Where scrum doesn't fit

Colm O’hEocha – AgileInnova4on

Slow Down to Speed Up Copyright © 2013 AgileInnova4on

1

Where Scrum won’t fit: Ge2ng Agile with Kanban in IT Ops and

Support

This material is Copyright Protected under the following Crea4ve Commons License: AGribu4on-‐ShareAlike 3.0 Please use this material freely and derive further work from it. BUT, aGribu4on must be given to the original author, and these same rights must govern all derived works. Please use the following aGribu4on: “Colm O’hEocha, AgileInnova4on Ltd. 2013: www.agileinnova4on.eu”

Scrum – CommiGed Sprint Work Items

Commitment: To Work Items

•  Well Groomed/Understood Items •  Acceptance Criteria/Tests Predefined •  Short Planning Horizon •  Evidence Based Planning •  Explicit Policies (e.g. DoD) •  Stable, Cross-‐Func4onal Team •  Fast Feedback (Inspect & Adapt)

Commitment to Sprint Plan: 100% Velocity means we’ll ‘fail’ half the 4me

Lets Protect Ourselves: Build in a ‘Con4ngency’ (or two)

Parkinson's Law: Work expands to fill 4me allocated to it

Work Takes Longer

How we Build Predictability (and commit):

Copyright © 2013 AgileInnova4on

Scrum – Forecast Sprint Work Items

Commitment: To Goal Forecast: To Work Items


Commitment to Sprint Goal: 100% Velocity means we’ll never ‘fail’

No need to build in a Con4ngency. Precise Es4ma4on not so important

Predictable delivery of Sprint Goal, not individual work items

Team Output Improves



Scrum –Some Un-‐Planned Demand

Commitment: To Goal Forecast: To Work Items (70-‐80% of Velocity)


Commitment to Sprint Goal: 70-‐80% Velocity means we’ll (probably) never ‘fail’

Con4ngency already built in (for un-‐planned work). Precise Es4ma4on not so important

Predictable delivery of Sprint Goal may be undermined (un-‐planned work)

Weakens Planning, dilutes Visibility & Predictability from the Team



Some Scrum Assump4ons (that we may not be able to sa4sfy)

•  All work items are known, understood and es4mable (for the next sprint) – Bugs (soeware, hardware, infrastructure,…) – Deploying to Customer Equipment –  Integra4ng Hardware from New Suppliers,…

•  All work items can be right-‐sized to a sprint •  All work Items can be delivered at Sprint End •  We work in cross-‐func4onal teams 7±2


Agile/Lean Delivery Models

Timeboxed Scrum, XP, FDD,… Flow Based

Kanban,…?

•  Work Planned based on Available Capacity •  Scope Fixed, Dura4on Variable •  Commitment -‐> Waste Elimina4on •  Macro & Micro Level Pull

•  Work Planned to fit the Timebox •  Dura4on Fixed, Scope Variable •  Commitment -‐> Es4ma4on & Planning •  Macro Level – Pull; Micro Level -‐ Push


Sprint Backlog

To Do In Progress Done

t1 t2 t3

t5 t6 t7 t4

t9

t8

t10

t13

t11

t12 t14

S1

S2

S3

S4

Scrum Board

Input Q (4)

Find It (2) Fix It (4) Test (3) Ready to Deploy

Deployed

In Progress Done In Progress Done

1

2

3

5

6

7

4 9 8 10

13

11

12

14

Kanban Mechanics in a Nutshell

IteraLon Based (Scrum) Flow Based (Kanban)

Focus on maximising work done in an sprint (max Velocity)

Focus on minimising the 4me to get a piece of work from start to finish (min Cycle Time)

WIP limited indirectly (per sprint) WIP limited directly (per workflow state)

Somewhat Predictable, Es4mable Work Interrupt Driven, Highly Variable Demand

Sprints Immutable (by Work Items or Goal) Finish something before you start something else (expedi4ng excep4onal)

Scope extends to work of the Team Scope extends across a Value Stream

Scope Can Vary Scope Oeen Fixed

Predictability Paramount Responsiveness Paramount

Any type of (Plannable) Work Design for the Work Expected: WIT, CoS, SLA

Coupled I/O cadence Decoupled I/O Cadences

Periodic Retrospec4ves Stop the Line – Fix as it Happens

X-‐Func4onal, Generalizing Teams (7±2) High Specializa4on with Hand-‐Offs Possible

Scrum & Kanban Summary -‐ Differences

10 Copyright © 2013 AgileInnova4on

•  Scrum –  Teams may struggle to break up work, deliver poten;ally releasable product, achieve transparency & joint responsibility

– Management may struggle to allow self-‐organisa;on, variable scope and to maintain sprint discipline

•  Kanban –  Teams may struggle with a lack of prescribed structure (e.g. cadence of events, Scrum roles)

– Management may struggle with a lack of commitment, milestones and ‘big-‐bang’ change

11

Both are Simple, Neither are Easy


•  Visualising your value stream •  Defining WITs, CoS, Policies •  Managing Flow – Balancing Demand to Capacity, Handling BoGlenecks

•  Stakeholder Engagement Models •  Input & Output Cadences •  Metrics & Repor4ng •  Scaling Kanban •  Managing Variability

Kanban for Complex Domains:


Colm O’hEocha AgileInnova4on Limited

www.agileinnova4on.ie [email protected]

•  Services –  Pre and Post Agile Adop;on Assessments –  Working with Senior Management on agile adop;on strategies, re-‐structuring, re-‐skilling

–  Kanban Bootstrap – a structured approach to implemen;ng Kanban

•  Training and Coaching –  Professional ScrumMaster and PMI Agile Cer4fied Prac44oner –  Succeeding with Agile and Scrum – an Agile Primer for the Team –  Geung Lean with Kanban –  Execu4ve Briefings on Agile and Lean in IT – benefits, challenges and strategies –  Product Owner Training –  Agile Requirements – User Stories and Es4ma4on –  Agile Tes4ng

13

Agile Innova;on Lean/Agile Training, Coaching, Consul;ng


This material is Copyright Protected under the following Crea4ve Commons License: AGribu4on-‐ShareAlike 3.0 Please use this material freely and derive further work from it. BUT, aGribu4on must be given to the original author, and these same rights must govern all derived works. Please use the following aGribu4on: “Colm O’hEocha, AgileInnova4on Ltd. 2010: www.agileinnova4on.eu”

Business

Where scrum doesn't fit