►
From YouTube: WorkItemsMentalModels Q3FY23
Description
Ben Leduc-Mills (Sr. UX Researcher) goes through the results of the Work Items mental model study performed during October of 2022.
Link to deck: https://docs.google.com/presentation/d/1jcuXr34KPjocJrXH5RH1murt7WO67m3JzjMQCM3QNjI/edit?usp=sharing
A
Hi,
my
name
is
Ben
Leduc
Mills
I
am
a
senior
user
researcher
here
at
gitlab,
and
today
I'm
going
to
share
some
work
that
was
done
on
work
items
and
mental
models.
A
To
do
this,
we
conducted
a
mental
model
study
and
which
basically
involves
users
creating
their
own
mental
models
of
different
work
items,
and
we
did
this
with
18
people,
some
of
which
were
developers.
Some
were
product
managers,
some
gitlab
users
and
some
of
them
out.
A
We
also
found
three
common
workflows
in
our
models
that
connect
across
different
stages
and
were
neatly
segmented
within
the
mental
model.
Maps
task
is
a
really
critical
noun.
It's
an
inflection
point
between
planning-based
tools
and
code-based
tools,
which
is
sort
of
the
point
at
which
product
managers
hand
over
the
work
to
Developers
story
and
roadmap
are
two
important
work
items
that
are
not
currently
a
central
part
of
GitHub,
but
are
Central
to
how
many
users
think
about
their
work.
A
A
So
we
looked
at
hierarchy
in
two
ways,
from
the
study
by
the
parent-child
verbs
that
were
used
in
the
maps
and
also
by
constructing
a
network
graph
and
breaking
it
down.
Both
of
those
methods
led
to
the
conclusion
that
there
are
a
core
group
of
nouns
that
form
the
backbone
of
work
items,
and
these
are
epic
story.
Issue
and
task
task
in
particular,
is
an
inflection
point.
A
It
was
the
only
term
that
appeared
on
every
single
map
that
we
collected
and
it
serves
as
a
bridge
between
project
managers
working
and
obtaining
Centric
environment
to
developers
working
in
a
code.
Subject:
environment
tests
typically
get
translated
from
things
called
the
tasks
in
planning
tools
to
things
like
issues
or
branches
in
gitlab
of
the
nouns
which
are
not
a
central
part
of
gitlab.
Currently
stories
and
road
maps
were
the
most
critical
to
you.
The
work,
in
fact
none
of
the
gitlab
users
that
we
interviewed
mentioned
using
roadmaps
at
all.
A
However,
it
was
very
critical
to
a
lot
of
product
manager
workflows
and
in
segmenting
the
these
Maps.
We
found
that
planning
producing
and
protecting
were
appropriate
labels
for
the
different
divisions
and
natural
segmentations
that
we
found
in
the
models,
in
particular
the
nouns
associated
with
the
planning
segment,
can
provide
clues
to
help
attract
plan-centered
users.
A
My
recommendations
following
This
research,
epic,
is
a
main
character,
so
what
I
mean
by
that
is
that
it's
the
most
Central
most
well-connected
and
most
relevant
work
item
by
a
good
margin.
So
our
future
efforts
in
this
area
should
try
to
look
at
work
items
from
an
epic
centered
point
of
view.
Task
is
our
Bridge.
A
The
universality
of
tasks
is
our
Bridge
from
git
lab
to
other
products.
It's
currently,
as
I
mentioned,
the
handoff
point
between
product
managers
and
Developers,
using
our
tool
and
story
and
roadmap,
are
our
missing
links
in
order
to
have
a
sustainable
planning
workflow
within
gitlab.
We
need
to
support
these
objects
story
is
so
widely
used
elsewhere.
A
It
will
be
critical
to
reconcile
how
it
works
currently
with
other
important
nouns
within
gitlab,
like
epic
issue
and
task
I'll
briefly
go
over
the
methodology
for
how
I
arrived
at
these
conclusions,
but
first
a
mental
note
or
a
note
about
mental
models.
A
So
a
mental
model
is
what
a
user
believes
to
be
true
about
a
system.
Nielsen
Norman
states
that
these
are
some
of
the
most
important
Concepts
in
human
computer
interaction
and
that
what
a
user
believes
about
a
user
interface
strongly
impacts
how
they
use
it.
So
it's
critical
that
when
forming
work
items
we
mirror
as
best
as
possible
what
our
users
believe
to
be
true
about
the
system.
A
Often
usability
mismatches,
cause
lots
of
user
frustration
and
performing
a
study
with
mental
models
is
one
way
to
get
a
sense
of
what
these
beliefs
are
and
avoid
the
pitfalls
of
Designing
a
new
Smash
product.
A
So
what
we
did
is
participants
were
given
a
link
to
a
mural
exercise.
You
can
see
here
that
there's
a
list
of
nouns
in
blue
and
verbs
in
green,
so
participants
constructed
a
map
connecting
two
nouns
with
a
verb
and
basically
use
all
of
the
nouns
that
were
common
in
there
in
their
work,
and
participants
could
also
add
their
own
nouns
and
verbs
as
they
wished.
We
also
did
some
pre-work
constructing
our
own
internal
map
with
our
group
to
compare
against.
A
So
the
work
that
we
did
on
the
map
sort
of
followed
a
three-stage
process.
You
know,
given
the
Raw
map,
we
cleaned
it
up,
put
arrows
cleanly,
Define
the
relationships
and
then
looked
for
segmentations
hierarchies
just
interesting
relationships
in
general.
We
also
exported
all
of
the
data
from
mural.
A
A
A
So
task
was
the
only
Universal
term,
which
lends
more
validity
to
the
fact
that
it
was
the
first
work
item.
It
also
came
up
as
an
inflection
point,
as
I
mentioned
earlier.
A
The
terms
like
story
and
road
map
were
very
prominent.
They
don't
currently
exist
or
not
used
by
many
of
our
participants,
but
they
are
Central
to
how
users
think
of
their
work,
suggesting
possible
areas
of
growth.
Here.
A
I'll
sort
of
skip
through
these
parts
here
feel
free
to
look
at
them.
As
you
wish,
I
break
down
the
noun
usage
over
Persona
versus
tool,
gitlab
or
non-getlab
I
also
break
down
the
verb
usage
between
personas
and
non-getlab
and
gitlab
users.
A
More
interesting
to
the
research
is
sort
of
what
these
Maps
can
tell
us
about
the
information
architecture,
and
this
is
where
we
start
incorporating
the
nouns
and
verbs
together
to
look
at
what
the
behind
the
scenes
structure
of
these
sort
of
collection
of
work
items
is
according
to
our
participants.
A
A
In
some
verbs
imply
the
same
kinds
of
relationships,
but
aren't
exact
duplicates
of
each
other,
there's
also
a
relatively
large
number
of
nouns
and
verbs
to
work
with
a
lot
of
possible
combinations
and,
of
course,
participants
added
a
lot
of
their
own
vocabulary
into
the
models.
So,
for
instance,
what
I
mean
here?
Is
it
epic,
as
a
parent
of
story
is
exactly
the
same
relationship
as
story
as
a
child?
Epic.
We
actually
have
the
same
relationship
here,
which
begs
the
question
of
how
we
can
find
these
hidden
duplicates.
A
So,
in
fact,
while
looking
for
this,
we
noticed
that
many
of
the
commonly
used,
verbs
reply
to
hierarchical
relationship
and
hierarchy
is
critical
to
sort
of
understanding,
information
architecture
as
a
whole,
and
it
was
very
important
to
the
work
items
team,
so
I
filtered
a
list
of
the
relationships
to
focus
only
on
hierarchy
and
basically
normalize
the
list
so
that
all
of
the
all
of
the
parent-child
relationships
were
sort
of
the
same.
This
gave
us
134
non-unique
parent-trail
relationships
which
allowed
us
to
build
a
tree
map,
and
this
is
what
it
looks
like.
A
You
can
see
roadmap,
epic
and
story
here
are
highlighted
as
sort
of
critical
branching
paths
through
the
hierarchy,
and
one
note
to
make
here
is
that,
in
the
case
of
clashes,
for
instance,
where
we
had
a
data
point
saying
epic
is
a
parent
of
board
and
also
one
saying
board,
as
a
parent
of
Epic,
so
sort
of
canceling,
each
other
up,
preference
was
given
to
the
relationship
that
occurred
more
frequently
in
the
data.
A
So
you
can
see
here
that
you
know
a
lot
of
things
are
as
they
are
in
in
git
lab
and
with
the
inclusion
of
some
of
the
terms
that
don't
exist
yet
just
gives
some
clue
as
to
how
we
might
incorporate
them
in
a
hierarchy
in
the
future.
But
it's
important
to
note
tree
maps
are
are
not
you
know
meant
to
be
statistically
valid.
This
is
not
the
only
way
to
think
of
of
hierarchy.
It's
just
one
way
that
we
constructed
it.
A
Based
on
this
particular
data
set,
I
think
the
main
takeaway
here
is
that
gitlab's
current
schema
was
largely
reflected
in
the
hierarchy
that
we
produced.
So
if
we
leave
out
the
nouns
that
don't
exist
in
gitlab
sort
of
hierarchy,
that
we
have
project
board,
epic
issue
task,
Maps,
quite
well
with
our
participants,
mental
models.
A
So
because
of
the
long
tail
of
the
terms
that
participants
used
and
the
small
number
of
relationships
that
naturally
occurred
more
than
once
there's
a
lot
of
data
we
might
be
missing,
even
though
we
did
find
some
hidden
duplicates
by
looking
at
you
know,
smashing
all
of
the
sort
of
hierarchical
verbs
together,
there's
a
lot
that
we
could
be
missing
so
by
taking
each
noun
verb,
noun
relationship
and
treating
it
as
a
connection
in
a
network
graph,
we
get
a
sort
of
a
node,
Edge,
node
relationship
which
can
generate
a
picture
and
that
can
uncover
interesting,
less
visible
properties
of
our
data.
A
And
what
I
mean
here
is
that
we
end
up
with
something
like
this.
Where
blue
circle
here
represents
noun
and
each
line
represents
a
relationship
from
one
noun
to
another,
the
bigger
the
circle,
the
more
relationships
it
has
with
other
nouns.
So
each
of
the
lines
here
is
a
verb
relationship
and
you
can
see
hopefully
the
arrows
there.
So
it
is
actually
a
directive
relationship
and
several
things
sort
of
jump
out,
but
there's
a
lot
of
ways
to
look
at
this.
A
One
way
of
looking
at
this
graph
is
by
counting
the
number
of
connections
that
it
has.
In
this
case,
epic,
which
was
only
the
third
most
frequently
used
noun,
is
actually
the
one
with
the
most
connections,
task
story,
roadmap
and
issue
round
out
the
top
five,
and
you
can
sort
of
easily
pick
those
apart
visually,
but
there's
a
lot
of
other
ways
that
we
can
measure
a
network
graph.
A
I
will
leave
it
to
you
guys
to
read
exactly
what
relevance
between
this
closeness
and
cornice
are,
but
they're
different
ways
of
measuring
the
importance
or
centrality
or
connectedness
of
nodes
in
a
graph,
and
what
we
find
is
that
by
most
of
these
other
measures,
the
same
nouns
always
tend
to
come
to
the
top.
In
this
case,
epic
is
at
the
top
of
all
four
of
these
measurements
story.
Tasks
and
issue
also
appear
at
or
near
the
top.
In
all
of
these
cases,
there's
a
few
interesting
side
notes
here.
A
A
But
basically
this
is
this
idea
is
that
we
can.
Basically,
there
are
a
course
set
of
nouns
that
are
a
center
of
the
user's
work.
There's
a
lot
of
evidence,
numerically
in
this
graph
to
suggest
that
that's
the
case,
you
know
we
do
end
up
with
some
solid
data
to
support.
A
You
know
some
of
the
hierarchy
like
epic
is
above
story
stories
above
task.
One
thing
to
tease
out
is
that
issue
and
task
are
sort
of
siblings
more
than
they
are.
You
know
a
parent-child
relationship,
and
so
some
effort
will
need
to
be
made
to
to
sort
of
make
sure
we're
reflecting
those
in
a
way
that
reflects
our
users,
mental
models
and
expectations.
A
A
Our
models
are
sort
of
naturally
segmented
into
into
three
areas,
so
kind
of
planning
or
thinking
about
the
work
performing
or
doing
the
main
chunk
of
the
work
and
protecting
or
fixing
securing
the
work.
Basically,
we
can
see
this
in
both
map
examples
and
the
graph
so
for
planning.
A
The
nouns
are
like
road
map
calendar
key
metric
things
like
that
they're
used
by
product
managers,
more
often
and
kind
of
represents.
The
top
of
our
hierarchy
in
our
tree
map
performing
is,
is
the
core
the
meat
and
bones
here.
That's
the
largest
segment
uses
nouns
like
epic
and
story
issue.
A
It's
used
by
both
project
managers
and
Developers
and
with
protecting.
We
see
nouns
like
alert
incidents,
defect,
vulnerability
and
maintenance,
and
it's
used
primarily
by
developers
more
than
product
managers.
A
A
Basically,
you
know
we
need
to
review
these.
You
know
early
on
in
our
in
our
process
to
ensure
that
the
alignment
of
our
product
is
aligned
with
our
users.
Basically,
that's
it.
The
next
steps
here
are
to
hold
an
actionable
insights
Workshop.
So
there's
a
lot
to
digest
here,
we'll
build
a
workshop
to
identify
and
prioritize
which
issues
to
tackle
first,
some
of
these
insights
will
likely
need
further
research,
we'll
need
to
collaborate
across
design
and
products
to
prioritize
those
research
efforts.
A
Thank
you
so
much
especially
like
to
thank
Blair,
Nick,
Gabe,
Adam
and
Anna
for
their
support
with
this
effort.
Thank
you.
So
much.