Book notes: Sooner, safer, happier: Antipatterns and patterns for business agility.

These are my notes on Sooner, safer, happier: patterns and anti-patterns for business agility by Jonathan Smart with Zsolt Berend, Myles Ogilvie and Simon Rohrer.

The book is like The DevOps Handbook, but less technical and more org-wide.

Key Insights

• Better Value Sooner Safer Happier (BVSSH):
  • Better == quality.
  • Value is unique. OKRs.
  • Sooner != faster:
    • Flow efficiency.
    • Lead time.
  • Safer: governance, risk, compliance.
  • Happier: none of the previous "at any cost".
• "It is not my problem. I did my bit. The hole is on the other side of the boat".
• SW binary is agile-created (complex) and the path to production is lean (complicated).
  • Milestones make no sense on emergent (complex) domains.
• Not "the business" but "our business".
• Imposing Agile is not agile.
• Edgar Schein: learning only happens when survival anxiety is greater than learning anxiety (fear of being a noob again).
• "Why" for change must appeal all primary motivators (autonomy, mastery, purpose).
  • You cannot over-communicate the why.
• Change is a social activity.
  • Invite over Inflict.
• Orgs are complex adaptive systems:
  • Any experiment that takes place in them are not experiments because you cannot undo them.
• You are never done improving.
• “We don’t have time. We need to go big.” is a fallacy and self-fulfilling prophecy.
• Cognitive overload: willpower is a limited resource, and when it runs out you fall back on habits.
• As more teams adopt the new ways of working, the impediments will be more systemic and harder to resolve.
• Pattern: Descale Before you Scale.
• Joseph Campbell: if the path ahead is clear, you are probably on someone else’s.
  • Impediments are not in the path; impediments are the path.
• Improvement over time is more important than the absolute value.
• Silence is unhealthy.
• Instead of reporting lines, supporting lines.
• People need to be recognized for the lowest cost and quickest time to failure.
• Water-scrum-fall:
  • Urgency paradox: ideas sit 12-18 months on big upfront prioritization, and when they hit the dev team they suddenly become urgent.
• Value streams are not oriented around customer personas.
• Funding is a constraint on which to maximize value.
• Applications and puppies have something in common: they are for life, not just for Christmas.
• Install Jira and Jenkins and you're agile.
• Don't manage dependencies, break team. > Ignorance is the single greatest impediment to throughput. Dan North at Introducing Deliberate Discovery
• Following the plan == not learning. > Humans need to belong to groups, but the tribal instinct is also an instinct to exclude Amy Lynn Chua at Political Tribes
• When a measure becomes a target, it ceases to be a good measure.
• Tacit knowledge can only be shared by socialization (pairing, face-to-face) and a bias for action.
• Learning is more effective when exercises are in the learner context.
• There is no such thing as best practice. Your context is unique.

List of Patterns and Anti-Patterns

Pattern	Anti-pattern
Focus on Outcomes	Doing an Agile Transformation
Start with Why; Empower the How	Using Old Ways of Thinking to New Ways of Working
Achieve Big Through Small	The Bigger the Capital “T” Transformation, the Bigger the Change Curve
Descale Before you Scale	Scaling Agile before Descaling Work
Scale Agility, not Agile, Vertically then Sideways	Grass Roots Hits a Grass Ceiling
Not one size fits all	One Size Fits All
Invite over Inflict	Inflict over Invite
Leaders Go First	Do as I say, not as I do
Psychological Safety	Psychological unsafe
Emergent Mindset with Servant Leadership	Deterministic Mindset
Optimize for Fast End-to-End Flow	Local Optimization
Outcome Hypothesis	Milestone-driven Predicted solutions
Intelligent Flow	Headless Chicken
Stop Starting, Start Finishing	Start Starting
Safety within Safety	Lack of Safety within Safety
Organize Safety by Value Stream	Role-Based Safety Silos
Intelligent Control	Fixed Mindset to Risk
Go Slower to Go Faster	Going Faster Leads to Going Slower
Continuous Technical Excellence	Agile Hollow Shell
Architect and Organize for Flow	Misalignment of Teams and Architecture
Smart People and Smart Teams with Robot Friends	Tools Over People
Optimize for Learning	Information and Learning Silos
Nested Learning with Built-in Feedback Loops	Output over Outcomes
Communicate, Communicate, Communicate	The Bubble Effect
Be Comfortable with Uncertainty	Applying a Deterministic Approach to an Emergent Domain
Measure for Learning	Weaponized Metrics

TOC

• A Sense Of Urgency
• How We Got Here
• Chapter 1 - Focus On Outcomes: Better Value Sooner Safer Happier
• Chapter 2 - Achieve Big Through Small
• Chapter 3 - Optimization over One Way; Invite over Inflict
• Chapter 4 - Leadership Will Make It or Break It
• Chapter 5 - Build the Right Thing; Intelligent Flow
• Chapter 6 - Building the Thing Right; Intelligent Control
• Chapter 7 - Continuous Attention to Technical Excellence
• Chapter 8 - Create a Learning Ecosystem
• Chapter 9 - The Best Time to Plant a Tree is Twenty Years Ago; the Second Best Time is Now

A Sense Of Urgency

• There is no such thing as best practice. Your context is unique.
• Better Value Sooner Safer Happier (BVSSH):
  • Better == quality.
  • Value is unique.
  • Sooner != faster.
  • Safer: governance, risk, compliance.
  • Happier: none of the previous "at any cost".
• Practices = principles + context.

How We Got Here

• Maximize outcomes with minimal output.
• Doing Agile vs being agile.
• "It is not my problem. I did my bit. The hole is on the other side of the boat".
• SW binary is agile-created (complex) and the path to production is lean (complicated).

Chapter 1 - Focus On Outcomes: Better Value Sooner Safer Happier

• They apply to the whole org:
  • Not "the business" but "our business".
• Value == OKRs.
• Sooner:
  • Flow efficiency:
    • work time / end to end lead time.
    • Typically, 10% or lower.
  • Lead time:
    • Measure the 85% percentile and its change over time.
    • Throughput:
      • Items per period.
      • Should increase when reducing lead time.
• Values balance each other.
• Cheaper == anti-pattern:
  • Increases hidden costs via a reduction in flow efficiency.
• Anti-pattern: Doing an Agile Transformation
  • Reasons to try an Agile Transformation:
    • “We are being disrupted. If we don’t change we won’t survive”.
    • Issues:
      1. Optimizes for the agile tool, not the outcomes.
      2. Cargo cult. “Doing agile”.
• Anti-pattern: Using Old Ways of Thinking to New Ways of Working.
  1. Imposing Agile is not agile:
     • Top-down, command-and-control.
     • Big upfront planning.
     • Deadlines.
  2. Emotional reactions:
     • Fear and resistance.
     • Loss aversion.
     • Agentic state:
       • Blindly obey orders, passing the responsibility for the consequences of the person giving the orders.
  3. Remove intrinsic motivators:
     • Autonomy: no choice on Transformation.
     • Mastery: new thing, hence becoming a beginner.
     • Purpose: If the reason for AT is reducing cost or increasing benefits.
• Pattern: Focus on Outcomes.
• Pattern: Start with Why; Empower the How:
  • Simon Sinek: people don’t buy what you do, they buy why you do it.
  • Edgar Schein: learning only happens when survival anxiety is greater than learning anxiety (fear of being a noob again).
  • “Why” for change must appeal all primary motivators.
  • You cannot over-communicate the why.
  • Orgs are complex adaptive systems:
    • Any experiment that takes place in them are not experiments because you cannot undo them.
  • Start with:
    1. Identify early adopters by running a voluntary Community of Practice.
    2. Create a central Ways of Working Center of Enablement.
• You are never done improving.

Chapter 2 - Achieve Big Through Small

• Anti-pattern: The Bigger the Capital “T” Transformation, the Bigger the Change Curve.
  • Change curve:
  • “We don’t have time. We need to go big.” is a fallacy and self-fulfilling prophecy.
  • Cognitive overload: willpower is a limited resource, and when it runs out you fall back on habits.
• Anti-pattern: Scaling Agile before Descaling Work.
  • Parkinson’s law.
  • Also Parkinson: bureaucracies expands at a rate of 5-7% each year irrespective of the variation in the amount of work.
    • Policies, standards and controls expand with the number of control staff employed.
  • Scale of the org itself can be an impediment for BVSSH.
• Anti-pattern: Grass Roots Hits a Grass Ceiling.
• Pattern: Achieve Big Through Small:
  • People have a limited capacity to unlearn and relearn.
  • The only way to determine the optimal course of action is by safe-to-learn doing.
  • Rule of one:
    • 1 experiment.
    • 1 customer or team.
    • 1 location.
    • In production.
• As more teams adopt the new ways of working, the impediments will be more systemic and harder to resolve.
  • Don’t continue until these are alleviated.
• Adoption as an S-curve:
  • One per business unit and key systemic enabler.
  • One small WoW CoE per business unit, reporting to the business unit, not the central CoE.
• Pattern: Descale Before you Scale:
  • Descale work and system of work.
  • An org is a network of interdependent services:
    • Descaling should include breaking dependencies.
• Pattern: Scale Agility, not Agile, Vertically then Sideways.
  • New middle management role:
    • Ensure that the mission and desired outcomes are clear, then listen, coach and remove impediments.
    • Coaching Kata.

Chapter 3 - Optimization over One Way; Invite over Inflict

• Anti-pattern: One Size Fits All
  • Joseph Campbell: if the path ahead is clear, you are probably on someone else’s.
  • Your scaling and cultural context is unique.
    • And culture will vary between teams, business units, geography,...
• Anti-pattern: Inflict over Invite.
• Pattern: Not one size fits all:
  • Unique VOICE:
    • Values and principles:
      • Behavioral guardrails that apply across context.
    • Outcomes and purpose:
      • Improvement over time is more important than the absolute value.
      • Targets should be avoided as they invite cargo cult and gaming the system.
    • Intent-based leadership:
      • High autonomy with high alignment.
    • Coaching and support:
      • Coach should be omnists: recognize and respect all bodies of knowledge without fundamentalism.
    • Experimentation:
      • Complex system: probe, sense, respond.
      • Edgar Schein: you cannot understand a system until you try and change it, and when you do try to change it, only then will the underlying mechanisms maintaining the status quo emerge.
• “No”, “Not yet” or “What do we stop in order to do this?”
• Impediments are not in the path; impediments are the path.
• Visualize (flow metrics), stabilize (limit WIP), optimize.
• Evolution or revolution (by invitation), depending on your context.
  • Only Kanban and Disciplined Agile support evolution.
  • Revolution if org is facing an existential crisis, there is support for a bigger dip in the change curve, there is sufficient personal psychological safety, people are volunteering and there is prior experience.
• Pattern: Invite over Inflict:
  • Improving BVSSH must be communicated as high priority from the board, and that taking action will lead to recognition.
  • Exemplary Community:
    • Voluntary.
    • Agree to seek to become exemplary.
    • Benefits:
      • Access to external speakers and experts.
      • First to access new tools.
    • Keep the innovators on the bleeding edge.
    • Members have x20 improvement on most measures.

Chapter 4 - Leadership Will Make It or Break It

• Leader: guiding on a journey, following is optional.
• Commander: control, obey is mandatory.
• Anti-pattern: Do as I say, not as I do:
  • Leaders go on the same journey:
    • Become a beginner, trying something new, failing.
• Anti-pattern: Psychological unsafe:
  • Boeing as a bad example.
  • Cost and schedule should not be the primary focus.
  • Leaders’ actions and ability to provide or withhold rewards communicate their preferences, which then become the preoccupation of the workforce.
  • Silence is unhealthy.
• Anti-pattern: Deterministic Mindset:
  • All non-trivial change and product development is emergent.
    • The notion that we can predict what is going to happen is […] wasteful, demoralizing and dangerous.
  • Can be caused by leadership being uncomfortable exhibiting vulnerability: not knowing.
• Instead of reporting lines, supporting lines.
• Pattern: Leaders Go First:
  • Transformational Leadership (vs transactional), from James Burns’ Leadership:
    1. Role model.
    2. Vision.
    3. Intellectual stimulation.
    4. Coach: care for individuals.
• Pattern: Psychological Safety:
  • People need to be recognized for the lowest cost and quickest time to failure.
  • Steps to build psychological safety:
    1. Set the stage:
       • Not personal incompetence but of the system of work.
       • Failure is learning.
    2. Invite participation.
    3. Respond productively.
• Pattern: Emergent Mindset with Servant Leadership.

Chapter 5 - Build the Right Thing; Intelligent Flow

• Anti pattern: Local Optimization:
  • Water-scrum-fall:
    • Urgency paradox: ideas sit 12-18 months on big upfront prioritization, and when they hit the dev team they suddenly become urgent.
• Anti-pattern: Milestone-driven Predicted solutions:
  • Milestones with RAG status are culturally toxic:
    • Red is shame and reprisal.
  • Milestones make no sense on emergent domains.
• Anti-pattern: Headless Chicken.
  • Feature factories disconnected from customer needs and company strategy.
  • Teams incentivized by output.
• Anti-pattern: Start Starting:
  • Reducing WIP increases the chances of identifying causality.
• Pattern: Optimize for Fast End-to-End Flow:
  1. Identify value-streams:
     • Value streams are not oriented around customer personas.
  2. Map products to value streams. Identify:
     • Duplication.
     • Monolith that straddles multiple value streams.
  3. Create small long-lived teams:
     • T-shaped, multidisciplinary.
     • Roles:
       1. Value Outcome Lead (VOL):
          • The What (value).
          • Outward focus.
       2. Team Outcome Lead (TOL):
          • The How for the system of work and people.
          • Inward focus.
          • Alleviate impediments, build continuous improvement as a daily habit, coach.
       3. Architecture Outcome Lead (AOL):
          • The How of the tech implementation within the org’s broader tech architecture and engineering principles and standards.
          • Solutions architect == hands on.
          • Enterprise architect provide support and governance across value streams.
       4. These roles at every level.
       5. Top level:
          • CEO = VOL.
          • CIO = TOL.
          • CTO = AOL.
  4. Funding the Flow of Value:
     • Fund by value streams, not by project.
     • Funding is a constraint on which to maximize value.
• Pattern: Outcome Hypothesis:
  1. Emergence: Toyota Kata.
  2. Nested outcomes: OKRs.
     • Outcome do not cascade, they align.
     • Levels:
       • Strategic objective (multi year).
       • Portfolio objective (<3 years).
       • Portfolio epic (<12 months).
       • Business outcome (<3 months).
       • Experiment (<1 month).
  3. Business Outcome: > Due to <this insight> > We believe that <this bet> > Will result in <this outcome> > We will know that we are on the right track when: > Measure 1,2,3,...: quantified and measurable leading or lagging indicators.
  4. Rolling roadmaps and fixed dates:
     • Planning is done continuously at the multiple nested cadences.
  5. From Project Management Office to Value Enablement Teams:
     • Coaches and supports measuring business outcomes.
     • Assist with prioritization.
     • Coaches the limiting of WIP.
     • Ensures alignment of nested outcomes.
     • Helps collecting lagging and leading data.
     • Consolidates data from value streams.
• From “is it done yet?” to “what value we observed, what we have learned, and what steps shall we take next to get closer to the desired outcome?”
• Pattern: Intelligent Flow
  • Horizontal nested value streams with the vertical nested outcome hierarchy == high alignment.
• Pattern: Stop Starting, Start Finishing
  • Reduce WIP.
  • Pull, don't push.

Chapter 6 - Building the Thing Right; Intelligent Control

• Anti-pattern: Lack of Safety within Safety:
  • Safety (governance, risk, compliance) people scared of raising concerns due to blame culture.
  • Lack of psychological safety in the Safety teams.
• Anti-pattern: Role-Based Safety Silos:
  • Each safety specialization (Infosec, Data Privacy, Fraud, ...) being its own silo.
    • It causes:
      • Duplication.
      • Information bubbles.
      • Infight between silos and silos vs delivery teams.
• Anti-pattern: Fixed Mindset to Risk:
  • Risk management (current/future threads) vs compliance (past issues).
• Pattern: Safety within Safety.
• Pattern: Organize Safety by Value Stream:
  • Cross-functional, long-lived safety teams aligned to value streams, with a Centre of Excellence for support, plus a risk catalog.
• Pattern: Intelligent Control:
  1. Product teams engage with their safety teams on a continuous basis.
  2. Safety teams reassess on the quarterly business outcome the frequency of engagement, depending on those business outcomes.
  3. Risk stories validated through automated test, or by Value Outcome Lead, or by Safety team.

Chapter 7 - Continuous Attention to Technical Excellence

• Applications and puppies have something in common: they are for life, not just for Christmas.
• Constant Improvement (Kaizen) and occasional step changes (Kaikaku): Punctuated gradualism.
• Anti-pattern: Going Faster Leads to Going Slower:
  • Features, and only features.
  • No time to follow good dev practices because:
    1. Lack of partnership with "Business".
    2. No time to apply lessons learned.
    3. System Entropy.
• Anti-pattern: Agile Hollow Shell:
  • Agile without technical excellence is Agile work management.
• Anti-pattern: Misalignment of Teams and Architecture.
  • Team topologies.
• Anti-pattern: Tools Over People:
  • Install Jira and Jenkins and you're agile.
  • Should be: People -> Process -> Tooling.
  • No technical career path for hands-on software developers.
  • Not having testers: too much automation.
• Pattern: Go Slower to Go Faster:
  • Visualize flow of features, defects, risks, debts.
  • Prioritize improvement of daily work over daily work.
  • Coach for technical excellence/practices is much rarer than "Agile process coach".
• Pattern: Continuous Technical Excellence.
• Pattern: Architect and Organize for Flow.
  • Cognitive Load.
  • Don't manage dependencies, break team.
• Pattern: Smart People and Smart Teams with Robot Friends:
  • Technical career, good example: UK Government digital services career framework.

Chapter 8 - Create a Learning Ecosystem

• Anti-patterns: Information and Learning Silos: > Up to 50% of info is lost on every hand-off. Poppendieck at Implementing Lean SW Dev.

Ignorance is the single greatest impediment to throughput.
Dan North at Introducing Deliberate Discovery

• Anti-pattern: Output over Outcomes:
  • Learning whether we have delivered something valuable or not comes from the customer.

• Anti-pattern: The Bubble Effect:

  Humans need to belong to groups, but the tribal instinct is also an instinct to exclude
  Amy Lynn Chua at Political Tribes

  • Silo mentality: not want to share with other tribes.
  • Poor learning retention: when people leave, their learnings are lost.
  • Duplication.

• Anti-pattern: Applying a Deterministic Approach to an Emergent Domain:

  • Following the plan == not learning.

• Anti-pattern: Weaponized Metrics:

  • When a measure becomes a target, it ceases to be a good measure.

• Pattern: Optimize for Learning:

  • Unlearn.
  • Tacit knowledge can only be shared by socialization (pairing, face-to-face) and a bias for action.
  • In Clear and Complicated domains: explicit knowledge.
  • In Complex and Chaotic: tacit knowledge.
  • Identify and break silos with Value Stream Mapping.

• Pattern: Nested Learning with Built-in Feedback Loops:

  • Learning is more effective when exercises are in the learner context.
  • Levels:
    1. Individual:
       • Sharon Bowman four Cs, from Training from the Back of the Room:
         • Connections: Learners connect to the topic exploring prior experiences and knowledge.
         • Concepts and concrete practices: Self-learn.
         • Conclusions: Learners do self-reflection on the learning experience and teach back to the class.
    2. Team:
       • Stand-ups, weekly reviews, conferences, unconferences, retros, show and tell, pair and mob programming.
       • Toyota Improvement Kata.
       • Dojos.
    3. Organization:
       • Internal meet-ups, conferences, unconferences, webinars, brown-bag sessions, show-and-tell.

• Pattern: Communicate, Communicate, Communicate:

  • Internal BVSSH Awards:
    • Team has to proved factual data and quotes from customers and colleagues.
    • Teams will connect with winners for help and advise
  • Communities of Practice:
    • Voluntary, open to all, Darwinian.
    • Meet once every two weeks.
  • ASREDS loop: Align, sense, respond, experiment, distill, share.

• Pattern: Be Comfortable with Uncertainty.

• Pattern: Measure for Learning.

Chapter 9 - The Best Time to Plant a Tree is Twenty Years Ago; the Second Best Time is Now

• How to continue from your current starting point and org memory to get better at fast learning in order to optimize for outcomes across the org?:
  1. Start with Why (why of change, not why agile).
  2. Focus on outcomes: BVSSH.
  3. Leaders go first.
  4. Create a Ways of Working Center of Enablement.
  5. Start small and create an s-curve of change.
  6. Invite over Inflict: Start with the natural champions.
  7. Involve everyone from top to bottom.
  8. Bias to action: communicate, communicate, communicate:
     • Change is a social activity.
  9. Become a Relearning org:
     • You are never done.
     • It takes years.