Breaking the hierarchy How Spotify enables engineer decision making - - PowerPoint PPT Presentation

breaking the hierarchy
SMART_READER_LITE
LIVE PREVIEW

Breaking the hierarchy How Spotify enables engineer decision making - - PowerPoint PPT Presentation

Breaking the hierarchy How Spotify enables engineer decision making Kristian Lindwall, Spotify Introduction 3 Section name Our employees are our most valued resource So where does it go wrong? 6 Section name Lets do this! 7


slide-1
SLIDE 1

Breaking the hierarchy

How Spotify enables engineer decision making

Kristian Lindwall, Spotify

slide-2
SLIDE 2

Introduction

slide-3
SLIDE 3
slide-4
SLIDE 4

3

Section name

slide-5
SLIDE 5

”Our employees are our most valued resource”

slide-6
SLIDE 6

So where does it go wrong?

slide-7
SLIDE 7

Section name

6

Let’s do this!

slide-8
SLIDE 8

Section name

7

Yes… or maybe that! Let’s do this!

slide-9
SLIDE 9

Section name

8

Oh no…

slide-10
SLIDE 10

Section name

9

slide-11
SLIDE 11

3 ways to break the hierarchy

  • 1. Slice the pie and give people a piece of it
  • 2. Don’t tell people what to do
  • 3. Be intentional about how to distribute decisions
slide-12
SLIDE 12

#1

Slice the pie and give people a piece of it

slide-13
SLIDE 13

”We need teams of missionaries, not teams of mercenaries”

  • John Doerr
slide-14
SLIDE 14

App Integrations Home consumer electronics Team X Team Y

PPX TRIBE

Deepen relationships between artists and fans by extending Spotify into their favorite apps Filling homes with music that inspires every moment Tailor the Spotify experience to platforms and partners

CORE EXPERIENCE TRIBE

Define and evolve Spotify’s core experience Create the most delightful music experience

CONSUMER EXPERIENCE MISSION

slide-15
SLIDE 15

Squad Squad

TRIBE TECHNOLOGY TRIBE TRIBE TRIBE

TSG TSG TSG TSG TAG

slide-16
SLIDE 16

Principle #1

Build the org around autonomous mini-startups and enable them to align effectively

slide-17
SLIDE 17

#2

Don’t tell the organisation what to do, tell them what you want

slide-18
SLIDE 18

Plans Actions Results

Why didn’t people follow the plan? Why didn’t the plan work? Why can’t we plan better?

From: ”The art of action”, S Bungay (2010)

slide-19
SLIDE 19

Plans Actions Results

Allow people/teams to define HOW to do it Give people/teams freedom to adjust actions in line with intent Create alignment on intent: what & why

slide-20
SLIDE 20

Photo credit: Fifth Discipline Fieldbook by Peter Senge

slide-21
SLIDE 21

Spotify use ”DIBBs” to debate and decide

Objective point or set of points with information about the customer, product, world

Data Insight Bet Belief

A learning, theory, or conclusion drawn from one or more data points. A hypothesis built on one or more insights that offers direction to a team or teams. A decision to test a belief or set of related beliefs in practice by establishing it as a goal and dedicating resources to it.

slide-22
SLIDE 22

Example: Mobile free tier

  • Mobile sales increasing
  • Mobile streaming increasing
  • Mobile staffing low

Data Insight Bet Belief

  • Mobile is overtaking desktop
  • Our mobile team is disproportiantely small compared to desktop
  • We need to become mobile first
  • To be competitive, we need a free tier on mobile
  • Launch a new product, mobile free tier
  • Ramp up the mobile team
slide-23
SLIDE 23

RFCs: Common tool at Spotify to clarify intent, and debate your proposed solution

slide-24
SLIDE 24

Principle #2

Express intent, not actions. And get really good at arguing about what to do.

slide-25
SLIDE 25

#3

Be intentional about how to distribute decisions

slide-26
SLIDE 26

Decision making power Brain power utilization

slide-27
SLIDE 27

Decision making power Brain Power utilization

slide-28
SLIDE 28
slide-29
SLIDE 29

Ideas on how to distribute decision making

slide-30
SLIDE 30

Farnam street blog: https://fs.blog/2018/09/decision-matrix/

slide-31
SLIDE 31

Farnam street blog: https://fs.blog/2018/09/decision-matrix/

CONSEQUENTIAL NSEQUENTIAL IN INCONSEQUEN CONSEQUENTIAL IAL IR IRREV REVER ERSIBLE IBLE REVERSIBL ERSIBLE

  • Should I be

pairing on this task?

  • Can I merge this

pull request?

  • Shall we hire this

person?

  • Shall I introduce a

new data processing framework?

slide-32
SLIDE 32

High consequence decision

Example 1 Crafting strategy

slide-33
SLIDE 33
slide-34
SLIDE 34

High consequence decision

Example 2 Large scale reorg

slide-35
SLIDE 35

Section name

34

150 people

slide-36
SLIDE 36
slide-37
SLIDE 37

High consequence decision

Example 3 What technologies should we use at Spotify?

slide-38
SLIDE 38

Java

Spotify tech radar

slide-39
SLIDE 39

Principle #3

Isolate and move fast on low consequence, distribute high consequence decisions.

slide-40
SLIDE 40

Summary

slide-41
SLIDE 41

Optimize for autonomy. Enable alignment across the

  • rganisation.
slide-42
SLIDE 42

Don’t tell people what to do, tell them what you need.

slide-43
SLIDE 43

Be intentional about how you make decisions and invest in the

  • nes that matter.
slide-44
SLIDE 44

Kristian Lindwall @klindwall

Thank you!