How organizations are learning to value learning and not just velocity!
We all hate wasting time and money. In the pursuit of cutting out waste, we've learned to systemically fool ourselves – to convince ourselves with very little evidence that the activities we're engaged in add value. And, further, activities that don't result in a product we can deliver are waste. But, the biggest leap of faith I continue to see most companies make is in believing that people want their new product, feature, or idea. They likely don't.
This talk is about the rise of learning as a valuable activity. I'll give examples of organizations that invest in experiments that take the cooperation of developers, testers, product mangers, infrastructure, sales, and marketing. At the end of these experiments organizations are left with no deliverable product and only the knowledge that the product they're thinking of should or shouldn't be built at all.
4. www.jpa:onassociates.com
If you’re not failing,
you’re not learning
-- Derek Sivers, sivers.org, you should watch the video
Why You Need to Fail: http://www.youtube.com/
watch?v=HhxcFGuKOys
41. Jeff
Pa:on
&
Associates,
jeff@jpa:onassociates.com,
twi:er@jeffpa:on
Stories
are
an
an+dote
to
“requirements”
So7ware
development
has
been
steered
wrong
by
the
word
‘requirement,’
defined
in
the
dicEonary
as
“something
mandatory
or
obligatory.”
The
word
carries
a
connotaEon
of
absoluEsm
and
permanence,
inhibitors
to
embracing
change.
And
the
word
‘requirement’
is
just
plain
wrong.
41
42. As a
I want
so that
www.jpa:onassociates.com
Agile
stories
are
about
the
future
world
we
imagine
63. Jeff
Pa:on
&
Associates,
jeff@jpa:onassociates.com,
twi:er@jeffpa:on
Get
out
of
the
building,
way
out
of
the
building
63
64. Jeff
Pa:on
&
Associates,
jeff@jpa:onassociates.com,
twi:er@jeffpa:on
Make
friends,
because
you
won’t
want
to
disappoint
your
friends
64
65. www.jpa:onassociates.com
Ateeq
has
an
epiphany
I’ve
always
been
confident
I
can
tell
you
precisely
what
users
do.
But
it’s
not
un9l
today
that
I
realize
that
I
could
never
tell
you
why.
72. If
you’re
really
good,
you’re
right
about
a
third
of
the
+me
www.jpa:onassociates.com
Your
product
decisions
are
more
likely
to
be
wrong
than
right
People
like
Marty
say
this
stuff
is
hard
(Marty
Cagan,
author
of
Inspired,
How
to
Create
Products
Customers
Love)
Typically
about
50%
to
80%
of
all
sobware
we
ship
fails
to
accomplish
its
objec+ves.
73. www.jpa:onassociates.com
Is
it
as
simple
as
building
only
the
features
people
will
use?
It
seemed
like
a
good
idea
at
the
Eme....
“Clippy”
-‐
Booed
off
the
Microso7
Office
stage
as
seldom-‐used
and
o7en
despised.
74. “There were plenty of weak
spots that led to Microsoft's
disastrous December quarter,
but one that didn't get much
attention Thursday was how
badly the Zune did.”
--Ina Fried, CNet News,
January 2009
opportunity:
integrated
music
management
and
portable
music
player
www.jpa:onassociates.com
It’s
only
aber
delivery
that
we
really
understand
value
75. www.jpa:onassociates.com
We’re
probably
right
about
2
Emes
out
of
10
We’ve
got
a
lot
more
flexible
architecture
-‐
easier
to
test
and
measure.
That’s
ow
we
know
that.
Eugene
Park,
Director
of
Product
Management,
Edmunds.com
78. www.jpa:onassociates.com
Nothing
leaves
the
board
unEl
there’s
been
a
discussion
on
what
we’ve
learned
Snag-‐a-‐Job’s
board
courtesy
of
David
Bi'enbender
Explicit
release
step
Explicit
measure
step
&
metrics
81. www.jpa:onassociates.com
Snag-a-Job’s architectural
journey is typical
(it just has a more entertaining
naming convention)
Thomas
Fredell,
Snag-‐a-‐Job’s
director
product
Our
current
placorm
allows
us
to
build
and
test
so7ware
with
users
faster
than
we
ever
could.
83. www.jpa:onassociates.com
Edmunds.com
optimizes its platform
around testing and
measurement
Eugene
Park,
Director
of
Product
Management,
Edmunds.com
We
built
our
CMS
with
AB
tesEng
in
mind.
We
used
to
mix
Test
and
Target
metrics
gathering
and
homegrown
metrics,
but
over
Eme
we
found
we
could
move
faster
with
our
own.
88. www.jpa:onassociates.com
Scale up prototypes to validate more.
Users say
they would
We can see
that they do
Eugene
Park,
Edmunds.com
Front-‐end
developers
can
quickly
build
and
deploy
tests
without
building
the
backend.
works reallooks real
90. www.jpa:onassociates.com
Edmunds
Price
Promise
involved
product,
development,
sales,
and
marke+ng
To
validate
Price
Promise
Edmunds.com
invested
months
of
Eme
and
lots
of
real
dollars
to
learn. Eugene
Park,
Edmunds.co
m
In
the
past
we’d
have
spent
months
researching
and
arguing.
Then
we’d
have
gambled
it
all,
or
not
tried
anything.
93. www.jpa:onassociates.com
“We had the misperception that the process, the
practice, or the methods will automatically
produce success.
Occasionally that’s the case, but most of the
time it isn’t.”
-- Eugene Park, Edmunds.com
“What process helps us to do is to not fool
ourselves”
-- Thomas Fredell, Snag-A-Job.com
95. WON’T GET FOOLED AGAIN
How
organiza+ons
have
evolved
to
value
learning
over
self-‐decep+on
Questions?
Yes, I’m joking. I’ve learned I never
leave time for questions. Please grab me
in the hall. I have lots of answers. Not
to YOUR questions... but still...
Jeff
Pa'on
jeff@jpa'onassociates.com
twi'er:
@jeffpa'on