QUALITY

AND THE

BOTTOM LINE

The Tip of the Iceberg

When accounting for quality, dont forget
the often hidden costs of poor quality
by Joseph A. DeFeo

C
MARIA KNIER

OMPANY X WANTED TO REDUCE

operating costs by 10%. It began with a
mission to have each executive identify
where costs could be cut in business
units. The executives created a list of 60
items, including things like eliminating
quality audits, changing suppliers, adding new
computer systems, reducing staff in customer services and cutting back R&D.
For example, the executives removed functions
that provided quality and services to meet customer needs. They bought inferior parts and
replaced computer systems at great expense. They
disrupted their organization, particularly where the
customers were most affected, and reduced the
potential for new services in the future.
After accomplishing this, most of the executives
were rewarded for their achievements. The result?
Their cost reduction goal was met, but they had dissatisfied employees, upset customers and an organization that still had a significant amount of
expense caused by poor performance.

Misconceptions about the cost of quality

The financial benefit to the
bottom line of an organiza-

tions balance sheet by improving the cost of quality with initiatives such as Six Sigma is not always
fully appreciated or understood. This misunderstanding stems from the old misconception that
improving quality is expensive.
This misconception is partially true. For example,
if my organization provides a service to clients for a
given price and a competitor provides the same
basic service with enhanced features for the same
price, it will cost my company more to add those
features that the competitor already provides.
If my organization doesnt add those features, it
will lose revenue because customers will go to a
competitor. If we counteract by reducing the price,
we will still lose revenue. In other words, the quality of my competitors service is better.
For my organization to remain competitive, it will
have to invest in developing new features. This positively affects revenue. To improve quality, features
have to be designed inor in todays terminology, a
new design must be provided at Six Sigma levels.
Because of this historical misconception,
organizations do not always support
the notion that a Six Sigma initiative will affect costs other
than add to them. They
M A Y

2 0 0 1

29

THE TIP OF THE ICEBERG

overlook the enormous costs associated with poor

performance of products, services and processesthe
costs associated with not meeting customer requirements, not providing products or services on time or
reworking them to meet customer needs. These are
the costs of poor quality.
If quantified, these costs will get immediate attention at all management levels. Why? When added
together, the costs of poor quality make up as much as
15 to 30% of all costs. Quality in this complete sense,
unlike the quality that affects only income, affects
costs. If we improve the performance of products, services and processes by reducing deficiencies, we will
reduce these costs. To improve the quality of deficiencies that exist throughout an organization, we must
apply breakthrough improvements.
A Six Sigma initiative focused on reducing the costs
of poor quality due to low sigma levels of performance and on designing in new features (increasing
the sigma levels) will enable management to reap
increased customer satisfaction and bottom-line
results. I have seen too many organizations reduce
costs by eliminating essential product or service fea-

tures that provide satisfaction to customers while

ignoring poor performance that costs the bottom line
and shareholders millions of dollars.

Another example

Company Y approached its situation differently

than did Company X, as described at the beginning of
this article. The executives identified all costs that
would disappear if everything worked better at higher sigma levels. Their list included costs associated
with credits or allowances given to customers because
of late delivery, inaccuracy or errors in billings, scrap
and rework, and accounts payable mistakes caused by
discount errors and other mistakes.
When this company documented its costs of poor
quality, the management team was astounded by the
millions of dollars lost due to poor quality of performance within the organization.
This total cost of poor quality then became the target. The result? Elimination of waste, a return to the
bottom line from planned cost reductions and more
satisfied customers. Why? Because the company eliminated the reasons these costs existed in the first place.
There were process and
product deficiencies that
caused customer dissatisFIGURE 1
Six Sigma and the Bottom Line
faction. Once these deficiencies were removed, the
quality was higher and the
Right features
Lower deficiencies
costs were lower.
While responding to
customer demands for
improved quality in eveCustomer loyalty
rything an enterprise does
is becoming essential,
organizations should not
overlook the financial
Fast cycle
Low warranty
Low scrap/
impact of poor perforGood price
Market share
time
costs
rework
mance. In fact, the cost of
poor quality should be the
driver of the project selection process for Six Sigma.
In other words, the cost
Higher revenue
Lower total cost
of poor quality provides
proof of why changes
must be made. The need
to improve an organizations financial condition
correlates directly with
the process of making
and measuring quality
How do we achieve these desired results?
improvements. Regardless

30

QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

of the objective you start

with, enhancing features as
well as reducing the cost of
poor quality will impact the
continuing financial success
of an operation.
While there is a limit to
the amount quality can be
improved when cost effectiveness and savings are
measured against the costs
of achieving them, its not
likely this will occur until
you approach 5- or 6-sigma
levels. A business must pursue the next level of quality
based on what is of critical
importance to its customers.
If customers demand something, chances are it must be
done to keep their business.
If they do not, theres time to
plan ahead.

FIGURE 2

Traditional Cost of Poor Quality

When quality costs are initially determined, the categories
included are the visible ones, as depicted in the iceberg below:

(4-5% of sales)
Waste

Customer returns

Rejects

Inspection costs

Testing costs
Recalls
Rework

Driving bottom-line performance

If you accept the reality that customers and the marketplace define quality, look at Figure 1, which
describes my basic message on how to grow the bottom line. If you have the right product or service features and lower your deficiencies, loyal customers are
developed.
With a competitive price and market share strongly
supported by fast cycle time, low warranty costs, and
low scrap and rework costs, revenue will be higher
and total cost lower. The substantial bonus that falls to
the profit column comes, in effect, from a combination
of enhancing features and reducing the costs of poor
quality.
Before getting into specific ways to identify, measure and account for the impact of costs of poor quality on financial results, lets take a quick look at what
to do first if you are trying to understand how the
costs of quality can drive a financial target.
If, for example, your organization sets a target to
save $50 million, there is a simple methodology to
determine how many improvement projects it will
take to reach that goal. The organization can then
manage the improvement initiative more effectively if
it puts some thought behind how much activity it can
afford. The answer will help it know how many
experts or Black Belts are needed to manage the
improvements and how much training will be
required.

The methodology includes the following six steps:

1. Identify your cost reduction goal of $50 million over
the next two years$25 million per year.
2. Using an average return of $250,000 for each
improvement, calculate how many projects are
needed to meet the goal for each year. For this
example, we would need an incredible 200 projects
(100 per year).
3. Calculate how many projects per year can be completed and how many experts will be required to
lead the team. If each project can be completed in
four months, that means one Black Belt on two projects per four months. Hence, one Black Belt can
complete six projects in one year. We will then need
about 17 Black Belts.
4. Estimate how many employees will be involved on
a part-time basis to work with the Black Belts to
meet their targets. Assume four per Black Belt per
four months. We would need about 200 employees
involved at some level each year, possibly for as little as 10% of their time.
5. Identify the specific costs related to poor performance, and select projects from this list that are
already causing your organization to incur at least
$250,000 per deficiency. If you havent created this
list, use a small team to identify the costs and create
a Pareto analysis prior to launching any projects.
6. Use this method and debate each variable among
the executive team to ensure the right amount of
QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

31

THE TIP OF THE ICEBERG

FIGURE 3

Cost of Poor Quality (COPQ)

As a company gains a broader definition of poor quality, the
hidden portion of the iceberg becomes apparent:

Customer returns

Waste
Rejects

Inspection costs

Testing costs
Recalls
Rework

Excessive
overtime
Excessive
field service
expenses

Premium
freight costs

Pricing or
billing errors

Expediting
costs

Excessive
employee
turnover

Complaint
handling

Development
cost of failed
product

COPQ ranges
from 15-25% of
total cost.

Overdue
receivables

improvement can be supported. All organizations

make improvements, but world-class organizations
improve at a faster rate than their competition.

Where to find costs of poor performance

To put targets of opportunity into perspective, look
at the traditional costs of poor quality and, even more
critically, the hidden costs of poor quality, as shown in
Figures 2 (p. 31) and 3. It is the hidden costs that must
be quantified to get a complete picture of losses due to
poor performance.
These costs of poor quality could disappear entirely
if every activity were performed without deficiency
every time.
Three major categories of costs of poor quality exist
32

QU A L I T Y P R O G R E S S

Planning
delays

M A Y

2 0 0 1

Customer
allowances
Late
paperwork
Lack of
follow-up on
current
programs

Excess
inventory

Incorrectly
completed
sales order

Unused
capacity

Time with
disatisfied
customers

Excessive
system costs

in organizations:
1. Appraisal and inspection costs.
2. Internal failure costs.
3. External failure costs.

Appraisal and inspection costs

Appraisal and inspection costs are costs associated
with inspectionchecking or assuring that deficiencies are discovered before customers are affected.
Examples include:
Testing products or checking documents before providing them to customers.
Reviewing documents and correcting errors before
mailing.
Inspecting equipment or supplies.

 Proofreading reports or correspondence.

Auditing customer bills prior to sending invoices.
Retooling due to poor design.
Discovering deficiencies at this stage avoids serious
failure costs later and helps develop more effective
and efficient inspection methods. There will always be
some costs in this category because some level of
auditing will be needed to assure consistent performance. The point is to avoid excessive costs.

Paying interest or losing discounts for late payments to vendors.

Providing on-site assistance to customers when
field problems occur.
Providing credits and allowances to clients for lack
of performance or late deliveries.
Efforts to correct external failures usually focus on
regaining customer confidence or lost sales. Both are
debatable costs that may or may not be fully calculated.

Internal failure costs

Interpreting the costs of poor quality

Failure costs within an organization are attributed

to the repair, replacement or discarding of defective
work the customer does not see.
Examples include:
Replacing metal stampings
that do not meet specifications
during production.
Repainting scratched surfaces.
Making up for unplanned
computer downtime.
Replacing components damaged when being moved from
one station to another.
Rewriting parts of a proposal.
Working overtime to make up
for slippage.
Correcting database errors.
Stocking extra parts to replace
defective components.
Scrapping products that do not meet specification.
Spending excess accounts payable time to correct
supplier invoice errors.
Engineering change notices to correct errors in specifications or drawings.
These costs may affect customer service indirectly.

The costs of poor quality at this stage are determined by educated estimates used to guide organizational decisions. They should not be part of a monthly
financial analysis, although understanding these costs
may affect the way financial and
cost accounting data are compiled and interpreted.
The precision required to
identify the costs of poor quality
varies depending on how data
are used. When used to help
select an improvement project,
data need not be as precise as
those used in developing new
budgets for a process after it has
been approved.
When you are evaluating projects, data on poor quality help
identify, charter and support
projects with the greatest potential for reducing costs.
Black Belts and teams may select some projects
because of the impact on customers or internal culture, but data must show where costs are highest so
focus can be concentrated on the vital few.
The amount of cost reduction provided by a remedy
is another indicator of project effectiveness. When
planning for a remedy, a task force should develop
supportable estimates of costs that will be eliminated
by the remedy and use those estimates to develop a
budget for the revised process.
There are four major steps in measuring the costs of
poor quality:
1. Identify activities resulting from poor quality.
2. Decide how to estimate costs.
3. Collect data and estimate costs.
4. Analyze results and decide on the next steps.

When you are evaluating projects,

data on poor quality help identify,
charter and support projects
with the greatest potential for
reducing costs.

External failure costs

External failures affect customers directly and usually are the most expensive failures to correct. External
failure costs may result from:
Satisfying warranty claims.
Investigating complaints.
Offsetting customer dissatisfaction with a recovery
strategy.
Collecting bad debts.
Correcting billing errors.
Processing complaints.
Expediting late shipments by purchasing more
expensive means of transportation.
Replacing or repairing damaged or lost goods.
Housing stranded passengers from cancelled
flights.

Identify activities resulting from poor quality

Activities are categorized as resulting from poor
quality only if they exist solely because of deficiencies
assessed when doing appraisals, inspections, and
internal or external cost estimates.
QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

33

THE TIP OF THE ICEBERG

A project team usually begins by measuring the

obvious costs of a problems primary symptom, such
as discarded supplies, customer complaints or erroneous shipments. After a flow diagram of the process
in question has been created and further analysis has
been conducted, additional activities are usually identified as those required, for example, to dispose of and
replace returned items.
Efforts to identify remedial activities are generally
more global since the focus is on costs of poor quality
throughout an organization. These efforts are best
undertaken by one or a small number of analysts
working with a team of mid-level and senior managers experienced in key areas.
The task force usually launches its efforts by identifying major organizational processes and their customers. For each process, the task force brainstorms
major activities associated with poor quality and
expands the list through carefully constructed interviews with individuals representing different levels
within the most critical functions. At this point, the
objective is to prepare a list of activities related to poor
quality, not estimate costs.
Project teams and task forces find it easier to explain
what they are looking for if they have a full list of typical examples associated with poor quality. The examples described earlier fall into major categories of poor
quality costs. Using key words such as rework, waste,
fix, return, scrap, complaint, repair, expedite, adjust,
refund, penalty, waiting and excess usually stimulates
a healthy response, too.

Decide how to estimate costs

When a specific activity related to poor quality is
identified, two strategies help estimate its costs: total
resources and unit costs. These strategies can be used
individually or together.
An example of the total resource approach is how
an operational unit calculates the human resource
time to process customer complaints and the dollar
value of that time. This approach requires two pieces
of data: total resources consumed in a category and
the percentage of those resources consumed for activities associated with poor quality.
An example of the unit cost approach is when a
project team calculates the annual cost of correcting
erroneous shipments. To find that cost, the team
should estimate the cost of correcting an average erroneous shipment, count how many errors occurred in
one year, and then multiply the average cost by the
annual number of errors.
Data for calculating the total resources used in a
34

QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

Calculating
Data for calculating the total resources
used in an expense category come from a
variety of sources:

Accounting categories
Financial and cost accounting systems
often contain specific categories that can
be allocated partly or totally to costs of
poor quality. Typical examples include
scrap accounts, warranty costs, professional liability, discarded inventory and
total department operating costs.

Time reporting
Many organizations routinely ask
employees to report how much time they
spend on specific activities. This makes it
possible to assign some or all of the time in
a category to a specific cost of poor quality.

Other information systems

Other information systems include cost
accounting, activity based cost accounting, materials management, sales or similar reports.
Data for calculating the percentage of
resources used for cost of poor quality
activities can be obtained through a variety of techniques, including:
Informed judgment. Supervisors and
experienced employees can make adequate judgments about what proportion of a departments time is spent on
an activity. This is especially true if the
unit performs very few distinct functions or the effort consumes a very
large or small portion of total time.
Special time reporting. This method
has been used to calculate costs for
processing computer complaints. A
special short-term collection of time
distribution data may be appropriate if
a department performs many different
functions, activity is neither unusually
small nor large, or there is uncertainty
or significant disagreement among

Resources Used
informed individuals as to the percentage of time or money allocated
to a specific activity. A significant
disagreement would typically be
one of more than 10% of the total
amount allocated.
Special data collections. Besides
collecting data on how much
employee time is spent on an activity, an organization might also collect data on the amount of time a
computer network is inoperative,
the volume of items consumed or
discarded, or the amount of time
special equipment or other resources are not used.
In all these examples, the general
calculation to determine costs of poor
quality is:
Cost of poor quality = (cost of total
resources in a category) X (percentage
of resources in category used for
activities related to poor quality)

Unit cost
An example of this strategy occurs
when a project team calculates the
annual cost of correcting erroneous
shipments. To find the cost, the team
should estimate the cost of correcting
an average erroneous shipment, estimate how many such errors occurred in
one year and then multiply the average
cost by the annual number of errors.
Focusing on unit cost requires two
pieces of data: the number of times a
particular deficiency occurs and the
average cost for correcting and recovering from that deficiency when it
does occur.
This average cost, in turn, is computed from a list of resources used to
make corrections, the amount used of
each resource and the cost of each
resource unit.
Unit cost is often the most appropri-

ate strategy when deficiencies occur

rarely and may be costly, when deficiencies are complex and require the
participation of many departments to
correct, or when deficiencies occur frequently and correcting them is so routine that those involved may not
realize their pervasiveness.
Data on the frequency of a deficiency
may come from any of the following:
Quality assurance.
Warranty data.
Customer surveys.
Field service reports.
Customer complaints.
Management engineering studies.
Internal audit reports.
Operational logs.
Special surveys.
Estimating the cost of a single
occurrence usually requires some
analysis. A flowchart showing various
rework loops associated with a deficiency can often help identify all
important resources used.
When searching for resources, consider hours worked by occupation and
level, contracted services, materials
and supplies, capital equipment and
facilities, and cost of money for borrowed or uncollected funds.
To find out how much of each
resource is used, check the following
sources:
Time reporting systems.
Cost accounting systems.
Various administrative logs.
Management engineering studies.
Informed judgment.
Special data collections.
When a team has identified the
amount of each resource used, it is
ready to calculate the cost for each
and add up costs for all resources. The
finance or engineering functions typically will have standard methods for

calculating the unit costs a team might

require.
Here are hints to remember when
calculating unit costs:
Include benefits as well as wages
and salaries.
Include allocated capital costs for
major equipment and facilities.
While this is a minor consideration
for many activities that can be safely
ignored, it is vital for some activities.
Do not be misled by the argument
that capital costs are fixed and would
exist even if deficiencies did not occur.
This is a typical example of the cost of
poor qualitys being hidden by standard practices. If computers were used
more efficiently, it would be possible
to process more jobs without buying
additional equipment. Idle capital or
misused capital resources are a cost of
poor quality just as surely as discarded paper from a faulty computer run.
Be sure to include penalties or misused discounts for late payments and
premium prices paid for rush orders
or shipments.

Other methods
Still other methods can be developed for special projects. For example,
in lost supplies the organization
should calculate the cost that would
have been consumed if there had
been no defects and the cost of supplies actually consumed. The difference between the two is the cost of
poor quality. This type of approach
might also be applied in comparing
actual outcomes with the best others
have achieved.
Special circumstances may lead a
team to develop still other approaches
that are appropriate to the specific
problem.

QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

35

THE TIP OF THE ICEBERG

category might come from a variety of sources such as

accounting, time reporting, other information systems, informed judgment, special time reporting, special data collections and unit costs. These sources are
described in the sidebar Calculating Resources
Used on pp. 34-35.

FIGURE 4

(assuming 1 million items produced and

a cost of $1,000 per defect)

Collect data and estimate costs

70

Analyze results and decide on the next steps

Collecting data on costs of poor quality helps make
decisions such as:
Selecting the most important quality improvement
projects.
Identifying the most costly aspects of a specific
problem.
Identifying specific costs to be eliminated.

$66,807,000
per million

60

Costs of poor quality (waste) per million items produced

(in millions of $)

Procedures for collecting data on costs of poor quality are generally the same as those for any good data
collection:
Formulate questions to be answered.
Know how data will be used and analyzed.
Determine where data will be collected.
Decide who will collect it.
Understand data collectors needs.
Design a simple data collection form.
Prepare clear instructions.
Test forms and procedures.
Train data collectors.
Audit results.
To estimate the costs of poor quality, it is sometimes
necessary to collect personal opinions and judgments
about relative magnitudes of time spent or costs. Even
though precise numerical data are not required for
such estimates, it is important to plan carefully. The
manner in which opinions are solicited affects
responses.
Sampling works when the same activity is performed often in different parts of an organization. All
field sales offices, for example, perform similar functions. If a company has 10 field sales offices, estimates
from one or two would provide a reasonable value for
calculating overall costs of poor quality.

Cost of Poor Quality as a Function

of Six Sigma Perormance Levels

50

40

30

20

10
$6,210,000
per million
$233,000
$3,400
per million per million

The results
Of note is that every organization that has adopted
Six Sigma and integrated the discipline throughout its
operations has produced impressive savings to the
bottom line. More customers were satisfied and
became loyal, and revenues, earnings and operating
margins improved significantly.
For example, Honeywells cost savings have
36

QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

Process performance level

(number of sigmas within specification)

exceeded $2 billion since it implemented Six Sigma in

1994. At General Electric, the Six Sigma initiative
began in 1996 and produced more than $2 billion in
benefits in 1999. Black & Deckers Six Sigma productivity savings rose to about $75 million in 2000, more
than double the prior years level, bringing the total
saved since 1997 to over $110 million.
A more revealing insight into the cost of poor quality as a function of Six Sigma performance levels is the following:
When 3 sigma of the process
that produces a part is within
specification, there will be
66,807 defects per million parts
produced. If each defect costs
$1,000 to correct, the total cost
of poor quality is $66,807,000.
When an organization improves the process to within 4
sigma, there will be only 6,210
defects per million at a cost of
$6,210,000.
At 5 sigma the cost of defects
declines to $233,000 per million,
a savings of $66,574,000 more
than the savings at a process
capability of +/- 3 sigma.
At the near perfection level of
6 sigma, defects are almost
eliminated at $3,400 per million
parts produced. (See Figure 4.)
After all data are collected and
tabulated and decisions are made,
no study of the cost of poor quality should end without a continuing action plan to eliminate a
major portion of the costs that
have been identified. There is no
need to use a complex accounting
method for measuring costs
because it would be expensive
and waste valuable effort. Simple
methods are sufficient.
The most important step in
developing useful cost of poor
quality data is simply to identify
activities and other factors that
affect costs. Any consistent and
unbiased method for estimating
costs will yield adequate information that will identify key targets
for quality improvement. More

refined estimates may be needed for specific projects

when diagnosing the cause of a specific problem or
identifying specific savings.
JOSEPH A. DEFEO is president and CEO of the Juran Institute, a
consulting and training organization headquartered in Wilton,
CT. He earned a masters degree in business administration from
Western Connecticut State University. QP

QU A L I T Y P R O G R E S S

M A Y

2 0 0 1

37