FMEA in a Service Setting

BASIC QUALITY
Using an FMEA
In a Service Setting
by Cecelia McCain
he use of corrective action—being reactive
to quality performance problems by iden-tifying
and eliminating the root cause of
nonconformities—is very common. Getting ahead
of potential problems and designing them out of
processes or preventing them from occurring is a
different challenge.
ISO 9001 includes item 8.5.3, which requires pre-ventive
action to eliminate occurrence and recur-rence
of nonconformities.
Compuware Corp. recently underwent an audit
conducted by a third-party ISO 9001 registrar and
24 I SEPTEMBER 2006 I www.asq.org
received certification for its product, support and
selected services delivery groups.
During the audit the auditor said, “Of course
you don’t use failure mode and effects analysis
[FMEA]” and then asked, “So, how do you address
the preventive action requirement?” She was quite
surprised to learn the process design activities for
Compuware’s nonmanufacturing SOPs did include
the use of an FMEA.
FMEA Components
At Compuware, project teams use an FMEA to
test a process design—or solution—during quality
planning and quality improvement projects to
ensure the process is capable of meeting features
and goals. The teams have found an FMEA addres-ses
the root causes of a problem, meets the desired
outcomes and deliverables and can be implement-ed
effectively.
An FMEA is a quality tool that allows the user to
examine the process from the viewpoint of its
potential for failure (what could go wrong in the
process) and plan for its control through service,
product or process design requirements. The
potential for failure is designed out of the process
before it reaches the customer.
If you have documented processes, regardless of
whether you are in a manufacturing or service
environment, an FMEA easily can be applied to the
steps of those processes.
T
In 50 Words
Or Less
• ISO 9001 requires preventive action to eliminate
occurrence and recurrence of nonconformities.
• Compuware uses failure mode effects analysis (FMEA)
to ensure its service designs and processes meet
desired outcomes and can be implemented effectively.
• FMEA also can help service companies continually
improve.

Take a closer look at the tool itself, starting with
Figure 1.
When completing an FMEA, begin by complet-ing
the header section, which includes the name of
the individual preparing the FMEA, the project
team responsible for designing the process and the
individual who owns the process that is being
designed.
Because working with an FMEA is an ongoing
task, not a one-time event, also assign responsibili-ty
and document the name
of the individual who will be
charged with maintaining
the FMEA after its original
creation.
Finally, document the
original date on which the
FMEA was created. As the
FMEA is revisited, reviewed
and revised throughout the
life of the process, the revi-sion
date portion of the
header section is updated
accordingly.
Column one of an FMEA
is the service feature, which
is something you anticipate
doing or creating to meet
customer needs, or process
feature, the steps necessary
to deliver the feature. At Compuware, project
teams transfer the features from a customer needs
clarification spreadsheet.
The customer needs clarification spreadsheet is
used to express customer needs in precise (measur-able)
terms. The spreadsheet breaks down each pri-mary
need so specific service or product and
process features and goals (responses to need) can
be developed.
In the example shown in Figure 2, a quality
QUALITY PROGRESS I SEPTEMBER 2006 I 25
Prepared by:
Project team:
Owner:
Responsibility:
FEMA date: Original:
Revision:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Service or
process
feature
Potential
failure
mode
Potential
effects
of failure
SEV
Potential
cause of
failure
OCC DET RPN Recommended
action
Who Acts
and target
completion
date
Action
taken SEV OCC DET RPN
FIGURE 1 FMEA Template
SEV = severity rating
OCC = occurrence rating
DET = detection rating
RPN = rating priority number
FIGURE 2 Customer Needs Clarification Spreadsheet
Project: service delivery process
Customer needs Response to need
Primary
needs
Clarified
needs
Service or product
feature (what)
Process
feature (how)
Process
feature goal
Establish and
maintain a
core group
of contract
employees
Sufficient
numbers of
employees
available for
long-term
assignments
On-site recruiting
On-site hiring
Preassignment
screening process
Develop process flow
Implement and
manage on-site
recruiting process
Implement and
manage on-site
hiring process
Manage hiring and
preassignment
screening process
Long-term orders
filled 100% of the
time
100% on-time fill
of long-term
orders
Improved ratio
of employees
available vs.
assigned

BASIC QUALITY
Effect Severity Rating Criteria
No effect 1 Failure would have no effect on the customer.
Slight effect 2 Customer is dissatisfied but still uses our products and services.
Moderate effect 4 Failure results in confrontation with customer, and additional cost is incurred.
Significant effect 6 Failure is significant, and customer rejects service.
Major effect 8 Customer experiences downtime or excessive failure rates, and large costs are incurred.
Extreme effect 10 Products or services do not comply with government regulations.
planning project team was assigned the task of
developing a service delivery process to ensure a
steady pipeline of contract employees would be
available to meet short- and long-term assignment
needs. The team used the customer needs clarifica-tion
spreadsheet to break down each primary cus-tomer
need requiring further clarification.
Brainstorming, a group technique for generating
new, useful ideas, is used to complete the next two
columns of an FMEA. Because project teams are
identifying potential failures, brainstorming is very
important to the successful completion of an FMEA.
Brainstorming gives each person on the project
team the opportunity to express his or her ideas
without being interrupted or criticized. It allows
more creative ideas to be generated quickly than
does an unstructured discussion, which often
focuses on who rather than what is right.
Column two in the FMEA template (Figure 1,
p. 25), potential failure mode, is about problems
that can result while designing the feature or imple-menting
the process—how the feature might fail.
Column three, potential effect of failures, is a state-ment
of the negative consequences of a failure.
Column four is used to determine the severity
rating (SEV) for each potential effect listed in col-umn
three. Using a scale (see Figure 3) of one (no
effect) to 10 (extreme effect), project teams brain-storm
to determine how severely the effect of the
failure impacts the customer. When possible, it also
is quite helpful to include customers in the FMEA
process, especially when determining severity.
Failure Mode Detection and Rating
Next, for each of the potential failure modes list-ed
in column two (Figure 1), the team brainstorms
to identify potential causes and list them in column
five.
Column six allows the project team to assign an
occurrence rating (OCC), which scores the likeli-hood
the cause documented in column five will
occur. Again, a rating scale from one (almost never)
to 10 (almost certain) is used to assign the rating.
If a process is performing very efficiently, the
likelihood the cause of the failure will occur is very
small—thus a low rating. If a process is performing
very inefficiently, with a high number of errors, the
likelihood that the cause of the failure will occur is
very high—thus a high rating.
A typical occurrence rating table for an FMEA in
a manufacturing facility would begin with one in
1.5 million occurrences, which is not realistic for
service companies.
As you develop an occurrence rating table for
your own service setting, examine the processes
with the highest productivity, such as paycheck
processing, invoice processing, order taking and
call handling. Determine which of those processes
have the highest level of productivity, such as
number of paychecks processed each month, and
set the failure rates accordingly.
In the Compuware example shown in Figure 4,
there are no processes with productivity more than
30,000 occurrences each month, so that is the num-ber
we chose to start with before moving down the
scale accordingly.
Column seven (Figure 1) is the detection rating
(DET). Project teams determine how likely it is cur-rent
process controls will detect the cause of the
failure listed in column five. As with the previous
columns, a rating (see Figure 5, p. 28) from one
(almost certain) to 10 (almost never) is used. If cur-
FIGURE 3 Severity Rating Table for FMEA

rent controls are almost certain to detect the cause
of the failure, a low rating is assigned. If there is a
remote likelihood current controls will detect the
cause or there are no controls in place to detect the
cause, a high rating is assigned.
Column eight (Figure 1, p. 25) is the rating prior-ity
number (RPN), which helps project teams de-termine
which issues need the most attention. The
RPN is calculated as:
SEV x OCC x DET = RPN
The remaining columns of Figure 1, nine
through 15, are used to lower the RPN on potential
failures with an RPN of 150 or higher. Column
nine, recommended actions, describes activities to
be taken or what can be done to prevent the poten-tial
failure. Column 10 identifies who is assigned to
complete the recommended actions and the time-frame
the actions are targeted to be completed.
Remember, the objective is to take action before
these failures reach the customer.
Column 11, action taken, describes what was
actually done to mitigate the effect or reduce the
impact of the failure on the customer. Columns 12
through 15 are used to recalculate the RPN by mul-tiplying
severity by occurrence by detection. If the
RPN is still too high, the project team needs to
return to column nine and identify additional
action to be taken until the RPN is reduced to a
level below 150.
FMEA Example
Let’s look at an example in
which a project team was develop-ing
a service delivery process to
provide contract employees for
assignments at external customer
locations. At Compuware, even
though we use the service, product
or process features listed on the
customer needs clarification
spreadsheet to conduct the FMEA,
we also could use the steps of a
high level process flow diagram.
The example shown in Figure 6
begins with the process feature
“implement and manage on-site
recruiting process” from the cus-tomer
needs clarification spread-
sheet illustrated in Figure 2 (p. 29).
The potential failure mode (how the feature
might fail) was determined to be “on-site recruit-ing
process not implemented,” and the potential
effect (negative consequences) was identified as
“insufficient numbers of employees available.”
The project team assigned a severity rating of
eight, meaning an insufficient number of employ-ees
available for assignment would have a major
effect on the customer. The potential cause of the
failure (why this might happen) was determined to
be “no one available to conduct on-site recruiting,”
with an OCC of 8 (high), and a DET of 3 (high).
Multiply severity, by occurrence by detection to
calculate the RPN:
8 x 8 x 3 = 192
Remember, an RPN of 150 or higher needs to be
investigated and the process improved until the
RPN is reduced successfully .
The recommended action was for branch man-agers
to cross train recruiters on the on-site recruit-ing
process by Nov. 15, 2006. The process owner
was responsible for ensuring the branch managers
completed the cross training and for notifying the
project team when training was completed.
Once the actions were completed, the project
team’s next step was to determine whether the
actions successfully reduced the RPN of 192. The
FIGURE 4 Occurrence Rating Table for FMEA
Occurrence Rating Failure Rate Criteria
Almost never 1 1 in 30,000 Process inefficiency very unlikely.
Remote 2 1 in 10,000 Remote number of process inefficiencies.
Very slight 3 1 in 4,000 Very few process inefficiencies.
Slight 4 1 in 2,000 Few process inefficiencies.
Low 5 1 in 400 Occasional number of process inefficiencies.
Medium 6 1 in 80 Moderate number of process inefficiencies.
Moderately high 7 1 in 20 Frequent process inefficiencies.
High 8 1 in 10 High number of process inefficiencies.
Very high 9 1 in 5 Very high number of process inefficiencies.
Almost certain 10 1 in 3 Process inefficiencies almost certain to occur.
History shows many process inefficiencies.

BASIC QUALITY
project team determined the actions taken did
reduce the OCC from an eight (high) to two
(remote). In other words, the action taken reduced
the likelihood the cause will occur by increasing
the pool of trained recruiters.
The project team also determined the action
taken improved the DET from a rating of high
(three), to a rating of very high (two), by improving
the likelihood it could detect the cause before the
customer could be impacted. Note that the SEV of
eight did not change, because the effect of failure
on the customer did not change. The new RPN was
calculated as follows:
8 x 2 x 2 = 32
Once the actions were completed and the new
RPN calculated, the revision date was entered into
the header section of the FMEA.
These steps were repeated until all the features
from the customer needs clarification spreadsheet
were included on the FMEA. Process flow diagrams
also were revised as needed to reflect the actions
taken as the project team works through the FMEA.
To maintain the FMEA, the process owner needs
to review and revise it anytime a change is made to
the process. Also, as processes are reviewed, which
should be at least annually, the FMEA should be
reviewed to ensure nothing has changed that
would impact the RPN of any of the features.
FMEA and Continuous Improvement
An FMEA also can be a useful tool to help ser-vice
organizations identify improvement opportu-nities.
Look at the service related processes within your
own company and determine which processes
have the biggest impact on your external cus-tomers.
Conduct an FMEA to identify features of
the process with the highest RPNs, create a Pareto
analysis to rank the features by RPN and initiate
quality improvement projects.
A good candidate for improvement in most com-panies
is the invoicing process. Improving invoic-ing
not only benefits the customer by increasing
such things as the quality, accuracy and timeliness
of invoices but also benefits the company by com-pressing
payment turnaround time and accelerat-ing
cash flow.
For example, let’s say one step or process feature
in the invoicing process is to enter information into
the billing system. To complete the FMEA, the next
step is to determine how this feature might fail
(potential failure mode), such as inaccurate data
entered into the system.
What is the potential effect of the failure? The
customer would receive an incorrect invoice, to
which an SEV of four (moderate effect) is assigned.
The rating of four is appropriate because the defin-ition
of this rating is, “Failure results in confronta-
FIGURE 5 Detection Rating Table for FMEA
Effect Rating Criteria
Almost certain 1 Controls are in place and almost certain to detect the failure mode.
Very high 2 Very high likelihood current controls will detect the failure mode.
High 3 High likelihood current controls will detect the failure mode.
Moderately high 4 Moderately high likelihood current controls will detect the failure mode.
Medium 5 Medium likelihood current controls will detect the failure mode.
Low 6 Low likelihood current controls will detect the failure mode.
Slight 7 Slight likelihood current controls will detect the failure mode.
Very slight 8 Very slight likelihood current controls will detect the failure mode.
Remote 9 Remote likelihood current controls will detect the failure mode.
Almost impossible 10 No controls in place to detect the failure mode.

allowed Compuware to meet the ISO 9001 require-ment
for preventive action. The tool allows for
determining action to eliminate the causes of poten-tial
nonconformities to prevent their occurrence.
FMEAs provide records of preventive action and
demonstrate a preventive action is appropriate to
eliminate the effects of a potential problem. They
also identify the potential causes of problems and
elimination of those causes—even in a service com-pany.
CECELIA MCCAIN is the director of quality program man-agement
at Compuware Corp. in Detroit and an ASQ
member. She has numerous certifications, including ASQ’s
Six Sigma Black Belt and the Software Engineering Insti-tute’s
tion with the customer, and additional cost is in-curred.”
The customer would confront the company to
resolve the incorrect invoice (especially if the error
is not in the customer’s favor), and the company
would incur additional cost by creating a revised
invoice.
A project team should brainstorm to determine
potential cause of the failure. One cause could be
that the data source (for example, the order form)
itself is inaccurate, making it impossible for the
billing clerk to enter the right information into the
system. Another cause could be human error—the
billing clerk entering the wrong data into the system.
For each potential cause identified, the team
then should brainstorm to determine the OCC and
the likelihood current controls will detect the fail-ure
before it occurs.
For each potential failure the project team then
should calculate the RPN by multiplying severity by
occurrence by detection and take necessary steps to
reduce ratings above 150. A simple improvement a
company could implement to improve the data
source might be to find ways to foolproof its order
form, perhaps through the use of drop-down boxes.
The use of FMEA as a preventive tool easily
capability maturity model.
Please
comment
If you would like to comment on this article,
please post your remarks on the Quality Progress
Discussion Board at www.asq.org, or e-mail
them to editor@asq.org.
Prepared by:
Project team:
Owner:
Responsibility:
FEMA date: Original:
Revision:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Service or
process
feature
Potential
failure mode
Potential
effects of
failure
SEV
Potential
cause of
failure
OCC DET RPN Recommended
action
Who acts
and target
completion
date
Action
taken SEV OCC DET RPN
Implement
and
manage
on-site
recruiting
process
On-site
recruiting
process not
implemented
Insufficient
number of
employees
available
8
No one
available
to
conduct
on-site
recruiting
8 3 192
Cross train all
recruiters on
the on-site
recruiting
process
Branch
managers
11/15/2006
Recruiters
cross
trained
8 2 2 32
FIGURE 6 FMEA Template for Compuware Contract Employee Process
SEV = severity rating DET = detection rating
OCC = occurrence rating RPN = rating priority number

FMEA in a Service Setting

More Related Content

What's hot (20)

Similar to FMEA in a Service Setting (20)

Recently uploaded (20)

FMEA in a Service Setting