Understanding Platforms

What They Are, Why They Work, When to Use Them, How to Build Them

Hazel Weakly (@hazelweakly@hachyderm.io)
Understanding Platforms: What They Are, Why They Work, When to Use Them, How to Build Them

Agenda

Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

What is Doing?

Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

IC
Systems engineering through technical work
Manager
Systems engineering through people work
Exec
Systems engineering through systems work
Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

How Do We Do?

One of the main hurdles when building a company is figuring out how to scale decision making

Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

IC Manager Exec
Thriving Outcomes Alignment
Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

Relationships change fundamentally based on their scope
1 3-5 15 50 150 ...
self group 3 groups ...

Every combination is a different relationship and collaborates differently.

  • The effectiveness of a collaboration model is influenced by the scope, the context, and the motivation under which it's utilized.
  • Ideal collaboration happens when the ownership of the work, the motivation to do it, and the authority to complete it are all mutually aligned.
  • There's another set of factors here: Time, and Stability.
    (They're less relevant than you'd think)
  • We think of Time as teams having a life-cycle, but that's not really true in the traditional "Forming, Norming, Storming, Dying" sense.
  • We like stability because it's an effective compensation mechanism for working around bad management, poor alignment, lack of agency, overload, and context switching.
Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

  • Vertical

    • Specialized Knowledge
    • Effective action in the organization
    • Cognitive overhead required for doing
    • Utilizes consensus

    Horizontal

    • Collective Knowledge
    • Effective learning in the organization
    • Cognitive overhead required for collaboration
    • Utilizes alignment
Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

Scaling Decision Making

  1. Shape motivation
  2. Understand relationships
  3. Scope context
  • Building a stable base of knowledge is how you scale collective learning as a society
  • Ignorance Limited Context is a feature, not a bug
Hazel Weakly (@hazelweakly@hachyderm.io)
What is Doing?

IC
Systems engineering through technical work
Manager
Systems engineering through people work
Exec
Systems engineering through systems work
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

What is a Platform?

Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Platforms Are a Way to Scale Doing

Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

What is a Platform?

IC
Platforms are the Dual of Libraries
Manager
Platforms turn Service Teams into Product Teams
Exec
Platforms derisk diversification by exploiting the economics of agglomeration
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Totes Platformsâ„¢

  • Kubernetes
  • Design Systems
  • Programs
  • Education Curriculum
  • Corn
  • Sales Funnel
  • Jazz Standards
  • Mission, Strategy, Vision, Objectives
  • Executive Alignment
  • Policy
  • Language
  • APIs
  • ISO Standards
  • Algebraic Structures Equivalent to Computational Models
    regex is a platform. die mad about it.
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Platforms Are Not

These are tools and a means to an end

  • Platforms
  • DevOps
  • Kubernetes
  • Libraries
  • Process
  • Culture
  • Infrastructure
  • Golden Paths
  • SDKs
  • Self-Service Platforms
  • Cookbooks
  • Enabling / Platform Teams
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Platforms ≠ Platform Engineering

Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Libraries vs Platforms

Libraries

  • Representation of domain knowledge
  • Abstracts over implementation
  • Interfaced with via process
  • Codifies semantics
  • Multiplier for platforms

Platforms

  • Representation of knowledge application
  • Abstracts over integration
  • Interfaced with via policy
  • Codifies workflow
  • Enables library creation
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Platforms Turn Services into Products

This isn't always good.

It's not always what you want.

You will always need both services and products.

Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

Platform Synergy

IC
Libraries are the building blocks of Platforms
Manager
Platforms synergize with executive alignment and team dynamics
Exec
Platforms operationalize Libraries the way Programs operationalize Policy
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

DevOps vs
SRE vs
Platform Engineering

Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?
DevOps
A culture of whole-system ownership motivated by broadening horizontal context
SRE
A specialist role in resilience engineering that naturally arises when horizontal context becomes too large for any one team
Platform Engineering
An empathy-driven approach towards sociotechnical organizational design
Hazel Weakly (@hazelweakly@hachyderm.io)
What is a Platform?

What is a Platform?

IC
Platforms are the Dual of Libraries
Manager
Platforms turn Service Teams into Product Teams
Exec
Platforms derisk diversification by exploiting the economics of agglomeration
Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

What Makes Platforms Work?

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

What Makes Platforms Work?

IC
Abstractions that are force multipliers
Manager
Interfaces between teams that reduce cognitive overhead
Exec
Consistent language and mental models that enable aligned innovation
Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Mapping Organizational Structures to Doing

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Slicing Context - Functional

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Slicing Context - Product

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Slicing Context - Matrix

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Directional Context

Vertical

  • Specialized Knowledge
  • Effective action in the organization
  • Cognitive overhead required for doing
  • Utilizes consensus

Horizontal

  • Collective Knowledge
  • Effective learning in the organization
  • Cognitive overhead required for collaboration
  • Utilizes alignment
Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Context + Collaboration

Vertical ctx size Horizontal ctx size Collaboration Complexity Market Agility
Functional high low low low
Product moderate moderate moderate moderate
Matrix high high high high
Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Cheating the Matrix

If you "aren't" a matrix organization, but...

  • You have teams that rotate projects
  • "Dynamic" or "rotating" teams are a thing
  • You keep "shifting left" everything
  • You compulsively CC 5+ people on every email

You may be entitled to financial compensation

Call H. E. Weakly (877 RE-ORG)

  • Building a stable base of knowledge is how you scale collective learning as a society

    But, how do you build a base of knowledge if scaling requires managing context?

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Shrinking Context

IC
Well designed abstractions
Manager
Thoughtful interfaces between teams
Exec
Consistent language and mental models
  • Well designed abstractions are both transparent and opaque

  • Thoughtful interfaces between teams result in effective collaboration models

  • Humans chunk concepts together to build on top of them

    Providing consistent language and mental models allows people to build their own chunks in a way that results in aligned growth of understanding

Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

Properties of Good Platforms

  • Reduce horizontal context
  • Enable emergent innovation
  • Enable experimentation
  • Minimize migration cost
  • Minimize collaboration model overhead
  • Improve feedback loops
  • Minimize context switching
  • Reduce IC context overhead
Hazel Weakly (@hazelweakly@hachyderm.io)
What Makes Platforms Work?

What Makes Platforms Work

IC
Abstractions that are force multipliers
Manager
Interfaces between teams that reduce cognitive overhead
Exec
Consistent language and mental models that enable aligned innovation
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

When to Platform

Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

My Sincere Apologies to Finance Teams and MBAs

Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Economics ⇔ Technology

Economics Technology
Scarcity Economy Post-Scarcity Economy
Fiat currency Social currency
Market pressure Consensus
Markets Tech Stacks, Codebases
Fiscal Debt, Inflation Tech Debt
Economies SDLC, DevEx
Context Cognitive Overhead
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Economic Model - Functional

  • Economics of scale
  • Tech Vertical - often viewed as a cost center
  • Value generation comes from removing waste from feedback loop
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Economic Model - Product

  • Economics of scope
  • Tech Vertical - often viewed as a profit center
  • Value generation comes from reuse of overhead when diversifying
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Economic Model - Matrix

  • Economics of density
  • Often viewed as too much paperwork multiple centers
  • Value generation comes from efficiencies generated by peer proximity
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Continuity, Economics

Economy Growth Strategy Continuity Risk Market Agility
Functional Scale Market Development low low low
Product Scope Product Development high moderate moderate
Matrix Density Diversification moderate high high
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Platform Economics

  • Libraries mature market development
  • Platforms mature product development
  • Platform Engineering derisks effective diversification

    • Blocks can be seen as services
    • Groups of blocks can be seen as product offerings
    • Clusters are economic optimization opportunities
    • Nodes are teams
    • Branches can be seen as codebases
    • Branch clusters can be seen as problem domain understanding
    • Clusters are economic optimization opportunities
    • Nodes are teams
Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

When to Platform?

Platforms make sense when the benefits of reducing horizontal context more than offset the overhead of adding a vertical to the organization.

Hazel Weakly (@hazelweakly@hachyderm.io)
When to Platform?

Platforms must be more than the sum of their parts

Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

How to Build High-Impact Platforms

Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

The Blueprint: LOVE

  1. Listen
  2. Observe
  3. Validate
  4. Empathize
Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

Not the Blueprint: EVIL

  1. Examine
  2. Vindicate
  3. Implement
  4. Launch
Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

Love vs Evil

They're surprisingly similar, really

Love

  1. Listen
  2. Observe
  3. Validate
  4. Empathize

Evil

  1. Examine
  2. Vindicate
  3. Implement
  4. Launch
Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

Empathy

Perspective taking
Recognize other people's perspectives as their truth
Staying out of judgement
Suspend judgement and be curious
Recognizing emotion in other people
Recognize and acknowledge other people's emotions
Communicating emotion
Communicate your understanding of other people's emotions
Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

Step-By-Step Recipe

Agglomeration-based development

  1. Ask people where the pain clusters are
  2. Find the pain clusters
  3. Compare results
  4. Solve the most painful cluster
  • IC

    1. Ask individuals on your team what their pain points are
    2. Look for duplication of effort, code, process, communication
    3. Compare results
  • EM

    1. Ask teams you interface with what their pain points are
    2. Look for duplication of workflows, interfaces, process, planning
    3. Compare results
  • Exec

    1. Ask organizational leaders what their pain points are
    2. Look for duplications of language, vision, alignment efforts, programs
    3. Compare results
Hazel Weakly (@hazelweakly@hachyderm.io)
How to Build High-Impact Platforms

Understanding Each Other

Motivation

IC
Thriving
Manager
Outcomes
Exec
Alignment
  • Priorities

    IC
    Reduction of toil
    Manager
    Optimization of team interactions
    Exec
    Long term alignment of innovation
Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Takeaways

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

A Platform Is An Abstraction That Platforms Like A Platform

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Platform teams are organizationally empowered to address cross-cutting changes

With a combination of technical, procedural, and cultural improvements

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Platform Engineering can turn culture into an intuition around how to use a platform

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Make Change Easy to Handle, Not Easy To Do

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Failure Modes

  • DevOps tends to fail because the sharp and blunt ends of a system are rarely aligned and addressing that requires a cultural shift
  • SRE tends to fail when implemented prematurely; it only makes sense when the horizontal context is too large for any one team
  • Platform Engineering tends to fail when implemented without empathy and diversity
  • Platform Engineering tends to fail when an agglomeration based approach isn't used to address optimizations
Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Platforms are a way to scale doing

Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

What is Platform Engineering

IC
Empathy Driven Development
Manager
Empathy Driven Collaboration
Exec
Empathy Driven Organizational Design
Hazel Weakly (@hazelweakly@hachyderm.io)
Takeaways

Platform Engineering is Empathy Driven Learning

Hazel Weakly (@hazelweakly@hachyderm.io)

platform engineer and platform engineering are unrelated much like SRE and resilience engineering