Professional Documents
Culture Documents
Par$cipant’s
Notes:
These
are
the
main
topics
that
will
be
covered
in
this
module.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
This
roadmap
shows
the
ac)vi)es
and
tools
of
the
measure
phase.
It
is
designed
to
give
you
a
good
overview
of
the
main
steps
in
Measure.
There
are
numerous
tools
and
techniques
within
Six
Sigma
that
help
teams
iden)fy
the
problems
and
solu)ons.
But
you
won’t
use
all
of
them
on
every
project.
If
you
are
feeling
lost
or
overwhelmed
by
all
the
tools,
come
back
to
the
high-‐level
road
map.
It
will
bring
you
back
to
the
big
picture.
The
ac)vi)es
in
the
Measure
phase
are
less
itera)ve
than
in
the
Define
phase.
For
instance,
you
should
always
create
a
data
collec)on
plan
before
you
can
collect
data
and
stra)fy
data.
However,
you
might
go
through
these
steps
two
or
three
)mes.
For
example,
if
you
have
exis)ng
data
and
verified
its
quality,
you
might
analyze
this
data
set
first
before
you
iden)fy
further
data
needs
and
go
through
the
Measure
steps
again.
Also,
in
some
projects
you
might
decide
to
apply
some
tools
that
we’ll
cover
in
the
Analyze
phase
like
detailed
process
mapping
or
cause
&
effect
diagram
before
comple)ng
your
data
collec)on
plan
to
get
more
ideas
where
to
measure.
This
is
only
to
serve
as
a
high-‐level
guide.
Par$cipant’s Notes:
Par$cipant’s
Notes:
Not
all
varia)on
impacts
us.
For
example,
in
most
cases
the
finish
on
a
steel
plate
is
adequate,
but
if
you’re
trying
to
create
an
air-‐
)ght
seal,
the
varia)on
in
surface
finish
could
cause
problems.
Almost
NO
customers
experience
the
average,
they
all
experience
something
else.
Trying
to
tell
angry
customers
that
“on
average”
your
performance
is
great
doesn’t
help
much.
Par$cipant’s
Notes:
If
you
)me
your
commute
to
work
on
a
daily
basis
the
)me
will
vary.
What
are
the
common
causes
for
that
variance
and
what
are
the
things
that
are
unusual?
Par$cipant’s
Notes:
If
you
)me
your
commute
to
work
on
a
daily
basis
the
)me
will
vary.
What
are
the
common
causes
for
that
variance
and
what
are
the
things
that
are
unusual?
Par$cipant’s
Notes:
If
you
)me
your
commute
to
work
on
a
daily
basis
the
)me
will
vary.
What
are
the
common
causes
for
that
variance
and
what
are
the
things
that
are
unusual?
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
We’ll
get
back
to
the
iden)fica)on
of
common
cause/special
cause
varia)on
at
the
end
of
this
module.
We
must
first
explain
some
fundamentals
aspects
of
Measurement
and
Data
Collec)on.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
This
is
a
very
important
concept
of
Lean
Six
Sigma.
Y=f(x).
Memorize
this
concept.
It
applies
to
the
improvement
of
any
result
at
any
level
of
the
organiza)on.
The
point
is
that
in
order
to
improve
an
output
(Y)
we
must
iden)fy
the
inputs
and
factually
iden)fy
which
inputs
are
enabling
a
desired
result
and
which
inputs
are
inhibi)ng
a
desired
result.
Par$cipant’s
Notes:
If
you
want
to
improve
the
Ys
you
must
focus
on
a
calculated
improvement
of
the
defec)ve
Xs.
Par$cipant’s
Notes:
What
is
a
result
they
would
like
to
improve?
What
are
the
key
variables
to
inves)gate?
Par$cipant’s
Notes:
It
is
easier
to
compile
a
list
of
ques)ons
and
then
seek
answers
through
data
than
to
look
at
a
spreadsheet
of
data
and
try
to
figure
out
what
it
is
telling
you.
Par$cipant’s
Notes:
Note:
You
can’t
stra)fy
data
later.
Make
sure
you
determine
stra)fica)on
informa)on
in
advance.
For
example:
You
are
collec)ng
invoices
to
see
if
the
correct
discount
amount
was
extended
to
customers.
You
look
at
1,000
invoices.
Aler
you
collect
all
of
the
data,
you
wonder
if
the
errors
are
a
certain
classifica)on
of
client
(i.e.,
interna)onal
origin).
If
you
didn’t
stra)fy
that
up
front,
you’ll
have
to
recollect
the
data.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
Discrete
Count
data
can
be
reported
as
a
Rate
(#
occurrences
per
unit
of
)me/space).
For
example,
#
calls
per
hour,
or
#
scratches
per
square
foot.
Rates
are
different
than
propor)ons
in
that
propor)ons
range
between
0-‐1,
while
rates
can
be
any
non-‐nega)ve
number.
Some)mes
rates
can
look
like
a
propor)on
(e.g.,
4
flaws
per
100
l
of
wire-‐>
.04)
–
that’s
why
you
need
to
understand
fundamental
data
types.
Par$cipant’s
Notes:
Take
a
couple
of
minutes
to
write
a
C,
DP,
or
DC
in
the
blank
indica)ng
what
type
of
data
this
is.
Par$cipant’s
Notes:
Take
a
couple
of
minutes
to
write
a
C,
DP,
or
DC
in
the
blank
indica)ng
what
type
of
data
this
is.
Par$cipant’s Notes:
Par$cipant’s
Notes:
Write
the
type
of
data
in
the
blank
space
at
the
boNom
of
each
column.
Par$cipant’s
Notes:
Write
the
type
of
data
in
the
blank
space
at
the
boNom
of
each
column.
Par$cipant’s Notes:
Par$cipant’s
Notes:
It
is
important
that
the
informa)on
you
use
to
make
decisions
is
accurate
or
you’re
likely
to
end
up
making
the
wrong
decision
rela)ve
to
the
solu)on.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
Let’s
review
the
best
way
to
create
an
opera)onal
defini)on.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
The
observed
data
is
the
result
of
both
error
in
the
measurement
process
and
varia)on
in
the
actual
process.
Use
MSA
to
check
if
your
data
is
reliable
or
if
you
need
to
take
correc)ve
ac)ons.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
FYI
for
those
who
are
not
scared
of
Greek
formulas:
n
_ ∑x
i =1
i
x=
n
In
the
formula,
“x”
refers
to
an
observa)on
or
data
point.
“i”
is
a
label
indica)ng
which
data
point,
for
example,
x1
or
x5.
“n”
is
the
total
number
of
data
points,
or
sample
size.
Par$cipant’s
Notes:
There
are
as
many
data
points
below
the
median
as
above
the
median.
Par$cipant’s
Notes:
The
mode
is
used
less
frequently
than
the
mean
or
median.
It
is
also
applicable
to
discrete
data,
e.g.,
white
is
the
most
frequent
car
color
–
the
mode.
Par$cipant’s
Notes:
The
average
devia)on
from
the
mean
for
this
data
set
is
6.13
minutes.
Par$cipant’s
Notes:
From
that
same
data
set
you
can
get
all
of
this
informa)on.
These
are
the
descrip)ve
sta)s)cs.
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
What
does
all
this
tell
us?
Once
our
data
has
been
normalized
to
give
us
an
indica)on
of
what
we
would
see
over
)me,
we
can
begin
to
make
predic)ons
by
knowing
the
mean
and
standard
devia)on.
Par$cipant’s
Notes:
For
example,
if
this
were
our
drive
to
work
we
can
conclude
that
there
is
a
68.26%
probability
that
it
will
take
us
between
30
and
40
minutes.
Par$cipant’s
Notes:
And
we
can
conclude
that
there
is
a
99.73%
probability
that
it
will
take
us
between
20-‐50
minutes.
Par$cipant’s
Notes:
If
we
con)nue
all
the
way
out
to
Six
Sigma
(6
standard
devia)ons
+/-‐
from
the
mean)
we
can
conclude
that
there
is
a
99.99966%
probability
that
it
will
take
us
between
5
and
65
minutes.
If
that
falls
within
our
customer
specs,
then
we
have
a
Six
Sigma
process.
If
not,
whatever
falls
outside
the
customer
spec
limits
are
considered
to
be
defects.
(Note:
This
is
a
simplified
explana)on
of
the
sigma
value
computa)on.
It
describes
the
main
idea
behind
the
sigma
value.
The
exact
rela)onship
between
the
normal
distribu)on
and
the
sigma
value
is
more
complex.)
Par$cipant’s
Notes:
The
goal
is
for
calls
to
last
no
more
than
300
seconds.
It
looks
like
the
process
is
not
capable
as
there
are
many
calls
that
last
longer
than
300
seconds.
The
process
needs
to
be
improved.
Why
do
you
think
the
goal
was
set
at
300
seconds?
Par$cipant’s Notes:
Par$cipant’s Notes:
Par$cipant’s
Notes:
Even
if
all
data
points
are
inside
the
control
limits
there
may
be
evidence
of
Special
Cause
varia)on
(non-‐random
paNerns).
Above
are
four
basic
tests,
or
rules,
for
detec)ng
special
cause
(there
are
eight
standard
tests).
Par$cipant’s Notes:
Par$cipant’s
Notes:
The
West
region
seems
to
be
stable.
The
East
region
seems
to
have
a
special
cause.
What
do
you
think
about
the
varia)on
in
each
chart;
is
there
too
much
varia)on
or
does
it
seem
to
be
ok?
Par$cipant’s
Notes:
Both
regions
are
stable;
but
are
they
capable?
What
are
the
advantages
of
looking
at
this
key
metric
by
region?
Par$cipant’s
Notes:
These
are
the
main
topics
that
were
covered
in
this
module.