spark 2016 - jason yip

48
An outsider assessment of Spotify Engineering Culture by an insider Jason Yip Agile Coach, Monetization Tribe, Spotify NYC [email protected] @jchyip https://jchyip.blogspot.com https://medium.com/@jchyip

Upload: spark-canada

Post on 13-Apr-2017

290 views

Category:

Business


1 download

TRANSCRIPT

Page 1: Spark 2016 - Jason Yip

An outsider assessment of Spotify Engineering Culture by an insiderJason Yip

Agile Coach, Monetization Tribe, Spotify NYC

[email protected]

@jchyip

https://jchyip.blogspot.com

https://medium.com/@jchyip

Page 2: Spark 2016 - Jason Yip

What I learned from Jerry Weinberg

Page 3: Spark 2016 - Jason Yip

https://flic.kr/p/eCfWC

Page 4: Spark 2016 - Jason Yip

What I learned from Dave Dame

Page 5: Spark 2016 - Jason Yip

It’s not the success that motivates but the willingness to show the struggle (paraphrasing)

Page 6: Spark 2016 - Jason Yip

What problem is Spotify trying to solve?

Page 7: Spark 2016 - Jason Yip

Everything comes down to “How do we deal with rapid growth?”

# of communication channels = n (n - 1) / 2

Page 8: Spark 2016 - Jason Yip

How might you maintain level up speed and quality even if you double in size?

Page 9: Spark 2016 - Jason Yip

The experience of working in Spotify Engineering is one of being free of friction, where everything makes sense, where you’re not surprised by strange decisions, where you can trust everything works as expected, and where there is a strong sense of safety that allows people to produce high quality outcomes, quickly, and efficiently.

Page 10: Spark 2016 - Jason Yip

The air war vs the ground war

Page 11: Spark 2016 - Jason Yip
Page 12: Spark 2016 - Jason Yip

“...those who master what venture capitalist Ben Horowitz calls ‘the black art of scaling a human organization’ act as if they are fighting a ground war, not just an air war.”

Page 13: Spark 2016 - Jason Yip

“Air war”

https://flic.kr/p/pbysvA

‣ Leadership buy-in‣ Messaging‣ Values‣ Etc.

Page 14: Spark 2016 - Jason Yip

‣ Individual buy-in‣ Day-to-day conversations‣ Day-to-day behaviour‣ Day-to-day routines‣ Policy‣ Etc.

“Ground war”

https://flic.kr/p/5ABvbF

Page 15: Spark 2016 - Jason Yip

Spotify has air superiority

Page 16: Spark 2016 - Jason Yip

The ground war is fought Tribe to Tribe, Squad to Squad, person to person

Page 17: Spark 2016 - Jason Yip

Here’s a taste of the ground war

Page 18: Spark 2016 - Jason Yip
Page 19: Spark 2016 - Jason Yip
Page 20: Spark 2016 - Jason Yip
Page 21: Spark 2016 - Jason Yip
Page 22: Spark 2016 - Jason Yip
Page 23: Spark 2016 - Jason Yip
Page 24: Spark 2016 - Jason Yip

‣ Speed of learning and execution lead us to create a best in class product

‣ The fewer technologies we are world class on, the faster we get‣ Shorter iterations yield faster learning, quicker value and higher

quality‣ Small empowered cross-functional teams (squads) is the

foundation of fast paced product development and innovation‣ Strong teams always beat rock stars‣ Your ability to learn and adapt is more important than your

knowledge, merit and experience

Spotify Engineering Core Beliefs

Page 25: Spark 2016 - Jason Yip

‣ Speed of learning and execution lead us to create a best in class product

‣ The fewer technologies we are world class on, the faster we get‣ Shorter iterations yield faster learning, quicker value and higher

quality‣ Small empowered cross-functional teams (squads) is the

foundation of fast paced product development and innovation‣ Strong teams always beat rock stars‣ Your ability to learn and adapt is more important than your

knowledge, merit and experience

Spotify Engineering Core Beliefs

How might this reliably

affect decisions?

Page 26: Spark 2016 - Jason Yip

‣ Role modeled by leaders and influencers‣ Supported by systems and policies‣ Communicated explicitly‣ Mentioned regularly

Good shared decision principles are:

Page 27: Spark 2016 - Jason Yip
Page 28: Spark 2016 - Jason Yip

‣ #technical-excellence○ Get more people talking about it

‣ “Fearless Client Development”○ Turn this into a prioritised TPD Bet

Existing initiatives to level up technical competence

Page 29: Spark 2016 - Jason Yip
Page 30: Spark 2016 - Jason Yip
Page 31: Spark 2016 - Jason Yip
Page 32: Spark 2016 - Jason Yip

A doesn’t quite work...

C depends on A...

Would be much simpler if F was done...

Page 33: Spark 2016 - Jason Yip

1. 2 week Feature Complete sync -> 1 week sync (2 week release cadence, 1 week sync cadence)

2. Feature Complete sync -> master is always releasable (aka Continuous Delivery)

Requires changes in ways of working and tooling support

From Release Trains to Continuous Delivery (client development)

Page 34: Spark 2016 - Jason Yip
Page 35: Spark 2016 - Jason Yip

When you can’t avoid the big project...

Page 36: Spark 2016 - Jason Yip
Page 37: Spark 2016 - Jason Yip
Page 38: Spark 2016 - Jason Yip
Page 39: Spark 2016 - Jason Yip

‣ Spotify Engineering Culture cannot be effectively decoupled from Spotify Product / Business Culture

‣ TPD Bets vs Company Bets

“Agile” must spread past TPD

Page 40: Spark 2016 - Jason Yip
Page 41: Spark 2016 - Jason Yip

It’s not Community vs Structure

Page 42: Spark 2016 - Jason Yip

Community

Structure

followsfollows

Page 43: Spark 2016 - Jason Yip

Summary: Recommended priorities

Page 44: Spark 2016 - Jason Yip

High effort

Low effort

Low Impact

High Impact

Change hiring criteria for non-Coach roles

More involved onboarding

Fundamentals training

Convert Engineering Beliefs into decision principles

Bet to level up client development

Ensure kickoffs and integration milestones for big projects

Agile outside of product delivery

Page 45: Spark 2016 - Jason Yip

Appendix

Page 46: Spark 2016 - Jason Yip

Who was that guy?

‣ Jason Yip‣ Agile Coach at Spotify since Feb 2015‣ Previously Principal Consultant at

ThoughtWorks (started in February 2001)‣ First encountered Extreme Programming in

1998

Page 48: Spark 2016 - Jason Yip

Thank you for your attention