Session Title: Scrumban comes to the rescue: A Case Study
Abstract: In this case study, we discuss the challenges faced by the customer and the project team and how Scrumban helped the customer navigate through these challenges. We highlight how Metrics helped the team in its planning, forecasting and identifying their Continuous Improvement steps.
9548086042 for call girls in Indira Nagar with room service
Lean Kanban India 2019 Conference | Scrumban comes to the rescue: A Case Study | Bharat Rajagopalan
1. 1
SCRUMBAN
TO THE RESCUE
A Case Study
Bharat Rajagopalan
Project Manager
Thanks to
Sudipta Lahiri for his guidance and
coaching and for his insights into the
metrics
The word-image logo is a registeredtrademarkof Software AG
2. 2
Rules of Engagement
1. Interact
Express yourself!
2. Ask questions!
Quick questions are most welcome, even if you
have to interrupt me!
4. 4
It’s about ScrumBan!
A case study of how Scrum and Kanban
co-exist – elaborated Quantitatively
An Amalgamation of two amazing
Schools of thought
Did you know?
ScrumBan was originally designed as a way to
transition from Scrum to Kanban! Stephen Covey
The 7 Habits of Highly Effective People
So then... What’s this all about?
Courtesy: www.wikipedia.org
5. 5 |
Voice of the
Customer
Voice of
the team
Tailoring applications to
client needs and focusing
on Building RIGHT
solutions vs. Building
solutions RIGHT!
Focus on “This is WHAT
we want to build” vs.
“This is WHAT we have
to build”
Kanban + Agile thinking
Bestof both worlds
ProfessionalServices
Value Streams
WIP
Visualize Workflow
Flow & Throughput metrics
Pull based execution
Scrum roles
Ceremonies and cadences
Time-boxing
Burndown Tracking
Agile Requirements (INVEST etc.)
KANBAN SCRUM
SCRUMBAN
The Origin Story
The word-image logo is a registeredtrademarkof Software AG
6. 6 |
What’s our approach?
Study/Gemba
Define/
Value Steam
Identify and
Create /
Visualize
Sustain
and Expand
Assess and
Improve
Operate
• Understand the
nature of project –
Study the
characteristics that
will help determine
a suitable delivery
approach
• Complexity,
expectations,team
structure, budget,
scope, timeline etc.
are some (of the
many butnot
limited) parameters
to be understood
• Identify what method
will enable most
effective delivery
• Create visual
boards/representation
to visualize work
• Identify and define
value streams best
suited for that
particular project
• Keep it simple –
identify what works
and use it!
• Remember thatnot
all projects are
similar – so
customize!
• ScrumBan – adopt
and implementthe
practices truly!
• Seek expertadvise
from Coaches from
time to time
• Arrange Gemba
walks – let others see
• Enable team to use
the visual
representation/board
• Identify key
practices to
improve and score
yourself(the team)
• Stay honest and
identify areas of
improvement
• Run retrospectives
and move
actionables to the
board
• Set milestones for
the next review
• Do 5 Why’s or
Brainstorm to take
actions on the
improvementitems
• Create them as
items on the board
and treat them
deliverables
• CoE, Newsletters,
Meet-up, process
champions - to
promote free flow
of ideas and
thoughts
7. 7 |
But still... why? – Our typical challenges
1
Estimation
Effort based
estimation while
calling it story
points (SP)
2
Committed
timeline for a
defined scope
from the pre-sales
cycle
Fixed Timeline
3 Visibility of
capabilities
required in future
time-boxes,
missing
Requirement
Visibility
4 Customers want
to work in a Scrum
model yet change
requirements mid-
sprint
Requirement
Volatility
7 Customers’Distributed multi-
vendoroutsourcing model
Creates
challenges of
trust and
hand-off from
one party to
the next
Hampers
continuous
flow of work
across
value
stream
Creates
defensive
barriers and
hinders
honest
collaboration
5
Tooldisparity
Customer’s tool
(e.g. JIRA) vs In-
house tool (e.g.
SwiftKanban)
6 A common language based
on data and metrics for
continuous improvement
between the customer and
Dev team, often missing
Metric driven
Improvements
8
PerceivedTransparency
No amount of WSR, MSR, Bi-weekly WSR, etc. seem to
provide necessary transparency the customer expects
8. 8 |
So, how did/do we do it?
By now, we assume that we have finalized ScrumBan as our management methodology. So, how do we proceed further?
1. Establish your Value Streams
Planning Value Stream | Pure Kanban
Story grooming and Design – High level architecture,
common module identification and reuse, NFR’s etc.
Execution Value Stream | Time-boxed Kanban
Lays out different value lanes within the Time-box (sprint)
such as planning, acceptance, demo etc.
2. Agile Requirements – Requirement gathering, grooming, backlog creation, prioritization etc.
• Focus on Customer Value
• Less Documentation
• More Collaboration
• Just in Time Definition
• INVEST
• ATDD (Acceptance Test)
• BDD (Behaviour)
• TDD
Critical Success Factor – Planning value stream:
Necessitates prioritization of requirements, defects and
production issues – sprint scope can change based on
prioritization.
Critical Success Factor – Execution Value Stream:
Defect triage, Defect severity judgement and availability
of clear requirements in stories, tech spike identification
and planning, etc.
9. 9 |
How did/do we do it? Contd...
3. Visualize your work
• Ensure ALL work
items are on the board
to enable team to use
the visualization
effectively
• Recommendation:
Resolve tool parity by
establishing Sync
between Customer’s
tool and in-house tool
10. 10 |
How did/do we do it? Contd...
• Time-box deliverables and plan the time-box in advance • We use Scrum roles, ceremonies and cadence to
stay relevant to our Customers
• Definition of Ready
(DoR) and
Definition of Done
(DoD)
• Set WIP Limits (2)
• While we have setup WIP limits – we have not
enforced WIP in our engagements yet
• We’ll tell you why later...
4. EstablishWorking modeland Rules of Engagement
11. 11 |
How did/do we do it? Contd...
Long-term planning and forecasting: For enforcing WIP
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
• Selected VS is from Ready lane to PO Acceptance
• Several lanes have flat lines
Recommendation: Exercise WIP Limits
Cumulative Flow Diagram (CFD)
12. 12 |
How did/do we do it? Contd...
Long-term planning and forecasting: Timeline vs. Backlog (Incl. Defects & Tasks)
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Cumulative Flow Diagram (CFD)
If this is the fixed backlog,
then Project completion date
is ~mid-SeptBut Demand > Capacity!
This will not converge if
this trend follows
Demand = 245 Stories/112
days = 2.18 cards/day
• Customer adds Stories
Next slide
CFD for Stories
CFD for Stories
13. 13 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Cumulative Flow Diagram (CFD)
• But the process
adds Defects
and Tasks
Any reduction here will directly bridge the
gap in TP of Demand vs Capacity…
The rate of this addition
is increasing every
month!
Recommendation
Needs offline
introspection!
CFD for Defects
Long-term planning and forecasting: Timeline vs. Backlog (Incl. Defects & Tasks)
14. 14 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Flow Efficiency How efficient is your system from Arrival to Departure of work item
FE is between the band of 40-60%
(if WIP = 1 for In-Progress lanes)
Assuming a WIP of 2-3 cards per person,
it perhaps is between 15-20%
Recommendation: Put as little
time on Estimation! It’s only 20%
of the full time.
15. 15 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Burn Down Charts Short-term (Sprint) and Long-term (Release) planning and forecasting
A well tracked Sprint Burndown not only gives you an idea of how is the
Sprint doing but also now gives you an accurate Actual Effort data for
your project!
Recommendation: Be ahead of
the burndown on Friday EOD to
absorb the impact of weekend!
Recommendation:
Extrapolate this to the rest of
Project.
Release Burndown – Similar to Sprint burndown, the Release
Burndown gives you a feel of how the release if faring. This is good
burndown for Release 0.2.
16. 16 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Cycle Time – Control Identify potential cogs that are pulling the process/flow down
A good looking Cycle Time – Control Chart,
however, the process is not consistent
17. 17 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Cycle Time – Distribution Identify potential cogs that are pulling the process/flow down
Given that we are working in time-box iterations, a
CT Average of ~12 days is good. However, the
variation makes it unpredictable!
Recommendation: If you implement WIP, it will
reduce CT, increase FE and reduce CT process
variation because of greater focus to complete one
card
CT Distribution consistent with Control Chart
18. 18 |
How did/do we do it? Contd...
5. Execute Measure Improve Execute ... – MetricsdrivenContinuousImprovement
Throughput/Velocity Productivity
Recommendation: Focus on execution variation to
control the Throughtput variation
Good data to suggest that there Throughput
variations
19. 19 |
So to Summarize...
1. Don’t hesitate go back to fix your roots
If its the value stream that needs tweaking – tweak it!
If its the WIP that needs adjustment – adjust it!
Do whatever it takes to refine the process
2. Measure and reflect
Having data is great – it gives you the power to make informed decisionand informed decisions
are far more stronger than perceptiondriven decisions
3. Experiment– Use what works butdon’tthrow away whatdoesn’t
Learn from the experiments that don’t work. Most importantly, don’t repeat them. Every well
thought process/method has something to offer– its up to yours’ and your team’s acumen to
create and succeedwith that secretingredient
“IF YOU ALWAYS DO WHAT YOU’VE ALWAYS DONE,
YOU’LL ALWAYS GET WHAT YOU’VE ALWAYS GOT!”
SO KEEP EVOLVING!