Scrumban

SCRUMBAN DEMYSTIFIED
FOR AGILE NEW ENGLAND
Ajay Reddy
Official Licensed Material Copyright CodeGenesys, LLCajay@codegenesys.com | @ajrdy

Official Licensed Material Copyright CodeGenesys, LLC
What Scrum and Scrumban are not
ajay@codegenesys.com | @ajrdy

Topics
•  Why and What Scrumban
•  Scrumban Foundations and Evolutions at
Mammoth Bank
•  Scrumban measurements (Optional)

Section 1 of 3
Why and what Scrumban

VersionOne’s State of Agile across the
industry

Scrum
Source: Scrum Alliance

75% of organizations using Scrum will not succeed in
getting the benefits that they hope for from it.“
“
-Ken Schwaber

Why Scrumban
managing variability in knowledge
Improves predictability
imposes no prescribed process
destination
provides more natural and
contagious mechanisms for
scaling improvements across an
enterprise
can germinate from anywhere
within an organization
product / organization
management features into the mix
such as risk, value & cost.

Mammoth Bank
Mammoth Bank is a large national (international) bank
headquartered somewhere in the US.
It has a very large IT organization (several thousand strong)
spread across multiple locations, including several offshore
partners.
ajay@codegenesys.com | @ajrdy Official Licensed Material Copyright CodeGenesys, LLC

History and current
state of Agile
transformation
3 revision of Agile transformation
In 2013, the bank was faced with
several difficulties:
Disruptive change not
working
Premature Success declared
Coaching with inadequate
results
Some improvements clearly
presentajay@codegenesys.com | @ajrdy Official Licensed Material Copyright CodeGenesys, LLC

Vague and False impressions

Bank’s problem areas
Productivity
Predictability
Politics
Profitability
Permanence
Plasticized
Prematurity

Bank’s goals
greater flexibility (agility)
improved predictability
(dependability)
increased productivity
higher quality
better alignment
greater customer satisfaction
GREATER MARKET SHARE
SUSTAINABLE PROFITABILITY

Scrumban Definition
Scrumban is the Kanban method applied to
Scrum as the starting and underlying
process framework.

4 Principles
•  Start with what you do now
•  Agree to pursue incremental, evolutionary change as opportunities are
discovered
•  Respect current process, roles, responsibilities and titles
•  Encourage acts of leadership at all levels of the organization
6 Practices
•  Visualize
•  Limit WIP (‘kanban’)
•  Manage flow
•  Make policies explicit
•  Develop feedback mechanisms at the workflow, inter-workflow and
organizational levels
•  Improve collaboratively, evolve experimentally (using models and the scientific
method)
The Kanban Method

Scrumban Definition
S c r u m b a n i s a p r a g m a t i c a n d
entrepreneurial application of The
Kanban Method to Scrum, using scientific
management theories often resulting in a
continuing evolution of Scrum based
systems into business valued, service
oriented, Agile, Lean and adaptive
systems at a pace moderated by the
thinking systems.
Scrumban is the Kanban method applied to
Scrum as the starting and underlying
process framework.

Scrumban
False Hybrids and
common ‘macro’ evolutions
“Scrum in Kanban”
“Kanban in Scrum”
Scrumban is not a process
process framework or a process
destination!

Scrum’s iteration/sprint
is an instance of
Kanban’s synchronized
two phased commit.

Some common examples of change
( not comprehensive):
•  Flow within an iteration (Two
phased synchronized commit)
•  Iteration planning
•  Prioritization
•  Role definitions
•  MVP/MMF and Release planning
•  Spikes
•  Story points (not official part of
Scrum)
•  Budgeting
•  Epics and story break down

Section 2 of 3
Scrumban evolutions

The Old Model (18th-20th Century)
Managers
Process
Workers
Process
Coaches
Designs
Or
DefinesImposes
Follow
Assign
Workers
To Tasks
Source: David Anderson, Agile Russia

The New Model (21st Century)
Managers
ProcessWorkers Process
Coaches
Facilitate
learning
Control
&
Evolve
Policies
Source: David Anderson, Agile Russia

RAD Group
Significantly delayed development of a consumer-facing mobile
application.
Bank was beginning to rapidly lose customers to other institutions.
5 separate team of 6-7 members each

Sam’s Team
•  Product Back log/Iteration back log/User stories
•  Sprints Burn downs and Burn ups
•  Scrum master, product owner, Agile coach
•  Push model- assignments
•  Consensus based points estimation
•  Priority determined by product owner
Mastery Level !
A3 Thinking
Cynefin
Systems thinking
Long-Tail
Learning Evolutions
Starting
Conditions
Predictability/Reliability
Business effectiveness of
results
www.scrumban.io Official Licensed Material Copyright CodeGenesys, LLC

Mobilize
The Kick start framework
Sources of dissatisfaction and
Defining Purpose & Success
Criteria
Who are our customers
Demand and capability Analysis
Basic Management

Moblize
•  work items / work cards (user
stories)
•  work size estimate (story points)
•  definition of done
•  daily standups
•  work type
•  workflow
•  pull
•  ready / buffer columns
•  blockers
•  classes of service
•  capacity allocations (explicit
WIP limits) [Proto kanban]
Already Familiar to Scrum Teams
New Concepts

System 1 (Sam’s Team)
Product
Backlog
Iteration
Commitment To do Doing Release Done
Learning Evolutions
Source

Des*na*on

DoD

Mastery Level !
A3 Thinking
Cynefin
Systems thinking!
Long-Tail Priorities
- Standard
- Bug Fix
Iden%ﬁed

Recogni*on
of
need
for

understanding
demand
and

capability

What
is
the
nature
of
demand

from
them?

Deﬁni*on
of
done


Systems Thinking

Systems Thinking
Dynamic systems
Amplifying or Dampening
Causal effects, Cascading effects
Changing systems
Skepticism
Complexity within Lateral
Thinking
Lean toward the weightier
understanding of the collecting of
the system

System 1 (Sam’s Team)
Iteration
Backlog Analysis Develop Test Deploy Done
Learning Evolutions
Mastery Level !
A3 Thinking
Cynefin
Systems thinking !
Long-Tail
DoD DoD DoD- Standard
- Bug Fix
How
works
is
done

Customers
Iden*ﬁed


Thinking Systems
Creates an adaptive capability in
your organization
Enables adaptability in your
business processes to respond
successfully to changes in your
business environment
Toyota’s quality circles
• Optional
• Collaborative
• Focused
• Empirical
• Common method
Toyota’s culture and adaptive capability's
outcome

A3
PDCA - a basic element of the Total
Quality Management movement.
The power of A3 reports and the
thinking behind them requires a good
grasp of the Plan-Do-Check-Act
(PDCA)
The PDCA cycle simply follows the steps of
the scientific method:
Plan is developing a hypothesis and
experimental design;
Do is conducting the experiment;
Check is collecting measurements; and
Act is interpreting the results and taking
appropriate action.
Logical
Objective
Means and ends important
Synthesis and presentation
Coherence-Continuity

Cynefin

System 1 (Sam’s Team) 6 Weeks
Analysis Develop Test Deploy
Learning Evolutions
Mastery Level !!
A3 Thinking !
Cynefin
Agile Budgeting
Long-Tail
- Bug Fix
Doing Done Doing Doing ReadyDone Done Done
DoD
Iteration
Backlog
Ready
•  Visualiza*on
aided
sharing

of
learnings
with
others-‐

Collabora*on
gains

•  Adopted
Buffer
Lanes

•  Not
handoffs
but
discovery

of
informa*on

•  Lean
Coffee


Effect of “Long Tail” on
Business alignment of the
system
•  Changes in Product Owner
Options and Priorities – Niche
features, Potential Value, Risk
of
•  Change Budgeting?
Differentiators, Spoilers, Cost
Reducers, Table Stakes
Innovative, Major Growth Market,
Cash Cow

System 1 (Sam’s Team) 4 Months
Learning Evolutions
Mastery Level !!!
A3 Thinking !!
Cynefin
Agile Budgeting
Long-Tail !!
- Bug Fix
DoD
§  Adopted Classes of Service & WIP Limits based on
experience shared by Bala’s team.
§  Collaborated with Bala’s team to apply A3 Thinking to
problem of business needing to release features to the
market as soon as possible.
§  Identified Continuous Flow & Modified Deployment mode as
solution (and successfully negotiated with the product
owner and other dependencies).
Analysis 3 Develop
3
Test 4 DeployBacklog
Doing Done Doing Doing ReadyDone Done
Ad Hoc
Standard
S
S

S
S

D
D

D

System 4 (Sri’s Team)
Analysis Develo
p
Test Deploy Done
Learning Evolutions
D

Mastery Level !
A3 Thinking
Cynefin
Agile Budgeting
Long-Tail !
- Bug Fix
Iteration
Backlog
•  Still practicing basic Scrum and using its
visualization as a task board.
•  Applying Team 1’s Long-Tail to prioritize
backlog items because it was mandated
by team leader.
C
S

8 Weeks
Differentiators, Spoilers, Cost
Reducers, Table Stakes

Scrumban system as a Queuing system
Source: Dimitar Bakardzhiev

Applying Little’s law
Departure rate
approximately equals arrival
rate (λ).
No work items that get lost or
never depart from the system.
It can be applied at an
iteration level of a project or
continuous flow.
Little’s Law holds in case of a project:

Z Curve Breakdown

Basic Sprint/Release
Planning with Little’s
law
We can calculate WIP to deliver for a
particular lead time using the formula:
(Use with much care - Use qualitatively to
influence/learn not make commitments)
WIPt = LTt
N
T

Managing uncertainty
when planning a
project
Account for the first and third legs of
the Z-curve
Account for Dark Matter
Account for Failure load
Uncertainty is reduced by aggregation. When
calculating Project lead time we need to
account for the uncertainty in terms of:
Source: Dimitar Bakardzhiev

On managing
uncertainty
Size Buffer
.
Scope Buffer
Time Buffer
Skills Buffer
Buffer Management
Focus on median lead time

System 2 (Bella’s Team) 8 Weeks
Learning Evolutions
Mastery Level !!
A3 Thinking
Cynefin
Agile Budgeting
Long-Tail !
- Bug Fix
DoD
•  Sharing its learnings with others, & applying
Team 1’s Long-Tail to prioritize backlog items.
•  Adopted Buffer Lanes. WIP Limits & Classes
of Service to differentiate ad hoc work
requests.
•  Unique in its efforts to plan with higher
confidence Little’s Law.
Analysis 3 Develop 3 Test 4 Deploy
Iteration
Backlog
Ad Hoc
Standard

System 2 (Bella’s Team) 4 Months
Learning Evolutions
Mastery Level !!!
A3 Thinking !
Cynefin !
Agile Budgeting
Long-Tail !
DoD DoD DoD
DoD
Analysis 3 Develop 3 Test 4 DeployBacklog
URGENT
STANDARD
Waiting
Simple
Complicate
dComplex
Chaotic
StdBug
s
•  Little’s Law used to influence time
expectation
•  Collaborated with Sameer’s team on
A3 Thinking to adopt Continuous Flow
& Modified Deployment
•  Adopted Cynefin framework from
Jivan’s team. Substituted Urgent class
of service for Ad Hoc items.
•  Added “waiting area” to visualize time
spent waiting on outside
dependencies.

System 5 (Raj’s Team)
Analysis Develop Test Deploy
Learning Evolutions
Mastery Level !
A3 Thinking
Cynefin
Agile Budgeting
Long-Tail !
- Bug Fix
DoD
•  Adopted Buffer Lanes
Iteration
Backlog
Ready
4 Months

System 1 (Sam’s Team) 6 Months
Learning Evolutions
D

Mastery Level !!!
A3 Thinking !!
Cynefin !
Agile Budgeting
Long-Tail !!!
S

Analysis 3 Develop
3
Test
4
DeployBacklog
URGENT
STANDARD
Simple
Complicated
Complex
Chaotic
Std
Doing Done Doing Doing ReadyDone DoneFIXED/ INT
•  Adopted application of Cynefin
framework & Classes of Service
from Jivan’s & Bala’s teams.
•  Adopts planning & forecasting with
Little’s Law including Buffering.
S

S
S

S
S

D
D

D

D

Design Thinking
Design thinking has come to be defined as combining
empathy for the context of a problem, creativity in the
generation of insights and solutions, and rationality in
analyzing and fitting various solutions to the problem
context.

Design Thinking
Board Design
Policy
Distribute teams
Electronic Tools
Design Thinking is "matching people’s
needs with what is technologically feasible
and viable as a business strategy"
Consider

Design Thinking
Uncluttered
Up to Date at all times
Universal
Uniform
Useful
‘5 U’ Boards

Options Theory
Never commit early unless you know why“
“
- Olav Maassen

Options Theory
Set your own constraints.
Review your options.
Collect information (and find more
options).
When an option is about to expire
review your constraints to see if
they are still valid.
Choose the best option based on
your current information at hand.

Options Theory
“Iteration commitment”
“Cost of Delay” of releasing a
product
“Last Responsible Moment”
Consider

System 2 (Bella’s Team) 6 Months
Doing 1 Done 1 Doing 1 Doing 1 ReadyDone 1 Done 1 Done
Analysis Develop Test DeployBacklog
Doing 2 Done 2 Doing 2 Doing 3 ReadyDone 2 Done
URGENT
STANDARD
Doing 1 Done 2 Doing 1 Doing 1 ReadyDone 2 DoneFIXED/ INT
•  Expanded dependency waiting areas
to all effected phases.
•  Negotiated SLAs with outside
dependencies to improve overall
system flow.
•  Pull Decisions based on Options
Theory. Constraints are not static.
Waiting Waiting
WaitingWaiting
WaitingWaiting
Learning Evolutions
Mastery Level !!!
A3 Thinking !!
Cynefin !
Agile Budgeting
Long-Tail !

Do More
ScrumDo’s new product release
Scrumbando.com
Official Licensed Material Copyright CodeGenesys, LLCwww.scrumban.io

Codegenesys.com/events
Foundations training
Advance Practitioner Training
Scrumban Guarantee program
Contact
info@codegenesys.com

Like Us
https://www.facebook.com/scrumban

Section 4 of 4
Scrumban fitness

Measurements
Measuring the right things and giving it
the appropriate weight
Measuring with appropriate degree of
certainty
Making sense of the data- Discipline of
collection and Analysis
Why measure ?

Any observed statistical regularity will tend to collapse
once pressure is placed
upon it for control purposes.
“ “
-Goodhart
Measurements

Measurement types
•  Qualification vs Quantification
•  Nominal vs Ordinal
•  Deterministic vs Probabilistic
•  Domain specific vs Domain agnostic

Scrumbando
•  Lead Time
•  Flow Efficiency
•  Customer Lead Time
•  Touch Time
•  Takt Time
•  LTH
•  CFD
•  Throughput
•  Aging of WIP
•  SLA Performance
•  Due Date Performance
•  Cause of Failure Load
•  Blocked CFD
•  Story Points CFD
•  Points to Lead time
•  More

Lead Time Historgram

CFD
CFD
shows
the

quan**es
of
work
in

progress
(WIP)
at
each

stage
in
the
system.

CFD is a visual
representation of
Little’s Law.
Average
Lead
Time
=

Average
Work
in

Progress
/
Avg

Throughput


Flow Efficiency
•  Calculated as touch time against
lead time.
•  It indicates how much room there
is for improvement by eliminating
waste without changing the
engineering methods.
5-40%

SLA Performance
•  Classes of Service spread risk

Due Date Performance
•  For Fixed Date class of service we should show
the answer of the question “Was the work item
delivered on time?”

Successful outcomes
at Mammoth bank
Though not all groups ultimately
applied Scrumban to their way of
working, it was ultimately adopted by
the vast majority of teams within the
RAD Group, and had also begun to
spread to other groups across the
larger IT organization within a few
short months. It was recognized as a
a simple, easy-to-use, yet powerful
framework for identifying and
managing impediments to workflow.

Successful outcomes
at Mammoth bank
Though the RAD Group did not make
up for all of its lost time on the mobile
banking app, it ultimately delivered the
second and final release just 5 weeks
past it’s original due date. As the first
release had been 10 weeks past its
due date, this was viewed as a major
achievement.

Monitoring Projects
in flight
Planned Project lead time is the sum
of the calculated project length and a
project buffer
The two essential measurements of
project performance are the
percentage of the project completed
and the amount of the project buffer
consumed.

The most important things cannot be measured“
“
-Deming, W. Edwards. Out of the Crisis.

More
•  How do roles, leadership work
out in Scrumban setups
•  What are some other catalysts
in maturing
•  Monte carlo simulations, Risk
calibration
•  Lots of contributions from many
Scrum and Kanban leaders who
really understand Scrumban
Management
Modeling
Maturing
Scrumban.io

Codegenesys.com/events
Foundations training
Advance Practitioner Training
Contact
training@codegenesys.com
http://codegenesys.com/events

Scrumban

More Related Content

Scrumban