You are on page 1of 10

Improving Customer Satisfaction through Six Sigma: A Paint Shop Case Study

by M. Kapadia, A. Mishra, S. Hemanth, and V. Limaye

Tata Auto Plastic Systems Limited (TAPS), Pune [a group company of Tata Auto Comp
Systems, (TACO), Pune, India], is the supplier of interior and exterior plastic components
like dashboards, front and rear bumpers, pillars, consoles, air ventilation nozzles, and door
panels for various automobile manufacturers in India and abroad. One of our major customers
was dissatisfied with ongoing quality problems in the painted parts we supplied.1 A Six
Sigma problem-solving team was formed in June 2002 to thoroughly investigate this issue
and improve quality.

As in all our Six Sigma projects, we began and ended our improvement journey with
our customer. A cross-functional team was formed comprising of paint shop personnel, our
customer interface group, and quality engineering.

In June 2002 this team visited the customer and conducted a “voice of the customer”
analysis. Feedback was disturbing and included verbal statements like “You do not supply us
with good quality parts,” “You are slow in responding to us,” and “Your systems are not
stable.” Written communication from the customer included a “Supplier Performance Rating
Index” for TAPS, as presented in Figure 1.

Fig. 1 Voice of the Customer

Lower Project Started

better 1320
Taps received the worst
Quality Rating out of all
Suppliers in June
950

495

Apr 02 May 02 June 02

After the customer visit, the team immediately set out on the Six Sigma voyage of five macro
phases: Define, Measure, Analyze, Improve, and Control—The DMAIC Cycle.

1
Define Phase

Objective: Identify what’s important to the customer and what we are failing to provide.

It was very clear from written and verbal customer feedback (Business Y) that rejection
levels for painted components at the customer end (Project Y) were causing the biggest
dissatisfaction. To reduce this dissatisfaction the team realized that they had to eliminate
defects in our painting processes (lower-level Ys). The linkages between the different project
metrics are schematically shown in Figure 2.

Fig. 2 Schematic - Project Metrics / Linkages

Business “Y”
Supplier PI Value

Project “Y”
Customer ppm

In House First pass yield Lower Level


“Y’s”

Defect 1 % Defect 2 %

The next task for the team was to narrow the scope of the project. Paint shop supplies a
total of nine parts to this customer. A Pareto analysis for rejections at the customer end
(Figure 3) indicated that almost all rejections were due to three components. Thus, the team
narrowed the scope of the project to focus only on these three specific parts.

Fig. 3 Project Focus


(Qty Rejected figures are coded to maintain confidentiality)
Component-wise Pareto Chart for Rejections at the Customer End
for Apr & May 2002
100

80
Qty Rejected

d 200
e tn
cte 60 e
Percent

je c
R re
yt
40 P
Q 100

20

0 0
Rr Frt
gril
l er er rs
mp mp he
D e f ec t Bu Bu Ot

Qty Rejected 119 78 72 2


Percent 4 3 .9 2 8 .8 2 6 .6 0 .7
Cum % 4 3 .9 7 2 .7 9 9 .3 1 0 0 .0

Project Scope narrowed on Bumpers and grills

2
Next, the team went ahead with defining the charter of this project. Figure 4 displays
pictorially the project definition and scope.

Fig. 4 Project Scope and Definition

4. Which characteristics
should be selected
for improvement?
Deliverable Primary Customer
Process Defect Free Customers Care-Abouts
“Painting” Bumpers/Grill •Parts are defect free (CTQ)
•Competitive Price (CTC)
•On time delivery (CTD)

5. What should be
1. Why does the 3. What process measured?
process exist? 2. Who receives/ relies deliverable
on the deliverable? characteristics
are critical
to customer Business Y: Customer Rating
satisfaction?
Project Y: Customer returns ppm
Lower Level Y’s:
First Pass Yield
6. What are the Individual defect level
process boundaries?

Start Painting End


Part Received from Part accepted by customer
Mould Shop / Customer
Process

A benchmarking study (to understand performance gaps) with other world-class paint
shops revealed the urgent need for improving if we had to survive in this competitive
environment. The team very painfully learned that our competitors'quality levels were at
least 4 to 5 times better than ours. This benchmarking exercise led the team to set an
aggressive goal of improving our quality levels by 4-fold in four months. A detailed timeline
was made to monitor the progress of the project, and our team committed to achieving this
target.

Measure Phase

Objective: Identify the characteristic to be measured and quantify the current level.

The process was mapped and flow-charted to clearly establish project boundaries.
Rejections at the customer end for April, May, and June 2002 were analyzed in detail and the
current level of Project Y (Figure 5) was accurately quantified. The actual customer ppm
levels are not shown for business confidentiality reasons; however, we can clearly see the
gap between where we were in the May-June 2002 time frame and where we wanted to go
(our target, based on world-class benchmarks).

3
Fig. 5 Project Y

! "
x2
ppm

x3
ppm
x1
ppm

Target ppm

Apr 02 May 02 June 02

Analyze Phase

Objective: Identify the input variables (key Xs) that affect the Ys the most; i.e., identify
input factors for successful performance of Y.

The team felt that in order to reduce the rejections at the customer end, our painting
processes had to be improved. Defects were prioritized based on rejections at the customer
end for two months. This helped the team to transition from Project Y to lower-level Ys. A
sample analysis for one of the three components (rear bumpers) selected for improvement is
displayed in Figure 6.

Fig. 6 Defect Prioritization (coded) for Rear Bumpers

16 120
14 100
12
10 80

8 60
6 40
4
2 20
0 0
R/D
Spot touch up not OK

Shade variation
Scratch / damage

Bubbles / blisters
Dust
Assly. Not OK

Edge uncover
Scratch painted

Rough surface
after Painting

The team realized that the defect “spot touch-up not OK” was the result of improper
reworking caused by defects like “dust” and “paint run-down.” The team spent many hours
brainstorming causes for these lower-level defects as preventing these would eliminate “spot
touch-up” permanently. Fifteen separate cause and effect diagrams were prepared to look for
potential root causes. A separate task force was also formed for investigating the defect
“assembly not OK.”

4
Based on the cause and effect analysis, extensive investigations for verification of causes
were undertaken to identify the real causes. For example, a 24 full factorial designed
experiment was conducted to investigate if color of paint (light or dark shade), fixture
condition (clean or un-clean), time sequence in a painting run (beginning or end), and type of
bumper (front or rear) had any statistical impact on dust levels on bumpers. Figure 7 shows
the main factor plots for the DOE.

Fig. 7 Main Effects Plot (results from the designed experiment)


n an
ac
k
lve
r g d ea c le nt ar
Bl Si Be En Cl Un F ro Re
2.4

P = 0.064 P = 0.125 P = 0.55 P = 0.001


2.0

1.6
Dust

1.2

0.8

Color Time Fixture Bumper


Condition

We learned that rear bumpers had statistically more dust than front bumpers. This led to a
series of experiments to find the root cause for this observation. For instance, the team found
that the filler particles which got deposited on the backside of front bumpers were falling
down and appearing as dust on rear bumpers. Similar probing was done for other defects.

Apart from these defects, the team also collectively agreed to focus their efforts on
“final inspection” efficiency before shipping to the customer. As plastic painting is a
complex process with very stringent aesthetic/cosmetic requirements, the team identified
measures for effectively and reliably discerning defects at the final inspection stage:

• Extensive training for final inspectors.


• Development and deployment of a “serial, stage-wise inspection” process for big plastic
parts.
• Attribute Repeatability and Reproducibility (R&R) investigation to verify if there were
any issues with our R&R.
• Measurement System Analysis (MSA) study, with our own in-house customer service
representative (who made several trips to the customer to critically understand their
cosmetic requirements) playing the role of the expert or the “voice of the customer.”

5
The result of our analysis phase was the identification of statistically validated Key
Process Input Variables (KPIV’s), as shown in Figure 8.

Fig. 8 List of Key Process Input Variables (KPIV’s)

• Dust
– Dust in paint
– Dust in paint shop
– Dust falling on rear bumper because of filler sanding
– Dust in gun/hose pipes
– Dust from surroundings
– Improper wiping
• Run-down
– Inexperienced painters
– Gun/pipe lines choke
• Scratch
• Grill assembly not OK
• Qualification of final inspectors

Improve Phase

Objective: Determine solutions to improve the Ys by controlling the Xs; quantify their
impact and compare to goal.

Having identified the real causes for various defects, the team brainstormed
counteractions either to remove the causes or at least to reduce the impact of the causes
on each of the defects. The team chalked out a detailed action plan and the process owner
took charge of the improvement actions. A periodic action item review by the Project
Champion and the project management skills of the process owner and Black Belt assisted
the team in keeping this project on the agreed time line. A small specimen of the
improvement action plan is presented in Figure 9.

Fig. 9 Improvement Action Plan (Sample)

As the improvement action items started getting implemented, the in-house quality
levels started improving. Figure 10 shows the “before” and “after” in-house improvements
in first-pass yield (including minor touch-ups, also), dust levels, and run-downs.

6
Fig. 10 Improvement: “In-House” Metrics (in lower-level Ys)

F irst pass yield

100
83.6
80

60

% FPY
47.51
40

20

0
B e fore A fter

Dust Levels Rundown

2.5 2.32 23.98


25
Dust Count per bumper

2 20

% Rundown
1.41
1.5 15

1 10

0.5 5
1
0 0
Before After Before After

As the in-house quality levels got better, customer rejection levels also started falling.
Figure 11 shows the trend chart for rejections at the customer end, revealing a drastic 4-fold
reduction from Apr-Jun 2002 to Sep-Oct 2002. Again, the actual customer ppm levels are not
shown for business confidentiality reasons.

Fig. 11 Improvement in Project Y

# $% & ' (
Project Started
Customer
PPM

)* '
% &# '

Target ppm

Apr to July Aug Sep Oct


Jun 02

7
With fewer rejections and disturbances on their production lines, our customer’s
confidence in TAPS greatly improved. Within four months, this Six Sigma project helped
TAPS to transition from being the worst supplier to one of the preferred suppliers for this
customer; this is amply reflected in TAPS’s supplier rating as shown in Figure 12.

Fig. 12 TAPS Supplier Rating trend

Supplier
Performance
Index of Project Started
TAPS
1320 Taps received the worst
Lower Better Quality Rating out of all
Suppliers in June
950
795 5.5 fold improvement
from June results
495

165
30 30

Apr 02 May 02 June 02 July 02 Aug 02 Sep 02 Oct 02

Control Phase

Objective: Implement control measures to ensure critical inputs (Xs) remain in control.

Having achieved the improvements, the next task for the team was to put a
comprehensive control plan in place in the shop. Controls were identified through Standard
Operating Procedures (SOPs) and visual aids. The process owner took the responsibility for
ensuring that the appropriate controls were in place.

Our team still felt the need for a feedback mechanism (for the process owner) in order
to ensure that all “touch points” in the control plan are being fully followed. The team
came up with a novel idea of developing the “Periodic Surveillance Audit” program for the
paint shop. The customer service group played a key role in developing this surveillance plan.
Also, a representative from our customer service group was given the responsibility of
carrying out these periodic audits.

The results of this weekly audit are being reviewed with the process owner. As this is a
customer satisfaction improvement project, the team thought of keeping the customer service
group closely involved through this activity. A sample of the surveillance audit plan is shown
in Figure 13.

8
Fig. 13 Sample of the Surveillance Audit Plan

Other Benefits of This Project

The fundamental thrust of this Six Sigma project was to improve key paint shop processes.
Many action items were generic in nature and soon the project’s effects were noticed on other
products also. For example, the customer ppm on painted products for one of our other major
customers dropped 2-fold in November 2002 from its historically steady levels. All of this
has resulted in cumulative savings of over US $66,000 per year for this paint shop.

The project also made us realize the importance of MSA. When we have many products with
delicate cosmetic features being inspected by various people working in different shifts, the
importance of having a common visual inspection platform cannot be overemphasized.

These gains in both quality and savings have reinforced key lessons for us about the
importance of a data-driven, customer-centered approach to improvement. Six Sigma proved
to be a great improvement strategy for our paint shop, not only bringing the project results we
sought but also paving the way for future successes.

Acknowledgements

We acknowledge the management support of Mr. Bakshi, CEO; Mr. Babu, VP Operations;
Mr. Vamburkar, VP Finance of TAPS; and Mr. Asokan, VP Supply Chain Business Group,
TACO.

Valuable contributions from team members Sanjeev Pandit and Prashant Wadekar are also
very much appreciated. Without their hard work and enthusiasm this project could not have
been completed.

1
Note: To preserve business confidentiality, the specific identity of the customer involved in
this case study and the actual quality levels are not disclosed in this article.

9
About the Authors

M. M. Kapadia is Head of the Quality Engineering Group and a 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.

A. Mishra is Head of the Paint Shop at Tata Auto Plastic Systems, Pune,
India. He has bachelor' s degrees in science and in chemical technology, both from Harcourt
Butler Technological Institute, Kanpur, India. He has a total of fourteen years of experience
in painting processes in scooter, motorcycle, and automobile component manufacturing with
specialization in process control and mass scale production.

S. Hemanth is a member of the Quality Engineering Group and a 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.

V. Limaye is Head of the Customer Service Group at Tata Auto Plastic Systems, Pune, India.
He has a diploma in Mechanical Engineering and a bachelor' s degree in Production
Engineering, both from Shivaji University, Kolhapur, India. He has a total of fifteen years of
industrial experience in the field of materials and logistics. In his current assignment at Tata
Auto Plastic Systems, Mr. Limaye is responsible for various aspects of quality, cost, and
delivery for customers.

10

You might also like