LKIN17: Enabling Enterprise Agility though a Hybrid Agile Implementation Model - Jasdeep Singh
-
Upload
innovation-roots -
Category
Education
-
view
38 -
download
3
Transcript of LKIN17: Enabling Enterprise Agility though a Hybrid Agile Implementation Model - Jasdeep Singh
1
© 2017 EdgeVerve Systems Limited, Bangalore, India. All Rights Reserved. EdgeVerve Systems believes the information in this document is accurate as of its publication date; such information is subject to
change without notice. EdgeVerve Systems acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document.
Except as expressly permitted, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing,
photocopying, recording or otherwise, without the prior permission of EdgeVerve Systems and/ or any named intellectual property rights holders under this document.
Jasdeep Singh Kaler Dr. Ronen Barnahor
Case Study : Enabling Enterprise Agility
through a Hybrid Agile Implementation Model
@jasdeep_kaler
2
• Context
• Target and Results
• Implementation Manifesto
• Kanban Practices in Action
Agenda
3
Who are we
94 countries
848 million customers
1.06 billion accounts
73k branches
16.5% of the adult banking
population
460+ installations
The industry-leading universal banking solution from EdgeVerve
4
Our Context
Enterprise
Solution –
• Multi
million
LOC
• Multiple
products
Expanding in
every geo –
multiple
commitments
~4000 + 7 Locations
5
Sounds familiar when you are driving ….
– Ability to manage change
– Predictability
– Alignmentand more ...
6
Mission
"Our mission is to adopt Lean & Agile mindset
and practices, become a learning organization
focused on continuous improvement to
provide better value to our customers."
- Dr. Ronen Barnahor,
Head of Agile Business Transformation, EdgeVerve
7
What we set out for?
ThroughputDoing more with same
QualityReduce cost of
warranty & variability
Time to MarketMarket Demand,
Early Feedback
30-50 %
Predictability Scope and
timely delivery
Employee
Engagement
Better Efficiency
Better
Effectiveness
Cycle time for Customer
Feedback Early feedback, eliminate sunk cost
30-50 %
25 % 30%
Scope Flexibility Ability to pivot and respond to opportunities
30-50 %
8
• Time To Market - ~50%
– From 18M to 6M in Large Products/solution , From 6M to 3M in smaller Products.
• Business flexibility
– Planning cadence of 10 weeks – ability to change scope with minimal cost
• Throughput/ efficiency
– Feature Cycle Time by more than 50%
– 8% cost reduction per feature point
• Quality
– System Test - 46% improvement (#defects/SPs)
• Prediction
– Better prediction based on data – agile dashboard
Results from Initial ARTs
9
Context - Revisited
• Initial Goals
– Transparency
– Alignment
– Global Optimization
• Challenges
– Integrated suite
– High dependencies
– Limited visibility
– Local optimization
– Small coaching team
– Mix of Agile, Waterfall
– Pressure for quick results
10© 2017 Scaled Agile, Inc. All Rights Reserved. 4.10
The SAFe “big
picture” for
enterprise
transformation
11
Our Hybrid Model of Implementation
All programs aligned to the following practices:
– Backlog management (Epic->Feature-Story) in central tool
– Organizational rhythm (Program Increment and Iterations)
– Program level ceremonies with managers (PI & Iteration Planning,
Demos, I&A/retrospectives)
– Agile testing – first step (testers into the teams. “Done is Done”, “zero
defects”, system testing in pull (not a phase))
– Agile Program management – replace all status/progress reports
– DevOps – environments and CI set-up
• Implement SAFe program by program (all the rest managers only –
no team coaching)
12
Our Implementation Manifesto
Managers first over Teams first
Predictability over Throughput
Effectiveness over Efficiency
Business Agility over Team Agility
Shu Ha Ri over Process Invention
We are uncovering better ways of implementing scaling agile by doing it and helping others do it.
Through this work we have come to value
That is, while there is value in the items on the right, we value the items on the left more.
13
Kanban principles in our Implementation
• Visualize the workflow
• Manage flow
– Focus on queue control ( not used WIP control yet)
• Make process policies explicit
• Improve collaboratively (using models and scientific
methods)
14
Visualize the workflow - The starting point
Map ALL the work we do – for entire Solution in hierarchical backlog
Economic prioritization using WSJF Visualize everything possible to visualize in all
levels
The Stockdale Paradox:
“Confront the brutal facts ……. yet never lose faith “
- Jim Collins: Good to Great
15
Visualization at Portfolio Level
Strategic Theme 1Strategic Theme 2Strategic Theme 3Strategic Theme 4Strategic Theme 5Strategic Theme 6
Strategic Theme 1Strategic Theme 2Strategic Theme 3Strategic Theme 4Strategic Theme 5Strategic Theme 6
Strategic Theme 1Strategic Theme 2Strategic Theme 3Strategic Theme 4Strategic Theme 5Strategic Theme 6
Value Stream 1 Value Stream 2Value Stream Value Stream
16
Visualize the workflow (Epic level) - Release Kanban
17
Epic Progress - Drill Down
Epic 4
Epic 5
18
Epic Progress - Drill Down
Epic 4
19
Value Stream - Feature
20
Visualize the workflow - Feature Kanban Board
Feature
Grooming
Feature
Preplanning Feature
Implement
Feature
Testing
System
Testing
Feature
Closure(Done)
Implementation System level - PullAnalysis - Pull
NewReady
Dev
Comp
lete
Done
Done
21
Impact of Queue Lengths and Wait Times
98% Product Developers
don’t measure queues
– Donald Reinertsen
Manage the flow - Managing Queues
22
Manage the flow with Queues- Feature Kanban
Feature
Grooming
Feature
Preplanning Feature
Implement
Feature
Testing
System
Testing
Feature
Closure(Done)
Implementation System level - PullAnalysis - Pull
NewReady
Dev
Compl
ete
Done
Done
Queue – Features in Ready, Dev Complete
Program Velocity – WIP Limit in Implementation
It is much more easier to implement than “Limit the WIP” concept
23
1. System Flows
2. Business
Acceptance
Testing
3. Zero Open
Defects
1. Performance
Testing/
Security
2. Testing/Docu
mentation
3. System demo
to the stake
holders
1. Feature
elaboration
2. Three
Amigos
alignment on
MVF
3. High level
UX work flow
( I/A)
4. Acceptance
Criteria
5. WSJF and
prioritization
1. Team/Team Rep review
2. Elaborate to User Stories
with Story points and
acceptance criteria
3. UX WF/ white boarding
4. Potential Spikes (if
needed)
5. Refine Feature
Acceptance Criteria,
JobSize, WSJF (If
applicable)
6. Mark for
NFRs/Architecture/UX/..
other conditions
7. Feature Testing
strategy/ automation
1. All stories
done
2. Code
development
and review
3. Zero Open at
Unit Test and
Story Testing
Defects
1. Zero Open
Defects
2. Acceptance by
PM/UX
Making flow policies explicit - Feature Kanban Board
Feature
Grooming
Feature
Preplanning Feature
Implement
Feature
Testing
System
Testing
Feature
Closure(Done)
Implementation - Scrum System level - PullAnalysis - Pull
NewReady
Dev
Compl
ete
Done
Done
Flow Policies – A sample
24
Managing Flow
Flow Management meetings in all levels
– Three Amigos, PMO/RTE, SMs and coach
– Review Kanban board & Dashboard
– Identify potential bottlenecks, starvation etc
– Brainstorm ideas to improve flow
– Focus on breaking silos and improving flow
25
Manage the flow - Release Kanban (Features)
26
Control the Flow-in – Just in time, Just enough, “full kit”
Backlog Readiness and Quality Index
Backlog Readiness• At any point Backlog should be ready
between 1-1.5 PIs (Based on last two PIs Feature velocity)
• Anything less is risk and anything beyond is waste
Backlog Quality• Acceptance Criteria Defined?• Prioritization completed?• Story breakup completed?• Feature Sizing completed?
27
Manage the flow - CFD
28
Improve collaboratively - Improving Flow using Control Charts
• Feature Cycle Time Control Chart
– Average Feature Cycle Time
– Drill down
• Purpose
– Analyze outliers
– Identify bottlenecks
– Arrive at optimum feature size to
improve the flow
29
Improve collaboratively - Dependency Visualization
• Dependent Features– ART Features who are dependent
on other features (of same or different ART)
• Blocker Features– ART Features which are blocking
other features (of same or different ART)
• Color Highlighting– Yellow – Dependent & blocker
features planned in same iteration
– Red – Blocker feature is planned later than dependent feature
30
Continuous improvement using continuous learning
• Hypothesis based experimentation approach for continuous incremental learning
– Quality vs throughput
– Ideal Feature Size
– Feature WIP limit
– Left shifting of System Testing
– Automation approach
– I&A – define problem statements supported by data (e.g. Ave. Cycle time)
• Contextual – what works there may not work here/vice versa
31
© 2017 EdgeVerve Systems Limited, Bangalore, India. All Rights Reserved. EdgeVerve Systems believes the information in this document is accurate as of its publication date; such information is subject to
change without notice. EdgeVerve Systems acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document.
Except as expressly permitted, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing,
photocopying, recording or otherwise, without the prior permission of EdgeVerve Systems and/ or any named intellectual property rights holders under this document.
• Jasdeep Singh Kaler ([email protected])
• Dr. Ronen Barnahor ([email protected])
•
•
32
Sample Implementation Roadmap