►
From YouTube: Digital Experience Retro - Mar 9, 2023
Description
Digital Experience Handbook: https://about.gitlab.com/handbook/marketing/digital-experience/
A
B
Yeah
I
just
wanted
to
say,
welcome
to
Trevor
we're
so
stoked
to
have
you
here
and
personally,
a
little
stoked
to
have
some
coverage.
We'll
Harrington
are
away
for
people
who
don't
know
Tina
and
I
have
both
separately
worked
with
Trevor
in
the
past
too.
So
we
know
he's
a
great
designer
and
a
great
person
and
we're
just
really
happy
to
have
you
here.
C
D
B
C
C
I
just
wanted
to
share,
like
I,
think
the
engineers
kind
of
see
like
the
end
shiny
product,
but
just
did
a
really
great
job,
navigating
a
lot
of
ambiguity
when
that
request
came
in
it
was
like
we
want
this
vague
idea
of
a
thing
and
just
tested
out
a
bunch
of
things
and
work
with
our
stakeholders,
and
now
we
have
this
like
great
calculator,
where
people
can
figure
out
if
premium
or
ultimate
is
worth
it
for
them.
So
I'm
super
excited
to
see
that
live
soon.
C
C
F
Yeah,
the
next
one
I
just
wanted
to
say
great
job
John
building
our
first
localized
page
in
the
buyer.
Experience
repository
got
that
German
sales
page
and
that
sales
form
is
linked
to
a
German
sales
funnel
in
Marketo.
It's
a
different
sales
form,
so
super
cool
and
then
also
big
thanks
Nathan
for
helping
us
figure
out
how
to
get
fastly
and
smart
lean
to
play
nice.
There
there's
definitely
some
ambiguity.
F
Okay,
we'll
move
on
to
the
next
one
things
to
improve
on.
We
were
supposed
to
launch
a
multivariant
nav
test
last
week
and
we
missed
the
deadline
and
therefore
are
not
running
the
test.
We
had
a
retrospective
issue,
there's
been
lots
of
activity
in
it.
So
thank
you.
Everyone
for
contributing
to
that
and
I
wanted
to
just
open
the
floor
for
discussion
and
help
us
figure
out
how
we
can
do
this
better.
F
F
So
we
built
the
Mr,
it
was
released
to
production,
it
was
broken,
didn't
work,
it
wasn't
collecting
the
data
for
the
A
B
test,
which
is
the
whole
point
of
the
A
B
test.
There
seems
to
be
some
blank
spots
for
how
we
verify
that
these
work
before
releasing
to
production
and
because
we
missed
the
deadline,
and
we
already
had
prior
commitments
it
we
just
had
to
forego
it.
C
And
just
add
some
context
in
terms
of
why
we
decided
to
not
launch
it
versus
like
waiting
to
launch
it.
The
test
that
we
were
doing
was
to
see
to
test
the
free
trial
CTA
on
the
header,
so
the
multivariants
were
one
was
our
control
one.
The
second
was
like
an
orange
red
color
and
then
the
third
was
a
gradient
with
a
gleam,
and
so
we
were
going
to
launch
that
test
and
conclude
that
test
and
then
go
forward
and
launch
the
new
home
page
illustrations
with
the
gradient.
C
H
I
I
was
gonna,
say
that
I
don't
know
if
if
it
could
work
or
if
it's
just
another
problem
but
maybe
having
a
an
environment,
that's
similar
to
production,
on
which
we
can
test
out
these
a
b
tests
and
then,
when
we
are
sure
that
it
is
working
in
an
in
an
environment
similar
to
production,
we
can
then
push
it
to
production.
Maybe
that
could
be
a
an
extra
step
that
we
could
take
for
a
b
tests.
I
I
It
looked
fine,
it
worked
as
best
as
it
could
in
the
review
app
we're
doing
all
these
Retros
to
to
get
to
the
bottom
of
it,
but
we
really
didn't
find
the
root
cause,
which
I
think
is
just
somewhere
like
there's
a
class
routine.
Maybe
that
the
navigation
package
and
the
test
I
have
no
idea.
But
it's
not
it
wasn't.
The
issue
wasn't
with
how
the
test
was
set
up,
because
that's
how
we've
set
up
tests
in
the
past
and
I
think
we
can't
really.
I
A
I
might
yeah
jump
in
I,
didn't
add
my
name
in
the
thing
I
posted
very
recently
in
the
retrospective
a
comment
that
I
think
there
were
two
parts.
So
the
technical
side
of
things
like
I
I
created
the
Mr
and
then
I
had
to
leave
so
Megan
was
the
reviewer
and
then
merged
it.
So
I
apologize,
you
know
I.
Obviously
there
was
some
misunderstanding
of
how
I
implemented
the
test
so
I
now
Megan
has
written
some
really
great
documentation.
A
I
didn't
realize
that
we
had
told
important
people
that
we
were
pushing
this
out
that
day
and
that
we
had
other
tests.
You
know
lined
up
to
to
go
forward,
so
you
know,
I
I
certainly
would
have
tried
to
get
that
out
sooner,
but
I
don't
think
that
that
allowed
for
an
adequate
amount
of
time
for
testing
and
figuring
out
these
bugs.
A
You
know
like,
like
Phil's
aside,
we
just
had
to
shut
the
test
down
and
move
on
so
trying
to
pad
that
time,
a
little
bit
more
I
get
that
it
was
a
friends
and
family
day
the
day
before,
and
then
we
needed
it
out
the
next
day.
So,
like
you
know,
just
a
timing
issue,
but
ideally
having
additional
time
padding
it
for
these,
these
bugs
that
are
inevitably
going
to
come
up.
A
It
sounds
like
with
the
the
startup
page
that
we
were
just
talking
about
like
a
similar
thing,
where
we
just
need
this
out
now
and,
like
you
know,
I
I
hope
that
we
can
get
to
a
point
where
we
pad
our
time
a
little
bit
better
for
and
fix
our
process
a
little
bit
more
but
but
yeah.
My
apologies
for
the
the
technical
side
of
things
for
sure.
D
Well,
I
wanted
to
say
that,
like
if,
if
the
valleys
were
collected
were
collected
properly,
we
still
didn't
solve
that
point
of
the
variance
not
displayed
so
like,
even
if
we
had
done
that
correctly
it's
like
well,
then
we
would
have
ran
into
it
anyway.
So
it's
like
in
that
sense.
Had
we
done
our
due
diligence
in
every
way
possible
like
it
still
would
have
failed,
so
that
I
think
that
to
me
that's
points
to
something
else:
that
side
of
our
control.
In
my
opinion,
foreign.
C
For
me,
I
think
part
of
the
message
is
the
communication
back
like
if
I
hear
that
hey
something
is
going
to
be
done
tomorrow,
if
it
doesn't
get
done
the
next
day,
most
of
the
time
it's
okay,
I
just
need
that
communicated
to
me.
So
then
I
can
go
and
adjust
other
plans
I'm,
not
as
bothered
like.
Oh
my
God,
it
didn't
go
out
and
like
it's
not
the
end
of
the
world,
there's
always
plan
BCD
and
I'm
happy
to
rearrange,
there's
certain
instances
where
that's
not
possible.
Obviously,
I
know
the
product.
C
Marketing
Group
gets
a
lot
of
that.
Where
there's
like
events
and
things
like
that,
where
it's
like
beyond
our
control-
and
we
need
to
get
things
out
when
we
need
to
get
them
out
but
I,
think
I
could
definitely
do
a
better
job
of
trying
to
provide
some
of
that
context
earlier
on
or
in
a
more
public
Channel.
C
Some
of
these
conversations
just
happen
one-on-one
or
in
different
meetings
that
I'm
a
part
of
so
I'll
from
my
end,
I'll
definitely
try
to
do
a
better
job
in
communicating
those
things
earlier
and
often
there's
that,
but
I
think
the
second
part
of
it
is
also
I.
Think
there
is
a
need
to
have
to
have
this
group
be
experts
in
a
b
testing
and
what
that
looks
like
I'm,
going
to
leave
to
Justin
and
Lauren
for
you
all
to
figure
out,
but
I
do
think.
C
This
is
something
that
we're
gonna,
Flex
more
and
more
with
our
okrs.
This
is
an
area
that
we
want
to
be
really
skilled
at
and
I,
don't
know
the
technical
issues
and
like
please
do
go,
spend
the
time
to
create
the
environment.
Do
what
you
need
to
do
to
be
the
experts
in
this,
because
we
need
to
be
relying
heavier
and
heavier
on
data
I.
Think
Dennis
is
doing
such
a
great
job
with
his
analytics
on
the
Google
front,
but
we
need
to
be
able
to
provide
more
of
that.
C
E
I
think
there's
like
this
is
a
we
didn't
do
a
lot
of
tests
for
a
while,
so
it
was
like
a
little
bit
of
rust
like
coming
off
like
make
sure
we
know
what
we're
doing
what
films
are
like,
we,
we
should
be
like
thinking
about
like.
Can
we
set
up
some
like
test
tests?
If,
if
our
limitation
is
that
review,
apps
may
not
work
like
we
know,
we
have
some
issue
with
Google
tag
manager
and,
like
certain
things,
not
firing
here
and
there
because
of
like
a
review
app.
E
Can
we
come
up
with
a
test
that
we
run
that
like
doesn't
really
affect
the
site
like
adding
an
extra
space
somewhere
we
can
make
sure
the
test
runs.
We
can
actually
click
data
and
be
like
cool.
You
know
how
to
work
it
on
this
page,
give
it
an
ABC
test.
Can
we
do
something
to
the
nav
that,
like
turns
off
a
hover
for
one
click,
No
One's
Gonna
know
about
it?
We
can
test
it
for
an
hour
or
two
make
sure
we
see
data
call
it
move
on.
E
A
That's
a
good
point
Nathan
that
maybe
we
should
have
on
the
recording
I.
Just
I
was
going
to
say
sorry
before
I
forgot,
I
was
gonna,
say
something
that
I
don't
know
at
what
point
like
I
agree
that
if
we're
going
to
be
running
a
lot
of
experiments,
we
should
be
like
really
really
good
at
experiments.
A
There
was
a
time
in
November
in
Palm
Springs,
where
Michael
said
we're
not
doing
AV
test
anymore.
Like
forget
experiments,
it's
never
happening.
I
I,
don't
know
what
changed
now,
we're
suddenly
doing
experiments
again,
like
that's
like
our
job
now
so
I,
don't
know
like
whoops.
You
know
I
kind
of
took
the
launch
Darkly
and
like
deleted
it
from
my
brain,
because
you
know
I
didn't
know
that
we'd
be
doing
them
anymore
and
now
here
we
are
but
yeah
I'm,
not
crazy
right.
That
happened
that
that
did
happen.
C
A
C
We'll
get
there,
that's
fair
I
dropped
from
my
memory
at
some
point,
I
see,
I
can
get
clarity
from
Michael
on
that
I
think
it's
I
can
see
in
the
context
in
which
you
may
have
said
that
I
think
he
was
Pro
like
there
are
definitely
things
that
we're
gonna
just
do,
but
I
do
think.
We're
also
gonna
need
to
test
certain
things.
This
test
idea
came
from
Michael,
so
I'm
a
little
confused.
J
Now
we
have
to
do
some
weird
click
handling
and
and
I
know
it's
it's
a
long
shot
like
there's
still
a
Blog
there's,
still
a
release
post,
there's
tons
of
stuff
on
www-
and
this
is
like
not
gonna,
happen
this
year.
But
it's
just
another
reason
that,
like
this
is
not
the
the
solution
but
anyway,
just
to
kind
of
say
that
it
it's
a
pain
to
deal
with
and
I
can't
imagine
trying
to
do
an
ABC
test
on
an
npm
package.
K
F
J
I'm
wondering
if
there
is
a
case
to
don't
quote
me
on
this,
but
I
feel
like
staging,
wouldn't
be
that
hard
to
get
back
up
and
running.
So
if
we
created
like
a
pipeline
job,
that's
manual
that
is
just
like
update
staging
and
then
staging
gets
updated
with
all
the
latest
prod
and
it
runs
on
a
gcp
bucket
with
the
same
configuration
and
fastly
as
they
about
docket.
A
D
C
Yes,
and
not
to
be
the
counterperson
to
that
hobby,
but
I
think
I
will
say,
gitlabs
quite
flexible
like
on
during
the
day
and
I
think
you
just
have
to
also
be
focused
on
like
absolutely
like.
If
you
worked
the
weekend
and
over
time
like
please
take
out
time
off
during
the
week
to
counterbalance
that,
but
it
is
just
also
making
sure
that
we
hit
our
business
results.
So
I
think
a
balance
to
that.
One
is
important.
B
I
think
to
that
point,
though,
like
the
problem
might
not
even
necessarily
be
people
working
overtime,
but
that
we're
accepting
jobs
that
are
like
two-day
turnover
and
that
also
kind
of
sets
a
bad
precedent
that,
like
any
job,
can
be
done
in
two
days
when
that's
not
really
realistic.
To
always
do
or
you
for
everyone
to
speak,
aren't
out
all
the
time.