Upload
srijan-technologies
View
181
Download
1
Tags:
Embed Size (px)
Citation preview
How to Deal with Emergencies on a Lean/
Scrum Team
Mr. Avienaash Shiralige @agilebuddha
Blog: www.agilebuddha.com
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Copyrights Protected - AgileBuddha
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
!• I have 18+ years experience in IT Product & Service
companies in various leadership roles !• Agile Transformation Consultant, Trainer and Agile Coach !• I share my opinions and experiences of applying Agile and
Lean thinking in transforming organizations on my blog: http://www.agilebuddha.com !
• Consulting clients in Australia, Europe, US and India !• Developer, Tester, Dev Manager, Product Owner, BU head,
Director..... !• No-schooling my daughter, Organic Farming, Natural
Birthing, No Vaccinations etc are my interest !
• I live in Goa, India
Me - Avienaash Shiralige - Founder, Writer - Agile Buddha
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Problem Context
• Traditional waterfall projects
• Development team never had to deal with maintenance - had a separate team to handle this
• Maintenance team had to deal with the non-sense developers created - “Developer had all the fun without hangover afterwards”
• Agile projects
• Teams start delivering early and often
• Maintenance/Enhancements starts post couple of sprints
• Planning becomes hard for the upcoming sprints
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
The Problem
• Production incidents and other unplanned mayhem consume team’s time and hence team unable to achieve sprint goal
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
The Goal
To absorb a reasonable amount of uncertainty, striking balance between robustness and speed.
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 1 - Perform Triage - And Reject
• Is this a real production emergency? Many times they are not.
Examples of non-incidents:
• Sales team coming with “the deal of the century”
• Stakeholders upgrading normal requests to production emergencies in an attempt to bypass backlog negotiation
• There may be a genuine need, but that does not make it classify under production emergency
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Strong Product Owner
• Is this is a genuine production request? Fix it.
• More often issues require very strong triage and hence strong, decisive PO who can convince stakeholders
You can always wait for 2 weeks to get a new feature.
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 2 - Perform Triage - And Defer Fix Until Next Sprint
• Some issues are there in the system for a long time, but when they are found, needs a immediate fix
• Such issues can be pushed to later sprints and having short sprint duration can be very useful in such negotiation
• If you can defer the problem to the next sprint, then team will pick it up as per regular process
• Next to the decision of reject, a good PO will make sure everything that can be deferred will be Deferred.
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 3 - Reserve a Buffer To Deal With Unexpected Issues
• If you have done Solution 1 and 2, whatever left are real issues that have to be fixed ASAP.
• Reserve a buffer of points or time for unplanned work
• Start with 80% capacity - planned work and 20% - unplanned emergency work
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Buffer Rules:
• If buffer is more than 1/5 of velocity, then it will leave big gap in the capacity planning. Keep it small.
• Buffer is NOT for backlog items
• Buffer needs to be protected from unintended use(Stakeholder smell a workaround in the regular process). So perform good triage!
• Buffer overflow/underflow: Measure how much buffer is used to plan upcoming sprints/releases
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 4: Fix Root Causes, Improve Quality
• Solution 1/2/3 are in logical order to control damage
• Fix issues so they stay fixed, build-in quality so that you don’t have emergencies at all!
• If you overflow your unplanned emergencies buffer, then you have no business in continuing the sprint. Abort Sprint!
• Instead, use the sprint to fix the root cause
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Tip: Size The Team Right
• Small team performs better as it has less overheads, but it is less robust against losing members
• 10 person team losing 1 person is 10% productivity loss
• 3 person team losing 1 person is 33% loss!
• Sweet spot is 7-9 people - Small enough to absorb overhead, large enough to absorb some loss
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 5 - Swap Not Started Items
• Apply MoSCoW principle
• Include all “must haves” and some “should haves” in a sprint
• Hence “should haves” become your buffer
• PO decides if emergency is more or less important than a “should have”
• Swap with not started similar size item
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Approach 6: Consider Using Kanban
• If team is doing more maintenance than new features, then Kanban
• Scrum Board - long term features
• Kanban board - short term changes, maintenance, support etc.
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Conclusion
• All these tricks are dependent on context
• Have a broad range of options to chose from
• If it is a large team working on the product, then you can try to break them into
• Scrum team - Working on strategic long term features
• Kanban team - Working on tactical short-term changes
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Product is in Pivot - Business Emergency
• Product has to go through a course correction in terms of vision, market segment and features
• Product team must be very good with product discovery
• Ability to-do as many iterations of idea as possible before you run out of money or patience
• The challenge lies in figuring out who cares about your incredible new product
Few famous business pivots - Twitter, Flickr, Groupon, Nokia, Star Bucks, HP, Instagram
http://www.forbes.com/sites/jasonnazar/2013/10/08/14-famous-business-pivots/
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Lean Start-up: Build - Measure - Learn
The primary measure of progress during discovery isn’t delivery velocity, it’s learning velocity.
When you’re doing discovery right, it could look a little like this team at the Nordstrom Innovation Lab.
Content Copyrights Owned By Agile Buddha - www.agilebuddha.com
Product is in Pivot - Engineering Impact
• Engineering teams also need to make course corrections to deal with this uncertainty
• Stop developing new features without validating markets
• Prototype new ideas with quick turn-around
• Sometimes building more demo-able Over shippable features
• Prefer “speed Over perfection” in engineering
• Engineering teams need to re-look at few practices like
• Automating tests for new features
• Hard coding Over building generic solutions/frameworks
• Product lacks roadmap during this stage. Hence moving to short sprints or to Flow model like Kanban helps reduce Build-Measure-Learn feedback loop.
Mr. Avienaash Shiralige @agilebuddha
[email protected] www.agilebuddha.com
Thank You!
Take this conversation online by tweeting using the hashtag #SrijanWW