Xp2015 Scaling Agility explored - LeSS SAFe comparison

© GOSEI
Scaling Agility Explored
Ran Nyman, Ari Tikka
XP2015 27.5.2015
Wednesday 27 May 15

Gosei&Oy&all&rights&reserved.
About this session
2
About
Root causes and basic assumptions
behind Scaling Agile
For you to evaluate LeSS and SAFe
How to analyze
Coordination
Organizational layers of control with
the theory by William G. Ouchi
Flow of work
Batch size and Queues
Corporate and business perspective
Wednesday 27 May 15

Speakers
3
Ari TikkaRan Nyman
Wednesday 27 May 15

Why to Scale Agile?
4
Wednesday 27 May 15

5
Don’t!
Adding more people just makes you slower.
-- one of the directors of SAGE program 1950’s
-- Frederick Brooks at Mythical man-month, 1975
Wednesday 27 May 15

Still want to scale up?
6
We have grown big, slow
and wasteful.
We are creating complex big
products. We don’t want to
become slow and wasteful.Jersey Shore www.joiseyshowaa.com
Wednesday 27 May 15

7
How do you end up slow and wasteful?
Common sense and fashionable solutions
Wednesday 27 May 15

“Hey, We have business! And it is
growing!”
“People just ﬁnd their roles.”
“Specialists are irreplaceable. We
need to optimize their individual
performance.”
In the beginning
8
Wednesday 27 May 15

“It starts to get messy. We need
someone to look after things.”
“Lets hire a coordination
specialist - the project
manager.”
Growing the using common sense
9
Wednesday 27 May 15

“The project managers really do
their job.”
“Obviously it is best to give
responsibilities to the
specialized people.”
Growth continues
10
Wednesday 27 May 15

“Do You understand what is
really going on?”
“Blessed the are project
managers. They get something
out of this mess.”
The coordinators become the heroes
11
Wednesday 27 May 15

“We are slow and expensive. Why are
projects no more productive?”
“People Resources are either idling or
overloaded.”
“The portfolio does not obey.
Dependencies and maintenance
dominate.”
But… too much to be coordinated
12
?
Wednesday 27 May 15

Symptoms of fragmented organisation
Wasting. Waiting. Scatter. Handovers. Loss
of knowledge. Hunting for resources. Bad
quality. Quick fix. Distress. Reorgs. Cost
management. Gaps between roles.
Nonproductive feedback. Misleading
measures. Unclarity. Bad atmosphere. No
time for learning.
Knowledge and power is always elsewhere!
13
?
Wednesday 27 May 15

Would scaling Agile up
help us?
14
We are slow and wasteful!
Wednesday 27 May 15

15
What to DO?
Wednesday 27 May 15

Scrum works for one team
16
Wednesday 27 May 15

17
Program Execution
Scrum team
Wednesday 27 May 15

18
Customer-centric learning
Wednesday 27 May 15

LeSS Framework
19
Wednesday 27 May 15

LeSS Huge
20
Wednesday 27 May 15

21
Control Systems and Coordination
Wednesday 27 May 15

Control Systems in organizations by
William G. Ouchi
22
Measure Input (€) and Output (€). Contractual between
parties. Exact contract!
Written rules and processes. E.g. Employment agreement
and supervision.
Informal value based rules that allow innovation and
collaboration. Only this works for unique, interdependent or
ambiguous task. E.g. SW Development
Market system
Bureaucratic
system
Clan system
Wednesday 27 May 15

William G. Ouchi
Inventor of management control mechanisms
Inventor of motivation Theory Z
Addition to well know Theory X and Y
Influenced by Japanese management style
23
Wednesday 27 May 15

Projects Coordinate Peoples Time and
Technology Dependencies
24
Wednesday 27 May 15

Gosei&Oy&all&rights&reserved. 25
Market
Control
Bureaucratic
Control
Clan Control
Project x Project Y Project z
Projects control people
Wednesday 27 May 15

Programs Coordinate Teams and
Technology Dependencies
26
Wednesday 27 May 15

Component teamComponent teamComponent teamComponent team
Market
Control
Bureaucratic
Control
Clan Control
Program
x
Program
Y
Program
z
Programs coordinate teams
Wednesday 27 May 15

Control by SAFe
28
Bureaucratic System
(process, written rules,role
descriptions)
Clan System (social rules)
Market System (€)
Wednesday 27 May 15

What changes Risk or challenge
Release trains and 3-month cadence
instead of parallel projects.
The parallel projects existed for a reason, which is not yet
solved. Still need for substantial planning, because the
underlying organizational design is unchanged.
From time based resource planning
to team output estimation.
Improved communication by all
hands release planning meeting.
The amount of dependencies, and queuing them for
solution remains a challenge and results in branching and
late integration. Contract game remains for planning needs.
Welcoming teams to middle management.
Training and consultation for Lean-
Agile best practices.
Culture follows structure. Focus in coordination, not value
adding work. Thinking and communicating in organizations
differs from what is actually happening. *)
*) Mats Alvesson and André Spicer: "A Stupidity-Based Theory of Organizations", Journal of Management Studies 49:7, November 2012.
André SPICER: 2013 "Shooting the shit: the role of bullshit in organisations" M@n@gement, 16(5), 653-666, Cass Business School, CU London.
Wednesday 27 May 15

What is not changing Risk or challenge
The numerous middle management roles are
renamed. As before, scarce resources are moved
from teams to management (e.g. UI design,
Architects).
The change does not happen.
The change is not useful.
Corporate layers of power and control legitimized
to be Agile.
No real change.
No business Agility developed.
Business decides, programs execute.
Contract game with business.
Little emphasis for structural change from
functional to feature teams.
Technical capability and competence limit the
effectiveness of the change.
Wednesday 27 May 15

Teams Coordinate Dependencies and
Technology
31
Wednesday 27 May 15

Market
Control
Bureaucratic
Control
Clan Control
Feature teamFeature teamFeature team
Feature teamFeature teamFeature team
32
People work with technology
Teams coordinate
Wednesday 27 May 15

Focus from Projects to Customer
33
Customer
Organizational design
Wednesday 27 May 15

Nooooooo!
We can not
change
everything.
Wednesday 27 May 15

Your Fear is Just
Expecting big improvements without significant change is
unreasonable!
Changing “everything” in a small independent part is the ONLY way to
real change.
Experiment and learn with limited risk
Resources for enough support
Moore’s chasm
36
Wednesday 27 May 15

Project Y Project x
Component teamComponent teamComponent team
37
Feature team
Feature team
Coordination cost
Investment
in learning
Wednesday 27 May 15

38
Flow of work
Wednesday 27 May 15

Three Layers in (large) Organizations
Business (top) management
39
Middle management
Value adding workers
Analyze
Coordinate
Intermediate
Execute
Technical reality
Economical reality
Internal reality
Reward power
Expert power
Dependent power
->Politics
Wednesday 27 May 15

Who is missing?
40
Business (top) management
Middle
management
Value adding workers
Analyze
Coordinate
Intermediate
Execute
Customer
User
Wednesday 27 May 15

Technology specailization leads to functional silos
41
Component
Teams
Customer
Problem
Product
Management
Feature ??
Release
Which team?
Wednesday 27 May 15

41
Component
Teams
Customer
Problem
Product
Management
Feature
Specifiers
Architects
UI
Release
Wednesday 27 May 15

41
Component
Teams
System
integration
Testing
Deployment
Customer
Problem
Product
Management
Coordination
Feature
Specifiers
Architects
UI
Project
Managers
Release
Wednesday 27 May 15

41
Component
Teams
System
integration
Testing
Deployment
Customer
Problem
Product
Management
Coordination
Feature
Specifiers
Architects
UI
Project
Managers
Release
Bloated
Backlog
Wednesday 27 May 15

Through backlog and specialists
42
Customer
User
Wednesday 27 May 15

From Technical To Customer Specialization
43
Wednesday 27 May 15

Flow of work with Customer Specialization
44
Customer
Product
Owner Deployment
Release
Wednesday 27 May 15

Flow of Work LeSS
45
Wednesday 27 May 15

Nooooooo!
It is too simplistic.
We are so many!
Wednesday 27 May 15

Yes, it is simple and not easy
Technology, competence, identites and culture need to develop.
Learning causes anxiety. Only survival anxiety is greater. (E. Schein)
Takes time, like any real change.
There will be worry and resistance.
Leadership challenge
47
Wednesday 27 May 15

48
Batch size and Queues
Wednesday 27 May 15

SAFe Batch Size
Planning cycle 3 months
Create big batch of work
to reduce total cost
49
Wednesday 27 May 15

LeSS Batch Size
Planning 2 week increments
Create small batches of work
that will enable fast feedback
50
Wednesday 27 May 15

Optimal
Batch Size
Optimal
Batch Size
Why is the Batch Size Problem?
51
Wednesday 27 May 15

Product Development and Big Batches
“We have found out that reducing batch size improves most
development projects significantly.”
-Six Myths of Product Development Stefan Thomke and Donald Reinertsen -
52
Wednesday 27 May 15

Queues SAFe
53
SAFe
Loads the system full of queues
for a Program Increment
Optimizes resource utilization
Wednesday 27 May 15

Queues LeSS
54
LeSS
Tries to keep queues outside of
system
Optimizes outcome after each
iteration
Wednesday 27 May 15

General Problem With Queues and Big Batches
55
Big Batches
Single Arrival
Wednesday 27 May 15

Product Development and Queues
“Queues delay feedback, causing developers to follow unproductive
paths longer. They make it hard for companies to adjust to evolving
market needs and to detect weaknesses in their product before it’s too
late.”
-Six Myths of Product Development Stefan Thomke and Donald Reinertsen -
56
Wednesday 27 May 15

57
Coordination Summary
Wednesday 27 May 15

Fundamental formula
When coordinating the work to be done,
the more technology-specialized the organization is
and the more you want to optimize utilization
the further into the future you need to plan.
58
Reach (length) of the plan = Utilization x Specialization
Wednesday 27 May 15

Coordination Approaches Compared
59
SAFe LeSS
Main control mechanism
Solving dependencies
Batch size
Cost of dependencies
Optimization
Customer contact
Organizational maturity
Requires stability in
Bureaucratic Clan
Coordinate people People work with technology
Big and slow tasks for scarce resources (people).
3-month releases needed to plan.
Fast, small and parallel technical
transactions. Sprint-long iterations.
Coordination is seemingly necessary waste
Learning to work with technology is
investment
Resource optimization (coordination) Outcome optimization
Intermediated Direct
Possible with lower skill
Learning for the role
“Natural” development
Higher skill needed
Learning what is needed
Skilled evolution, leading learning
Component organization functions in
unchanging environment.
Long-living teams adapt fast to changes in
environment.
Wednesday 27 May 15

60
Business view
Wednesday 27 May 15

Growth of the middle management
63
Customer,
user
From Agile Manifesto:
Individuals and interaction
Business and developers work
together daily
Face-to-face conversation
Simplicity
Self-organization
Learning from reality
Owner, capital,
top management
Front-line
worker
Wednesday 27 May 15

1. New process and best practises by SAFe
64
Customer,
user
together daily
Simplicity
Self-organization
Owner, capital,
top management
Front-line
worker
Wednesday 27 May 15

2. Dis-intermediating by LeSS
65
Customer,
user
together daily
Simplicity
Self-organization
Owner, capital,
top management
Front-line
worker
More with
Wednesday 27 May 15

Nooooooo!
It will break!
Wednesday 27 May 15

Find your product to enable direct
customer interaction.
Build customer-oriented feature
teams.
Learning away from coordination
chaos. Decoupling in practise.
The Product Owner decides, customer
interaction clarifies.
The line management grows the value
of the organization.
LeSS Organizational design
Wednesday 27 May 15

Your key questions
What do you want?
What do you dare?
70
Wednesday 27 May 15

SAFe LeSS
Slogan
Framed problem
Value proposition for
“Scaling Agile”
Solution
Main control mechanism
Real-time delivery
Adoption scope
Program Execution Customer-centric Learning
Internal efficiency Optimal response to customer demand
Improved program execution
Lean-Agile ways of working
More with less:
Effective and agile value-adding work
Minimal bureaucracy
Program process and best
practises
Organizational design: principles, guides, rules and
600 experiments for inspect and adapt
Bureaucratic Market, Clan
Detailed planned 3-month cycle. Continuously improve real-time delivery
Program level One product first
Questions
71
Wednesday 27 May 15

Thank You
72
Wednesday 27 May 15

Xp2015 Scaling Agility explored - LeSS SAFe comparison

More Related Content

What's hot (20)

Similar to Xp2015 Scaling Agility explored - LeSS SAFe comparison (20)

More from Gosei Oy (14)

Recently uploaded (20)

Xp2015 Scaling Agility explored - LeSS SAFe comparison