Vous êtes sur la page 1sur 23

Spotify Rhythm

How we create focus


Agile Sverige, 2016-06-01
Consultant
www.crisp.se Henrik Kniberg
henrik.kniberg@crisp.se Dad
@HenrikKniberg

Organizational
Refactorist

Author
2000

# of employees
1500

1000

500

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Henrik Kniberg
Fall 2014 research: Whats slowing us down?

Misalignment! WIP
overflow!

Henrik Kniberg
This doesnt scale

WTF!

#@%&?!
Huh?

?!?!

Hey!!

Henrik Kniberg
Misalignment WIP overflow
WTF! You guys are
building a TUNNEL?
WTF! You guys are
building a BRIDGE?

Henrik Kniberg
Goal: Clear priorities, Data-driven decisions, Learning loop

Were working on X.
Because we think its
going to give impact Y,
Were working on X.
which matters to
Because Sam said its
Spotify because of Z
important.

Were done when the


metrics have moved
Were done when Sam
is OK with it.

Henrik Kniberg
Spotify
Rhythm
OKR P&A

Objectives & Key Results Priorities & Achievements

Henrik Kniberg
Henrik Kniberg
Visualize the Big Bets
Innovate on
C Build
Build new infrastructure
Launch in product for D
Country A capability B

Henrik Kniberg
High-Medium-Low = a broken way of prioritizing
Innovate on
C Build
Build new infrastructure
Launch in product for D
Country A capability B

HIGH
prio

MEDIUM
prio

LOW
prio

Henrik Kniberg
Stack rank North Star goals
Company
Launch in Beliefs
Country A
1
Only ONE thing gets X
to be priority one! active
users
Innovate on
C 2

If we only can do
Build
one of these two infrastructure 3
World
class
things, which one for D employer

would we do?
Build new
product 4
capability B
Why?
5

Henrik Kniberg 8
Dont do it all at once

Now Next Later

Launch in Build new


Country A
1 product 1 1
capability B

Innovate on
C 2 2 2

Build
infrastructure 3 3
for D

Henrik Kniberg
The Bets Board is a Google Spreadsheet
Open to everyone at the Spotify

Henrik Kniberg
Now Next Later

Each bet needs a 2-page brief Launch in


Country
1
Build new
product
capability
1 1

(google doc)
A B

Innovate on
C 2 2 2

Build
infrastructure 3 3
for D

DIBB
Lead team sponsor Road Manager
Data Insights Belief Bet

Key stakeholders Success metrics

Related bets .........

Henrik Kniberg
DIBB an argument framework
Example:
(Something like this happened a few years ago)

Data Insight Belief Bet


How people listen to music
Mobile is overtaking
Desktop u
sers desktop as primary WTF were Hire a bunch of
music gadget! optimized for the mobile devs
wrong thing!
Train a bunch of
Mobile use
rs For long term our desktop devs
survival, we need into mobile devs
Time We have very few to become
mobile devs mobile-first Build
compared to infrastructure for
How were staffed
desktop iterating fast on
Mobile devs
mobile

Feedback loop
Desktop devs

Henrik Kniberg
More Bet Boards started popping up
Tribe X Bet Board

TPD Bets Board Now Next Later

(Tech-Product-Design) 1 1 1

Now Next Later

Company Bets Board


2 2 2

1 1 1
3 3

Now Next Later


2 2 2
Build new
Launch in 1 product
capability
1 1
Country
A B 3

Innovate on
C 2 2 2 4

Build
infrastructure 3 3 Tribe Y Bet Board
for D
Now Next Later

Content,
4 1 1 1

Revenue, 2 2

Marketing, 3

etc 4

Henrik Kniberg
Bet = alignment point
#1 Bet

Sony
Playstation
Integration

Henrik Kniberg
Stack-ranked bets enable cross-company alignment
#1 Bet
Sony
Playstation
Integration

#2 Bet

Innovate
on C

Henrik Kniberg
Spotify Rhythm = cadence for deciding what to focus on
Strategy team every quarter

Company Bets Company Bets


Now Next Later Now Next Later

1 1 1 1 1 1

2 2 2 2 2 2

3 3 3 3

TPD prioritization group every 6 weeks

TPD Bets TPD Bets TPD Bets TPD Bets


Now Next Later Now Next Later Now Next Later Now Next Later

1 1 1 1 1 1 1 1 1 1 1 1

2 2 2 2 2 2 2 2 2 2 2 2

3 3 3 3 3 3 3 3

Henrik Kniberg
Now
Tempo markers
Launch in
Country A
1

= Moving as fast as it could


Innovate on
C 2

Build
= Moving slower than it could infrastructure 3
for D

= Blocked
5

Henrik Kniberg
What does Spotify Rhythm mean for squad autonomy?

Henrik Kniberg
Squads are (still) responsible for figuring North star goals

out how to make best use of their time Now


Company Bets
Next Later

- Bet Boards & DIBBs provide context TPD Bets


1 1 1

Now Next Later

Tribe Bets
2 2 2
1 1 1
Now Next Later 3 3

2 2 2
1 1 1

3 3
2 2 2

3 3

Hey, we need
your help

Hey, you need to upgrade


to the new build server

User data
Backlog IO

Henrik Kniberg
Take-aways
Agile techniques can be used to lead a Current Challenges:
company How does innovation
Autonomy only scales if leaders provide fit in?
high-level context How do we minimize
Simple tools = fast evolution waste / process
Use different time scales at different overhead?
levels
Keep experimenting
but dont forget to follow up!
VISION

Henrik Kniberg

Vous aimerez peut-être aussi