Book notes: The Unicorn Project

#books

These are my notes on Gene Kim's The Unicorn Project.

Note that this is not your usual IT book, but a novel about it (or IT). Nonetheless there is a lot of practical advice, as Gene Kim is one of the authors of Accelerate and The DevOps Handbook.

Being a novel makes it a more enjoyable and easy read, one that even your pointy-haired boss could enjoy!

The Unicorn Project is a self-contained book, but related to The Phoenix Project, which I would also be happy to recommend.

Tip: Each chapter has a bunch of very interesting references!

Tip 2: Gene Kim did three "Ask Me Anything" on the IT Revolution Book Club. You can watch them here, here and here.

Key Insights

As Gene Kim is also the author of Accelerate, a lot of the key insights from Accelerate can be found in The Unicorn Project. See my Accelerate notes for completeness.

Is not small vs big, but fast vs slow.
Developer productivity should be the concern of the most senior developers.
To speak clearly, you have to think clearly. To think clearly, you have to write clearly.
Employee engagement and customer satisfaction are the only things that matter. The rest will take care of itself.
Five ideals:
- Locality and simplicity.
- Focus, Flow and Joy.
- Improvement of Daily Work:
  - Elevate improvement of daily work over daily work itself.
- Psychological Safety:
  - Safety is a precondition of work.
- Customer Focus.
People are trying to help. It is the process that is helpless. Reminds me of Gerry Weiberg's Becoming a technical leader.
When engineers think of "the customer" in the abstract instead as a real person, you rarely set the right outcomes.
CEO keeps reminding of who the customers are and the vision/mission.
Architecture is about enabling Focus, Flow, Joy and Fast feedback.
Transformational Leader:
- Expects excellence.
- Pursuit of perfection.
- Urgency to achieve mission.
- Dissatisfaction with status quo.
- Zeal for customers.
Geoffrey Moore's Three Horizons:
- Horizon 1 - Cash-cows. Needs bureaucracy.
- Horizon 3 - Future growth ideas. Needs speed.
Geoffrey Moore's Four Zones:
- Invest in Core, outsource Context.
- Rack funerals:

Image from Gene Kim's presentation The Unicorn Project And The Five Ideals.

Part One: September 3-September 18

Chapter 1

Mistakes and entropy are a fact of live.
Culture of fear (reference to Westrum org):
- Kills innovation.
- People are careful on how they phrase things.
- Avoid "forbidden words".
The Phoenix Project building is silent. Immediate sign that there is something wrong with the development teams.

Chapter 2

Maxine, the hero, keeps a daily work diary:
- What and how long work takes, and lessons learned.
Time tracking systems are useless.
When engineers think of “the customer” in the abstract instead as a real person, you rarely set the right outcomes.
CEO keeps reminding of who the customers are and the vision/mission.
Safety is a precondition of work.

Chapter 3

Busy != productive != effective.
People are trying to help. It is the process that is helpless. Reminds me of Gerry Weiberg's Becoming a technical leader.
Small changes == more productive.
Program state is the enemy.
Change is a fact of live.
Changes should be done easily, quickly, safely.
Agility is never free.
Focus, flow, joy. Fast feedback. Thanks to architecture.

Chapter 4

How can you create anything of value if you don't have feedback on how it is used?

Chapter 6

Too much information makes every thing slow down.
Coordination, communication, approvals maybe more expensive than redundancy and duplication.

Chapter 7

Dev productivity as a main theme of the book.
Complected. Sensei Rick Hickey.
Dev productivity should be the concern of the most senior devs.
Pay down technical debt as part of daily work.
First ideal: Locality and simplicity.
Second Ideal: Focus, Flow and Joy.
Third Ideal: Improvement of Daily Work:
- Elevate improvement of daily work over daily work itself.
Forth Ideal: Psychological Safety:
- Safe to talk about problems.
- Solving problems requires prevention, which requires honesty, which requires lack of fear.
- Ask "what caused a problem" instead of "who".
Fifth Ideal: Customer Focus
Technology is too often a caste system ( Arch > Dev > QA > Ops)

Chapter 8

Before: Leader roles was to design and decompose work and verify it was done by armies of interchangeable workers, who were paid to user their hands, not their heads.
Now: Leaders should be guiding, enabling and removing obstacles.
Toyota Andon cord: people are thanked for raising problems, so they can be solved and daily work improved.
Servant Leadership != Transformational Leadership.
Transformational Leadership:
- Understand the org vision.
- Question how work is performed.
- Inspirational communication.
- Personal recognition.
- Supportive leadership.
- Characteristics:
  - Expects excellence.
  - Pursuit of perfection.
  - Urgency to achieve mission.
  - Dissatisfaction with status quo.
  - Zeal for customers.

Chapter 11

You build it, you run it

Chapter 12

Lead time of ideas: 2 years
- Small features, rolled up to larger ones than then need to be approved and specced

Chapter 13

Blameless postmortems:
- Rule: you cannot say "I should have done" or "If I have known about that, I would have done Y".
Disconnect between Product Development and front-line users.
To speak clearly, you have to think clearly.
To think clearly, you have to write clearly.

Chapter 15

Monitoring FTW!
Graceful degradation is hard.
Synthetic traffic for Load testing is very hard.
Customer focus:
- Improve customer's live.
- Create value for them.
- Whether they will pay for it
  - If not, we shouldn't be building it.
Trust.

Chapter 16

Employee engagement and customer satisfaction are the only things that matter. The rest will take care of itself.
Only one in three strategic ideas have a positive outcome. And only a third of those have a big enough impact.
Goofrey Moore - Four Zones and Three Horizons:
- Horizon 1 - Company’s cash-cow.
- Horizon 2 - High growth. Next cash-cows.
- Horizon 3 - Pool of ideas to explore. Need to find asap:
  - Market risks: does the idea solve a real customer need?
  - Technical risk: Is it technically feasible?
  - Business model risk: Is there a financially feasible engine of growth?
Horizon 1 needs:
- Consistency.
- Compliance.
- Bureaucracy.
- Rules.
Horizon 3 needs:
- Speed.
- Experimentation.
- Fast feedback.

Chapter 17

Moore’s Four Zones. Core vs Context:
- Core: what the customer is paying for.
- Context: everything else. What is needed for operate (cafeteria, HR, email, …).
- Context is a burden.
- Invest in Core, outsource Context.

Chapter 18

Blameless post-mortem even without an outage: they are the best opportunity to learn.

Chapter 19

Teaching Thursdays is 2 hours.
Leaders must model the behaviours that they want.
Adults hide the fact they are learning because they don’t want to be seen as incompetents.
Event sourcing FTW!
"Rack funerals"