►
From YouTube: Progressive Delivery UX/PM meeting 2020-09-01
Description
Meeting to improve alignment between Product Management and UX for progressive Delivery
B
A
What
this
issue
is
and
how
it
started,
so
we
just
completed
a
competitor
analysis
on
harness
and-
and
it
raised
some
really
interesting
questions
and
linking
this
back
to
our
previous
competitive
analysis
and
survey
that
we
did
with
spinnaker.
A
A
I
had
a
hard
time
formulating
the
questions
and
that's
where
nadia
came
in
to
help
me
because
the
question,
the
big
question
is:
why
are
we
losing
business
and
how
can
we
get
more
customers
to
use
cd
within
gitlab
and
not
go
to
external
tools,
but
going
specifically
into
some
of
the
their
approaches
to
cd
is
that
they
treat
cd
as
a
first
class
citizen
and
the
way
that
they
do?
That
is
with
their
several
different
methods,
specifically
for
harness.
There's
two
that
come
to
mind.
One
is
in
the
pipeline
view.
A
They
also
show
the
jobs
themselves
and
you
can
like
deep
dive
into
it.
It
gives
you
on
the
job
itself
like
full
context.
The
logs.
You
can
see
the
images
in
the
issue
itself
here,
but
you
can
see
like
the
logs
in
front
of
your
face.
You
can
see
what's
going
on
in
the
job,
you
can
see
metrics
associated
to
it.
So
it's
really
really
awesome.
A
So
that's
one
thing,
and
the
second
thing,
which
is
something
that
I
think
vitica
is
working
on,
is
that
they
allow
you
to
do
they
kind
of
say.
Cd
is
basically
the
same.
It's
like
a
template.
You
need
to
choose
if
you're
doing
canary,
if
you're
doing
a
blue
green,
if
you're
doing
some
kind
of
other
thing
and
then
they
just
let
you
you
know,
select
from
a
drop
down
what
you
want
to
do
and
enter
the
specific
details.
A
So
you
don't
have
to
build
everything
as
one
huge
gitlab
cieml
file
and
just
have
the
deploy
as
one
job
inside
the
the
ci
ammo.
So
the
question
is:
do
we
can
we
treat
cd
as
its
own
thing
and
not
just
a
continuation
of
ci,
and
maybe
these
two
approaches
is
the
nice
way
to
do
it?
So
that's
kind
of
the
background
to
the
specific
issue.
C
So
it
sounds
like
our
hypothesis
here
is
that
our
competitors,
spinnaker
and
harness
are
providing
a
better
I.e.
First-Class
citizen
experience
with
cd,
which
is
why
we
are
losing
our
our
cd
business
to
those
competitors.
That's
our
hypothesis
right!
So
what
we
need
to
do
is
we
need
to
go
out
and
prove
that.
So
we
want
to
understand
what
do
people
like,
because
just
because
they
do
that
doesn't
mean
that
they're
doing
it
well
or
doesn't
mean
that
there
isn't
room
for
improvement
in
what
they're
doing
so.
That's,
oh
yeah,
but.
A
C
Yeah
yeah,
I
know
I'm
not
saying
like
the
hypothesis
is
wrong
or
anything
it's
just
more
of
like
making
sure
that
we're
covering
it
here
in
the
discussion
guide.
Okay,
so
we're
at
so
we're
asking
the
questions
that
we
need
to
in
order
to
get
the
information
so
yeah,
I
know
totally
so.
We've
got
this
spinnaker
information,
and
now
it
sounds
like
we
want
to
focus
on
harness.
C
A
So
is
that
really
the
approach,
or
is
our
approach
like
just
regardless
to
our
competitors,
even
though,
like
we're
losing
business
to
them?
Obviously,
but
how
do
we
get
cd
to
be
its
own
individual
thing?
You
yourself
now
that
you
finished
the
ci
thingamajig,
which
I
don't
remember
the
name
you
were
like.
What's
the
difference
between
cis
ncd
and
nadia
was
like
looking
for
some
docs
on
how
to
start
with
cd
and
everything's
tied
together
with
ci,
it's
all
bundled
together.
A
B
Question
users
need
to
to
that,
like,
however,
yeah
like,
maybe
maybe
they
do,
because
ci
and
cd
are
really
different
and
different
types
of
organizations
are
using
each
of
them.
But
this
is
one
of
the
questions
that
I
would
like
to
like
ask,
and
this
is
maybe
why
it's
good
to
take
into
account
some
of
the
hardness
and
spinnaker
stuff
is
competition,
but
also
like
what
you
are,
or
they
are
saying,
is
that
I
would
like
to
understand
actually
for
myself
like,
like
do
people
understand
the
difference?
B
You
know
and
those
like
and
also
experience
of,
setting
up
like
pipelines,
you
know,
like
the
templates
that
you
were
mentioning,
I
remember
from
exhibit
labs
years,
that
those
templating,
where
those
help
with
configuring
the
pipelines,
what
big
huge
ideas,
even
in
those
years,
so
I
it
sounds
like
from
company
to
company.
That's
a
trend,
that's
something
that
people
will
think
that
people
need.
B
I
don't
know
if
that's
reality,
because
I
don't
remember
us
doing
any
meditation
in
those
years,
but
this
is
what
I
would
like
to
learn,
because
that
sounds
like
making
it's
very
simple
also
by
the
competitors,
data
that
I've
seen
they
have
really
nice
visual
ui
that
will
allow
you,
okay,
choose
the
type
of
the
deployments
that
you're
using
and
then
we
you
don't
even
actually
need
to
know
what
you
need
to
do:
ci
or
cd.
We
will
guide
you
through
that.
Maybe
we
should
not
even
bother,
but
you
know.
D
A
So
so
a
few
more
background
points
that
came
up
regardless
to
this
specific
issue.
So
one
of
them
is
the
fact
that
sometimes
it's
not
the
same
people
that
are
doing
the
ci
bit
and
the
cd
bit
so
the
cd
bit
is
deployment
right
and
deployment
is
production
and
production
is
sacred,
so
only
some
people
have
permissions
to
it,
but
ci
is
like
more
internal.
A
So
it's
for
internal
use
cases
and
therefore
the
permissions
schema
is
a
little
bit
different
and
when
I
talk
to
enterprises-
and
this
is
more
on
jackie's
side-
but
I've
been
to
my
share
of
conversations
where
this
came
up,
where
enterprises,
specifically
like
banks
and
those
kind
of
sectors,
are
looking
for
some
way
to
do
job
permissions.
D
D
A
C
D
That
say
that
you
have
three
repositories
and
each
repository
contains
a
yaml
file,
the
build
file,
the
test
file
and
the
the
deploy
file.
And
then
you
have
a
fourth
repository
that
includes
all
three
ammo
files
into
the
main
gitlab
ciemo
file.
Then
you
could
say
towards
the
test.
I
I
only
want
these
developers
to
have
access
to
that,
the
same
for
built
to
save
for
deploy
and
that
way
have
permission
management
as
a
workaround.
Basically,.
A
D
C
Like
making
it
a
first
class
citizen,
I
would
agree
that
we
need
to
do
that,
but
I
don't
want.
I
feel
like
there's
information
missing.
Just
like
dimitri
said
earlier.
We
need
to
know
more.
We
need
to
understand,
maybe
more
about
how
like,
if
I
wanted
to
separate
my
cd
and
my
ci,
because
they're
due
different
people
and
nobody
can
touch
production
and,
except
for
my
anointed
production
people
yeah,
I'm
going
to
use
a
different
program
because
they
only
have
access
to
that
program.
Is
that
the
case?
That's,
my
hypothesis.
C
C
Can
we
also
offer
them
the
same
capabilities
within
our
tool
instead
of
having
them
use
a
separate
tool
altogether?
So
I
think,
there's
still
merit
in
talking
with
people
who
use
competitive
competitors
in
understanding.
Why
like?
Why
did
you
choose
that
one?
Why
do
you
continue
to
use
that
one?
What
are
the
things
that
work?
Well
for
you,
with
with
this
particular
tool,
what
are
the
things
that
you
wish?
It
did
that
it
doesn't
do,
and
we
can
have
good
conversations
around
that.
C
We
can
also
talk
to
people
who
use
git
lab
either
use
it
for
cd
or
I
don't
know.
I
don't
know
if
we
should
talk
to
people
who
don't
use
it
for
cd
but
would
like
to,
and
maybe
they
use
a
competitor,
I'm
not
sure
that
seems
like
it
might
be
harder
to
find
those
people,
but.
D
C
B
Also,
that's
probably
a
question,
probably
I
don't
know
but
yeah.
It
also
sounds
to
us
like
lori,
you
said,
like
information
is
missing
yeah.
I
think
we
even
like
need
a
little
bit
of
problem
validation
like
what
exactly
are
we
trying.
D
C
Yeah
yeah
definitely
problem
validation
very
a
little
bit
more
generative
than
our
typical
problem
validation,
because
we
there's
more
to
uncover,
I
think
and
or
you're
probably
right.
You
know
everything
that
you've
said
you
and
dmitry
have
said:
you're,
probably
right,
but
let's
get
some
data
behind
it,
just
to
make
sure
that
we
are
on
the
right
track
and
have
the
opportunity
to
ask
more
questions
about
other
things
that
might
help
you
to
make
some
decisions
for
the
ui
and
future
wise
later.
A
Yeah,
I
definitely
agree.
That's
why
I
kind
of
asked
for
help
on
this
one,
because
I
think
I
don't
even
know
what
I'm
trying
to
uncover.
If
that
makes
sense
like,
I
know,
there's
something
to
uncover.
C
D
D
Going
too
fast,
let
me
let
me.
C
D
My
screen,
then
we
have
at
least
some
something
to
go
about,
should
be
this
one.
D
We
got
two
times
the
same
one
all
right,
so
I
found
the
to
start
at
the
top
a
little
bit
the
business
decisions
I
was
wondering
about
this,
like
this
seems
to,
from
my
perspective
point
at
like
all
right.
Is
there
an
opportunity?
Canvas
like
what
are
the
questions
we
have
from
a
business
perspective?
This
is
more
like
this
is
the
business
decision.
Why
we're
even
doing
this
like.
A
At
this
point,
because
we
kind
of
don't
even
know
the
questions
yet,
but
definitely
along
the
way,
I
think
the
business
decision
here
is:
we
have
competitors,
we're
losing
business
to
them.
We
want
their
money.
C
And
you
know
ori,
that's
a
good
point
about
the
opportunity
campus.
This
is
this
is
this
is
truly
generative
research
like
we
don't
really
know
what
we're
going
to
find
out.
So
it's
kind
of
hard
to
write
the
opportunity
canvas
at
this
stage.
I
could
see
after
we
do
this,
that
you'll
probably
have
everything
that
you
need
to
write
opportunity,
campus
yeah.
C
The
problem
is
some
we're
doing
something
wrong
or
not
enough
and
they're
going
to
competitors
right.
That's
that's
really
the
problem
and
we
just
have
to
figure
out
what
what
it
is
that
we're
not
doing
or
what
it
is
that
we're
not
we're
doing
wrong.
So
we
can
fix
it.
So
we
can
start
capturing
some
of
that
business.
D
Yeah
all
right,
so
it's
basically
kept
your
business
because
we're
losing
losing
business
like
we
want
to
recapture,
because.
D
A
I
have
one
thing
to
say
that
I
learned
a
lot
from
the
ci
experience
and
that's
that
we
found
out
that
anyone
that's
using
ci
in
git
lab
is
like
90
sure,
to
have
retention
next
year,
and
we
don't
have
that
same
number
for
cd.
So
I
want
to
make
sure
that
we
can
continue
on
stages,
increase
retention
based
on
whatever
comes
up
from
here
in
terms
of
the
business
I'm
not
sure
like.
D
And
then
yeah,
so
I
I
think
the
rest
is
less
business
related
right,
like
it's
more
like
all
right.
How
is
what
are
the
higher
hypothesis
in
terms
of
the
visual
versus
the
other?
You
know
like?
Do
we
need
to
separate
those
things?
I
think
that's
all
hypothesis
based
nothing
of
that
is
like
business
related.
As
far
as
I
can
see
it
all
right
onto
our
boxes.
D
Our
pipeline
view,
which
was
built
for
ci,
is
limiting
users
who
are
using
them
for
releases.
So
this
is,
I
think,
a
pretty
good
hypothesis,
but
I
was
saying
here
like
what
is
it?
What
is
the
difference
of
using
a
pipeline
for
release
versus
mix
usage
versus
only
ci
like
when?
Where
do
we?
Where
do
we
draw
the
line.
D
A
B
To
do
the
cd
today,
it's
like
it's
just
one
step
too
far.
They
want
to
like
be
manually
validating
what
they're
delivering
to
production.
So
some
companies
are
even
not
ready
to
do
cd
today
continuous
delivery,
like
they
don't
want
to
deploy
the
productions,
trade
so
yeah.
There
is
part
of
the
permission
question,
but
yeah.
It's
also
like
what
how
far.
A
Yeah
I'll
even
add
to
that
there's
a
lot
of
companies
out
there,
including
my
previous
one,
where
a
we
were
deploying
to
on
on-prem
and
not
to
cloud
so
continuous
delivery
didn't
really
exist.
What
you
did
was
you
created
a
build
and
you
uploaded
it
to
some
cdn
and
told
the
users
to
take
it
from
there.
So
it's
not
really
a
cd
and
I'm
sure,
there's
more
companies.
A
You
know
with
that
same
model
and
the
other
one
is
that
people
are
okay
with
automating
everything
as
long
as
it's
internal,
but
they
still
want
those
manual
checks
at
the
end.
So.
A
C
This
is,
we
should
just
have
like.
Maybe
one
would
be
good
or
or
maybe
two,
because
if
you
have
a
lot
more
like
you
should
have
a
goal
for
each
hypothesis
and
then
like
several
objectives
for
each
one
of
them.
So
it's
almost
like
we're
getting
really
specific
here
with
our
hypothesis
hypotheses,
where
we
don't
really
have
to
be.
That
specific
here,
like
the
hypothesis,
is,
if
we
add,
features
and
increase
the
ease
of
use
and
the
increase,
the
trust
there's
another
word
there.
C
In
our
cv
feature,
we
will
get
more
people
using
our
cd
feature.
That
is
our
hypothesis
right.
That
is
our
reason
to
go
and
find
out
what
are
other
people
doing
to
do
those
things?
How
are
they
increasing
trust
and
effectiveness
and
efficiency
in
in
their
cd
option?
And
how
can
we
do
that?
That's
all
we
really
have
to
say
here.
B
Two
sub
points
are
going
into
too
much
details
because
it
depends
on
the
type
of
the
work
we
talk
about,
which
would
be
which
we
will
learn
anyway
in
the
process.
D
Yeah,
but
I
I
mean,
like
those
like
it's
it's
nice-
to
have
a
general
hypothesis
that
is
encapsulating
all
the
things
that
we
want
to
know,
but
if
it
doesn't
go
into
specifics,
it's
harder
to
kind
of
ask
the
right
research
questions.
D
C
That's
why
that
doesn't
go
in
the
hypothesis,
it's
probably
in
our
goal
or
our
objective.
So
we
want
to
understand
you
know.
So
maybe
I
see
these
as
maybe
go
objectives
because
they
are
much
more
specific
with
the
pipeline
view
like
understand
how
our
pipeline
view
limits,
users
like
in
which
ways
so
they're
they're,
very
much
objectives,
though
not
not,
goals.
B
I
have
to
drop,
who
is,
who
will
be
the
dri
of
this
doc,
just
like
on
aligning
on
the
next
actions
or,
if
any,
help
needed
for
me.
Let
me
know
I
would
love
to
collaborate,
but.
A
I
think
I'll
take
it
since
dimitri
is
gonna,
be
off
for
a
month
anyway,
yeah
let.
B
C
Yeah
great,
I
can
help
too
just
let
me
know
what
you
need
and
if
you
want
emily
to
help
recruit
we'll
get
that
started
too.
Yeah
cool
cheers
all
right,
bye,
guys.