Scrum has proven to be a successful framework for many companies in complex delivery domains to transition to Agility from more traditional delivery methods. At some point during these transitions many companies have experienced a stall to what have previously been ongoing and exponential improvements. While team-level performance might have seen a substantial improvement, end-to-end delivery as perceived by customers and other stakeholders may not bit fit enough. In other cases, no matter how much teams work within the Scrum framework to improve their work, recurring “impediments” cause the Scrum implementation to plateau, or even regress. We have learnt that complexity brings with it challenges in the "white spaces" that exist between teams where local improvements become disconnected from the delivery of customer value. In essence, Scrum is good, but doesn't provide enough to meet the needs of the whole system. This talk will summarize key lessons distilled from that realization. We will highlight some of the limitations of team-centric Agile approaches and discuss Lean & Systems Thinking concepts to improve the system as a whole and bring your organization back to a place of continuous improvement that is connected to customer value.

1. Amp up your Agile implementation in
complex environments with Systems
Thinking
@martinaziz
@fer_cuenc
a
Fernando A. Cuenca Martin Aziz
fernando.a.cuenca@gmail.com martinaziz@gmail.com

2. A company responding to the
environment
The
Company
Customers

3. Complexity
of Problem
Considering the problem domain
Causal
Loop
Part 1

4. Dedicated
ScrumMasters
Dedicated
Product
Owners Cross-
functional,
small teams
Agile Pattern
Part 3
Introducing Decentralized
Control & Capacity Management
The
Company

5. Complexity brings scale
Number
of
Teams+
Complexity
of Problem
Causal
Loop
Part 2
+

6. Customer Recognizable
Features, Fitness for Purpose
and Lead Time
The
Company
Customers
Product
Owners

7. Tribal Identifies, Inwards-looking
local optimization
The
Company

8. Teams and the local agenda
Number
of Teams
Tribal
Behaviou
r
+
Local
Improvement
Effort
Complexity
of Problem
+
+
Causal
Loop
Part 3

9. Where is my
stuff?
What am I
going to get
at the end?
Why does it
take this
long?
Seems to
work for
me….
6
End to End Measurement for
Fitness
The
Company

10. How does this impact Lead Time and
Fitness for Customer Purpose?
Lead
Time
Fit for
Purpose
-­‐
Number
of Teams
Tribal
Behaviou
r
+
Local
Improvement
Effort
Complexity
of Problem
+
+
?
Causal
Loop
Part 4
t

11. Comes from
different
sources
Comes in different
types, requiring
different processing
It has different
frequencies of
arrival
It has different
levels of urgency,
Importance, and
cost of delay It has different
perceived
value
Cost of Delay and
Heterogeneous demand
The
Company

12. Senior
Stakeholder
Customer
Commitment
Push
Queue
Commitment, Push and
overburdening
The
Company

13. Potential for overburdening
Number
of
Teams
Tribal
Behavio
ur
+
Local
Improvement
Effort
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
Overburdening
Complexity
of Problem
+
+
+
+
Causal
Loop
Part 5

14. Delivery
“chains”
Shared team
members
Internal
“shared
services”
Person with
very
specialized
skills “floating”
around teams
“External” team
members
White Spaces
The
Company

15. The impact of Teams & Overburdening on
Flow Efficiency
Number
of
Teams
Tribal
Behavio
ur
+
+
Flow
Efficienc
y
-­‐
Local
Improvement
Effort
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
Overburdening
+
Complexity
of Problem
Coordination
effort +
White Space
+
+
+
-­‐
-­‐
Causal
Loop
Part 6
Wait Wait Wait WaitWork Work Work Work
Flow
Efficiency
=
work
time
work
time
+
wait
time
x 100%

16. Decomposition and Batch Sizes
The
Company

17. Some stay the
same
Some are re-
aggregated
and batchedSome describe
business
functionality
Some are
purely
technical tasks
Some have
internal
dependencies
Some are sent
to other team’s
backlogs
11
Recognizability and Transaction
Costs
The
Company

18. The impact of batch size on lead
time
Number of
Teams
Tribal
Behaviour
+
+
Flow
Efficiency
-­‐
Local
Improvement
Effort
Lead Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
Overburdening
+
Complexity
of Problem
Coordination
effort +
White Space
+
+
+
-­‐
-­‐
Batch
Size
+
+
Transaction
Cost
+
Causal
Loop
Part 7

19. ?
?
Local OptimizationThe
Company
??
?
?
?
?

20. Decomposition and Customer Recognizability, the final
links
Number
of
Teams
Tribal
Behaviour
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficiency
-­‐
Batch
Size
Local
Improvement
Effort
+
+
Lead
Time
Fit
for
Purpose
-­‐
Implicitness
of
Commitment
Point
Distance
between
Teams
&
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of
Problem
Coordination
effort
+
White
Space
+
-­‐
-­‐
+
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Causal
Loop
Part 8

21. Sense
&
Promise
Push
Stories
Scrum
Team
Scrum
Team
XP
Team
Story
AccumulaBon
Features
Ideas
Kanban
Team
Customers
Delivery
Customers
Doing
White
spaces
between
teams
sources
of
greatest
delays.
Stories
hard
to
recognize
by
customers
or
actors
on
the
leR
end
of
the
flow.
Unconstrained
demand
Early
Commitment
without
connecBon
to
capability
Key challenges of the systemThe
System

22. Constraining the work
Number
of
Teams
Tribal
Behaviour
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficiency
-­‐
Batch
Size
Local
Improvement
Effort
+
+
Lead
Time
Fit
for
Purpose
-­‐
Implicitness
of
Commitment
Point
Distance
between
Teams
&
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of
Problem
Coordination
effort
+
White
Space
+
-­‐
-­‐
+
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐ -­‐
Breakin
g the
loop

23. Making commitments explicit
Number
of
Teams
Tribal
Behaviour
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficiency
-­‐
Batch
Size
Local
Improvement
Effort
+
+
Lead
Time
Fit
for
Purpose
-­‐
Implicitness
of
Commitment
Point
Distance
between
Teams
&
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of
Problem
Coordination
effort
+
White
Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐ -­‐
Explicit
Commitment
-­‐
Breakin
g the
loop

24. Know your delivery capability and implement Pull
Number
of
Teams
Tribal
Behaviour
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficiency
-­‐
Batch
Size
Local
Improvement
Effort
+
+
Lead
Time
Fit
for
Purpose
-­‐
Implicitness
of
Commitment
Point
Distance
between
Teams
&
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of
Problem
Coordination
effort
+
White
Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐ -­‐
Explicit
Commitment
-­‐
Pull
Policies
Measurement
-­‐
-­‐
Breakin
g the
loop

25. See your world as a service to a customer. Maintain
customer recognizability at all times.
Number
of
Teams
Tribal
Behaviour
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficiency
-­‐
Batch
Size
Local
Improvement
Effort
+
+
Lead
Time
Fit
for
Purpose
-­‐
Implicitness
of
Commitment
Point
Distance
between
Teams
&
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of
Problem
Coordination
effort
+
White
Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐ -­‐
Explicit
Commitment
-­‐
Pull
Policies
Measurement
-­‐
-­‐
Service
Orientation
+
Breakin
g the
loop

26. Flow – getting to a system that is in control
Sense
Pull
Features
Features
Do Next
Ideas
Customer
s
Delivery
Customer
s
Features
Do
Features
Good
Constrained delivery
pipeline
Upstream Downstream
Progress
Customer
Recognizable
Delivery
Improvements
aligned to optimize
for value delivery
Work is pulled into
delivery pipe
automatically as capacity
becomes available. Push
is avoided to prevent
overburdening.
A new
system

27. 0. Understand the purpose of the System and identify Services
1. Understand Sources of Dissatisfaction
2. Analyze Demand and Capability
3. Model the Knowledge Discovery Knowledge
4. Discover Classes of Service
5 . Design Kanban Systems
6. Roll-out
One way to get there: STATIK
the Systems Thinking Approach to Introducing Kanban
A new
system
1. Who are my
customers?
2. What do they
expect?
3. How do I meet
those
expectations?

28. Kanban at the System Level
A new
system
Service 1
Service 2
Service 3
Discovery
(N)
Construction
(N)
Deployment
(N)
Commitment
Points
Upstream Flow
Downstream
Flows
Constraints to
enable PULL
Validation
(N)
Different
visualization
fidelity

29. Learning
More Some resources that have influenced us
Antifragile: Things
That Gain from
Disorder
Nassim Nicholas Taleb
Kanban from the
Inside
Mike Burrows
The Principles of
Product Development
Flow
Donald G. Reinersten
Lessons in Agile
Management: On the
Road to Kanban
David J. Anderson
Goal: A Process of
Continuous
Improvement
Eli Goldratt
Management 3.0
Jurgen Appelo
Toyota KATA
Mike Rother
Actionable Agile
Metrics for
Predictability
Daniel S. Vacanti

30. Learning
More Some people that have influenced us
David J. Anderson (LKU,
Kanban)
leankanban.com/blog
Alexei Zheglov (Kanban)
connected-­‐knowledge.com
Patrick Steyaert (Flow)
www.okaloa.com
Chris Chapman
(#SystemsThinkingTO)
@DerailleurAgile
Esther Derby (Agile
Management)
www.estherderby.com/category/
insights
Dave Snowden (Cynefin)
cogniQve-­‐edge.com/blog

31. About us How to reach us to keep the conversation going
@fer_cuenc
a
fernando.a.cuenca@gmail.com

32. Scratch
Area
aRer
this
point

33. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐
-­‐

34. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐
-­‐
Explicit
Commitment
-­‐

35. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐
-­‐
Explicit
Commitment
-­‐
Pull
Policies
Measurement
-­‐
-­‐

36. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐
Constraints
-­‐
-­‐
Explicit
Commitment
-­‐
Pull
Policies
Measurement
-­‐
-­‐
Service
Orientation
+

37. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
+
-­‐
+
Transaction
Cost
-­‐

38. Complexi
ty
of
Problem
Number
of Teams
+
-­‐
Tribal
Behaviou
r
+
Coordinatio
n effort +
White
Space
+
Decompositio
n
+
Customer
Recognizabilit
y
Flow
Efficien
cy
-­‐
Batch
Size
Transacti
on Cost
Local
Improvement
Effort
+
+
+
+
+
Lead
Time
Fit for
Purpose
-­‐
-­‐
+
Implicitness of
Commitment
Point
Distance
between
Teams &
Commitment
Point
+
+
-­‐
Overburdenin
g
-­‐
+
+
+

39. Number
of
Teams Tribal
Behavio
ur
+
+
Decomposition
+
Customer
Recognizability
Flow
Efficien
cy
-­‐
Batch
SizeLocal
Improvement
Effort
+
+
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
+
Overburdening
-­‐
+
Complexity
of Problem
Coordination
effort +
White Space
+
-­‐
-­‐
+
+
+
+
+
+
-­‐
+
Transaction
Cost

40. Many people and groups are required
to achieve the company’s goals.
The
Company

41. The impact of batch size on lead
time
Number
of
Teams
Tribal
Behavio
ur
+
+
Flow
Efficien
cy
-­‐
Local
Improvement
Effort
Lead
Time
Fit for
Purpose
-­‐
Implicitness of
Commitment
Point
Distance
between Teams &
Commitment
Point
Overburdening
+
Complexity
of Problem
Coordination
effort +
White Space
+
+
+
-­‐
-­‐
Batch
Size
+
+
Transaction
Cost
+
Causal
Loop
Part 7