comprehensive initiative helped TYA to achieve about US

$730,000 of additional sales for the current financial year. Also the team realized how
each operation in the organization is linked to overall business operation (shown in figure
10). In other words, improving the Xs improved process Ys, which in turn resulted in
huge financial gains for this company.

www.sixsigmaforum.com
 9

Fig. 10 Linkages between metrics

X1: No tracking of X1: Buses arrive late X 1 : M ism a tch in X1: Mismatch in
crates
X1: Loose contact q u a n tity actual and system
X2:First board not in
X2: No addition of X2: Pin Breakage X 2 :D e la y in c u sto m quantity
crates with business . home position c le a ran ce
. . . . .
. . . .

Sample Root Causes (X’s)

C h e c ke r B

Oc t No v
90
25
UC L=8 8 . 1 2

80
e 20
u 70
Sets Lost

alV Me a n =6 7 . 2 9

Sets Lost
60
al 15
du

Hours Lost
50
vii LC L=4 6 . 4 5
dn
I
40 10
30
0 0 0 0
20 5
Before Nov Dec Jan Feb M ar
0 10 20 Data Not
(Sep-Oct)
Obs e rva tion Numbe r 0 Available
Con 1 Con 2
Aug Sep Oct Jan Feb Mar Dec'02 Jan'03 Feb'03 Mar'03

Crate Shortage Checker Output Production start up time Material Shortage Spares Shortage

Lower level Y’s

Y1 = Monthly Engine Inspection % Non-Compliance (-ve) Y2 = Monthly Main Inspection % Non-Compliance (-ve) Y3 = Monthly Floor Inspection % Non-Compliance (-ve)

Lower the
Lower the
better Lower the
better better

Sept O ct Nov Dec Jan Feb M ar

Sep O ct Nov Dec Jan Feb M ar Sept O ct Nov Dec Jan F eb M ar

Project Y’s

Monthly % FPDN Non-Compliance FPDN

Lower the
Sets
better Define Measure Analyse / Improve / Control Integrate into daily work Before
% Variance FPDN

After Plan
Prod
July 02
May to

Nov
Sept
Aug

Oct

Mar

Apr

May

Jun

Jul

Aug

Sep
Feb

A 02
Se 2

Ja 2
Fe 3

M 03

A 3
3
pt
02

O 2
N 02

D 2

03

03
Ju 2

Ju 3
Dec

Jan 03

l'

'
n'

'

'0
n'
0

n'

l'0
'0
p'

'0

b'

'0
'0

'0

Se
ar
ct
ug

ug
ec
ov
ay

ay
Ju

Ju
M
M

Customer Y Business Y

www.sixsigmaforum.com
 10

As corporate MBB/BB resources and TYA production manager (project champion), the
authors facilitated and guided this crucial Six Sigma project. Since its success, TYA has
begun to institutionalize Six Sigma, pursuing further quality and delivery performance
improvements resulting in even better financial performance.

TYA is actively developing in-house Six Sigma expertise, training champions, Green
Belts, and Black Belts to assist in the Six Sigma transformation from “corporate
dependence” to “self-sustenance” or “independence.” Time will bear witness to the
efficacy and extent of this transformation.

Acknowledgements

The authors would like to thank Mr. Pradeep Tewari, CEO–TYA, for his support and
guidance during this project.

The authors also thank T. Sundaresan, Materials Manager, TYA, for his active support
during this project. Valuable contributions from team members Chobey, Jain, Maruti,
Prashant, Satyan, Shinde, and Venkatesh are also very much appreciated. Without their
hard work and enthusiasm this project could not have been completed.

About TYA

Tata Yazaki AutoComp (TYA), a subsidiary of TATA Auto Components Systems

(TACO), is a supplier of automobile wiring harness assemblies to many Indian and
overseas customers. Major customers include Tata Motors, BEHR, Honda, General
Motors, and Toyota.

About the Authors

M. M. Kapadia is Head of the Quality Engineering Group and a Corporate Six Sigma
Master Black Belt at Tata Auto Comp Systems, Pune, India. He has a doctorate in
Mechanical Engineering with specialization in Quality Engineering from the University
of Mumbai, India. He also has a master' s degree in Mechanical Engineering from
Pennsylvania State University, USA, and a master' s degree in Business Administration
from St. Joseph' s University in Philadelphia, USA. He has over eighteen years of
experience in quality control and engineering in various industries in the USA and India.
Dr. Kapadia is also an American Society for Quality (ASQ) certified quality and
reliability engineer.

www.sixsigmaforum.com
 11

S. Hemanth is a member of the Quality Engineering Group and a Corporate Six Sigma
Black Belt at Tata Auto Comp Systems, Pune, India. He has a bachelor's degree in
Mechanical Engineering from Regional Engineering College, Surathkal, India. He also
has a diploma in Total Quality Management from the University of Pune, India, and is a
member of ASQ. Mr. Hemanth has been working in the field of quality for the last three
years and has successfully completed several quality improvement projects.

B. Sharda is Deputy General Manager of the Production Department at Tata Yazaki

Auto Comp (TYA), Pune, India. He has a bachelor' s degree in Mechanical Engineering
from Regional Engineering College, Silchar, Assam, India. He has a total of eighteen
years of industrial experience in automobile and white goods manufacturing. Currently,
Mr. Sharda is responsible for manufacturing of wiring harnesses in both the domestic and
export units of TYA employing about 850 people.

References
1
Metric linkage model based on Johnson Controls Black Belt Training Program.
2
To preserve business confidentiality, the specific identity of the customer involved in
this case study and the actual delivery levels are not disclosed in this article.
3
Variance is defined here as the gap in actual dispatch vs. planned dispatch expressed as
a percentage of planned dispatch.
4
The actual % negative variance of FPDN levels is not shown for business
confidentiality reasons; however, we clearly were nowhere close to our target.

www.sixsigmaforum.com