5. Scrum : Summary
Product Owner
Team
Scrum Master
Product Backlog
Sprint Backlog
Potentialy
Shippable
Product
Burn-down Chart
Product Planning
Sprint Planning
Daily Scrum
Sprint Review
Retrospective
Cardinal Rule: Work on the highest priority item first
6. Product Owner
6
Define features of the product
Decide on release date and content
Responsible for product profitability / ROI
Prioritize features according to market value
Adjust features and priority every iteration,
as needed
Accept or reject work results
17. Yahoo-Eurosport: 2008 Event Schedule
TDF
Euro
Paris-Dakar
January
Tour de France
February
Rugby 6 Nations
March
April
Rolland Garros
FOOT:
Olympic Games qualifiers
World Cup qualifiers
17
7-Nov-13
May
June
Wimbledon
Moto GP
Golf, Athletics, Cycling
Basketball
Boxing
Horse Racing
Hockey, etc
Year long project, 2-wk sprints
18. Managing the Product Backlog
Initial
Ready
Refined
R
1
End of S1
S
1
R
1
S
2
S
2
S
3
S
3
S
4
S
4
R
2
R
2
R
3
R
3
R
2
R
3
R
2
R
3
20. Example - Release Plan – initial version
Sprint 1
Sprint 2
Sprint 3
Sprint 4
Sprint 5
Sprint 6
Sprint 7
Sprint 8
Planned
Go Live
Test Env’t
Mega
Menu
Top
Nav
Bottom
Nav
Top Right
Nav
Global
Nav
(Toolbar)
Bottom
Nav
Left Nav
Authoring,
Content
Mgmt
Search
News
Rollup
Prep for
Cutover
MAT
Portal
Integration
People
Picker
Left Nav
Breadcrumbs
Ongoing activities: update taxonomy
7-Nov-13
version
VST
VST
Feedback
MAT
Feedback
Wizzard
Comms
Panel
Part 1
Comms
Panel
Part 2
Comms
Panel
Part 3
Actual
Go Live
23. Delivering Business Value
Scope Definition
Specs
Develop
QA
Regression
Deploy
Business Value
deliver business value all at once
Waterfall
ROI
ROI
Scrum
Sprint 1
Sprint 2
deliver business value in
ea iteration: Dev, QA,
docs, integration test
23
Sprint 3
Sprint 4
Sprint 5
Time
24. 64% implemented features are
rarely or never used
Sometimes
16%
Rarely
19%
Often
13%
Always
7%
Never
45%
Ref: Jim Johnson, Chairman of Standish Group, quoted in 2006 in:
http://www.infoq.com/articles/Interview-Johnson-Standish-CHAOS
Sample: government and commercial organizations, no vendors, suppliers or consultants
26. 26
Release Planning
• Projects may have multiple Releases (optional)
• Each release likely has multiple Sprints
• Stories are started and completed in one Sprint
• Stories are executed based on value and priority
28. Agile deals with
Ziv’s Law:
Humphrey’s
Law:
Wegner’s
Lemma:
Langdon’s
Lemma:
28
•Specifications will never be fully understood
•The user will never be sure of what they want
until they see the system in production (if then)
•An interactive system can never be fully
specified, nor can it ever be fully tested
•Software evolves more rapidly as it approaches
chaotic regions (without spilling into chaos)
36. Scrum Origins
Jeff Sutherland
•
•
Initial scrums at Easel Corp in 1993
IDX and 500+ people doing Scrum
Ken Schwaber
•
•
Co-presented Scrum at OOPSLA 96
with Sutherland
Author of three books on Scrum
Mike Beedle
•
Scrum patterns in PLOPD4
Ken Schwaber and Mike Cohn
•
Co-founded Scrum Alliance in 2002,
initially within the Agile Alliance
36
37. Agile & PMBOK
Agile processes reflect PMBOK process
groups: Initiating, Plan, Execute,
Monitor/Control, Close
In each iteration:
• Plan, Execute,
Monitor, Control
• Scope, time, cost and
quality management
38. Reading List
•
Agile Project Management with Scrum by Ken Schwaber
•
Agile Estimating and Planning by Mike Cohn
•
Agile Retrospectives by Esther Derby and Diana Larsen
•
Agile Software Development with Scrum by Ken Schwaber and
Mike Beedle
•
Scrum and The Enterprise by Ken Schwaber
•
Agile and Iterative Development: A Manager’s Guide by Craig
Larman
•
User Stories Applied for Agile Software Development by Mike Cohn
Sites:
•
ScrumAlliance.org
•
JeffSutherland.com
Editor's Notes
adaptive planning as opposed to traditional planning, evolutionary development and delivery, and encourages rapid and flexible response to changeRead more: http://www.crunchbase.com/company/scrumstudy-com-a-leading-brand-from-vmedu-inc#ixzz2jrzwhgqE Follow us: @crunchbase on Twitter | crunchbase on Facebook
http://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar Eric S. Raymond on software engineering methods, based on his observations of the Linux kerneldevelopment process and his experiences managing an open source project, fetchmail. It examines the struggle betweentop-down and bottom-up design. It was first presented by the author at the Linux Kongress on May 27, 1997 in Würzburgand was published as part of a book of the same name in 1999.Cost of delay is central to decision making. It drives the right conversations and decisions around cost and revenue, rather than just the unbalanced conversations of cost many IT divisions suffer today. Cost of delay is the right tool for proper prioritization decision.Lifetime ProfitabilityThe cost of delay isn’t a simple calculation for most projects. Individual project deliverables have highly speculative and uncertain value propositions (until a customer pays, uptake rate is subjective). Also its speculation how any delay will cause competition to take more of your market share impeding lifetime value. Given its hard on a single project, calculating the compounding impact on other delayed projects in your portfolio (given your staff are still finishing this one) makes for a mind-blowing forecast equation for cost of delay. This is what we want to solve for you. Using the same models you use for planning and managing projects we build a consolidated picture of revenue and cashflow considering all aspects of cost of delay.Our Cost of Delay RoadmapThe software industry is new to calculating cost of delay. Whilst we believe we are the leaders in this area (combining probabilistic forecasts for cost of delay calculations) it is an unpaved road we are travelling on. We are willing to put our plans on the table to show where we intend to be in the next few years and where we currently are -Now: Linear lost revenue calculations plus cost of developmentWe implemented this in our KanbanSim and ScrumSim v1.3.1. We correctly calculate the lost revenue from a given target date. Revenue is modeled as a fixed value per unit of time (day,week,month,year). This simple calculation gives an accurate portrayal of lost immediate revenue but UNDERESTIMATES cost of delay by ignoring compounding impact on other projects and lost market share from being late.2013: Compounding Dependent ProjectsDependent projects that are either pre-requisites or delayed by another project should add to the cost of delay calculations. We are beginning to model dependency trees in order to cascade cost of delays and the latest delivery forecasts into the cost of delay calculations. We feel that any portfolio planned without considering these cascading dependencies is flawed and looking forward to offering the first viable solution that helps muddle through the complexity of this planning process.2013-2014: Lost product life-time profits (market share, competitive pressure losses)How much would have google made (or Apple lost) if Google had Android on the market 12 months earlier? This is the type of analysis we are adding to our modeling language to obtain a more accurate picture of life time product harm by a delay. This is the holy grail of portfolio planning and prioritization and we are VERY excited to have the right groundwork to do this analysis with statistical rigor.
Risk: Uncertainty that matters,Uncertain eventProbability and Impact
Cost of delay is central to decision making. It drives the right conversations and decisions around cost and revenue, rather than just the unbalanced conversations of cost many IT divisions suffer today. Cost of delay is the right tool for proper prioritization decision.Lifetime ProfitabilityThe cost of delay isn’t a simple calculation for most projects. Individual project deliverables have highly speculative and uncertain value propositions (until a customer pays, uptake rate is subjective). Also its speculation how any delay will cause competition to take more of your market share impeding lifetime value. Given its hard on a single project, calculating the compounding impact on other delayed projects in your portfolio (given your staff are still finishing this one) makes for a mind-blowing forecast equation for cost of delay. This is what we want to solve for you. Using the same models you use for planning and managing projects we build a consolidated picture of revenue and cashflow considering all aspects of cost of delay.Our Cost of Delay RoadmapThe software industry is new to calculating cost of delay. Whilst we believe we are the leaders in this area (combining probabilistic forecasts for cost of delay calculations) it is an unpaved road we are travelling on. We are willing to put our plans on the table to show where we intend to be in the next few years and where we currently are -Now: Linear lost revenue calculations plus cost of developmentWe implemented this in our KanbanSim and ScrumSim v1.3.1. We correctly calculate the lost revenue from a given target date. Revenue is modeled as a fixed value per unit of time (day,week,month,year). This simple calculation gives an accurate portrayal of lost immediate revenue but UNDERESTIMATES cost of delay by ignoring compounding impact on other projects and lost market share from being late.2013: Compounding Dependent ProjectsDependent projects that are either pre-requisites or delayed by another project should add to the cost of delay calculations. We are beginning to model dependency trees in order to cascade cost of delays and the latest delivery forecasts into the cost of delay calculations. We feel that any portfolio planned without considering these cascading dependencies is flawed and looking forward to offering the first viable solution that helps muddle through the complexity of this planning process.2013-2014: Lost product life-time profits (market share, competitive pressure losses)How much would have google made (or Apple lost) if Google had Android on the market 12 months earlier? This is the type of analysis we are adding to our modeling language to obtain a more accurate picture of life time product harm by a delay. This is the holy grail of portfolio planning and prioritization and we are VERY excited to have the right groundwork to do this analysis with statistical rigor.
Scope Creep
Focusing on customer needs ensures:the right features are builtnot wasting effort (and resources) on features that are not neededThe figures may be different in Nestle,Principle remains:Only build the features that the client/users need