2. Distributed Teams – Key Principles
Among SAFe and Scrum principles, there are
three key principles that Distributed Teams
focus on:
1. Close, daily cooperation between business
people and developers
2. Principle: Apply cadence and synchronize
teams
3. Build incrementally with fast, integrated
learning cycles
Adapt and apply SAFe
practices to distributed
teams
3. Principle: Close, daily cooperation between business people
and developers
Co-located Distributed
Team
Partially-
dispersed
Distributed Team
Fully-dispersed
Distributed Team
Recommended Distributed
Team Structure
• Co-located Product Owner,
Technical Lead, Scrum
Master, and team
• Optionally, use Lead
Developer as interface
between Distributed Team
and on-site ART
Alternate Partially-dispersed
Distributed Team Structure
• Co-located Product Owner,
Technical Lead, Scrum
Master, and partial team
• Some members may
reside onshore
Not recommended
4. Principle: Apply cadence, synchronize teams
Distributed Team operated in
the same cadence as ART
Synchronize onshore/offshore
Team and ART ceremonies
Common Program and Team
Backlogs
1. Pre-PI Planning
2. PI Planning
1. Distributed Scrum Team Ceremonies
2. Scrum of Scrums (SoS) & Program Increment level
Ceremonies
1. Common Program Backlog of features
2. Team Backlog of user stories
3. Team PI Objectives that align and rollup to Program PI
Objectives
5. Pre-PI Planning process encompasses the below key steps:
T-6 Week T-5 Week T-4 Week T-3 Week T-2 Week T-1 Week
Goals/Objectives for PI, Capabilities to
Achieve PI Goal and Feature refinement
• Acceptance criteria
• Product Management review
T = PI
Planning
Product Management
input
• Product requirements
• Product Management
review • Multiple grooming sessions
• User story maturity rating
User Story Build-out & Refinement by Scrum Teams
Milestone:
product plan
reviews
with key Leads
Draft a PI Plan
• PI Capacity
• Story Allocation
to sprints
Offshore Team only
6. Time
Meeting
Participant
s
Expectation
Tel Aviv
El
Segundo
(PDT)
Tuesday 9am-
5pm
Monday
PI Planning Day 1 Tel Aviv Scrum
Teams
1. Review Features (Business and Enablers)
2. Review user stories, functional and non-functional
acceptance criteria
3. Conduct User Story Maturity Rating
4. Create Dependencies objects in Agile Craft
5. Record assumptions
6. Record risks and mitigation plan
7. Asks
Tuesday 5pm-
6pm
Tuesday
7am-8am
PI Planning Day 1 Sync Dev Leads, RTE,
Scrum Masters and
Pos
Communicate:
1. Features (Business and Enablers)
2. User stories, functional and non-functional
acceptance criteria
3. Dependencies
4. Assumptions
5. Risks and mitigation plan
6. Asks
Wednesday
9am-5pm
Wednesday
8am – 2pm
PI Planning Day 2 Tel Aviv Scrum
Teams
1. Document velocity for sprints 1-5
2. Plan features and stories for sprints 1-5
3. Record Team PI Objectives
4. Record assumptions
5. Record risks and mitigation plan
6. Asks
Wednesday
5pm-6pm
Wednesday
7am-8am
PI Planning Day 2 Sync Dev Leads, RTE,
Scrum Masters and
Pos
Communicate:
1. Present velocity for sprints 1-5
2. Present features and stories for sprints 1-5
3. Present Team PI Objectives
4. Present assumptions
5. Present risks and mitigation plan
6. Asks
Thursday
Midnight – 1AM
Wednesday
2pm – 3pm
PI Planning Day 2
Readout
Dev Leads, RTE,
Scrum Masters and
POs
Communicates:
1. Present changes to team PI Objectives
2. Present changes to PI Scope (features and stories)
3. Confirm Assumptions
4. Confirm Risks and mitigation plan
5. Asks
El Segundo PI Planning Days: Tuesday and Wednesday every 10 weeks
PI Planning for Distributed
Team
• In lieu of face-to-face PI Planning,
Distributed Team prepares PI Plan
two-days prior to onsite PI Planning
event.
• Set clear expectation, meeting time
and list of participants for these
“Distributed Team PI Planning” days.
• Setup “PI Planning Day X Sync”
meetings to align outcomes from
previous ART PI Planning day.
Sample PI Planning for Tel Aviv Team
7. Scrum Ceremonies: Plan and publish your Distributed/Onsite Scrum Ceremonies and SoS
Monday Tuesday Wednesday Thursday Friday
(T-5d) Planning Part I: Grooming
session with the team and PO on
user stories maturity ranked
(T-1d) Final prioritized list of user
stories identified and ready in
JIRA with maturity ranked and
estimated
Day 1 (T)
Start of Sprint
• S: Planning Part II – Capacity
Planning
• S: Set up Jira stories and
estimation
Day 2 (T+1d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• P: Dev/QA Post Scrum
Day 3 (T+2d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• S: Final Test Cases Id
• S: Scrum of Scrums (RTE,
Proxy RTE, Scrum Masters)
(7:30AM)
Day 4 (T+3d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• S: Mid Sprint Review
Day 5 (T+4d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• S: Scrum of Scrums (RTE,
Proxy RTE, Scrum Masters)
(7:30AM)
Day 6 (T+5d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• Backlog Refinement
(Recorded story review, pre-
estimation, slotting)
Day 7 (T+6d)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
• 60% of user stories delivered
to test team
• Recorded Tech Review for
upcoming stories.
Day 8 (T+7d)
• S: Scrum of Scrums (RTE,
Proxy RTE, Scrum Masters)
(7:30AM)
• S: (by 10:30AM PDT)
Documented Daily Scrums
(onshore and offshore)
Day 9 (T+8d)
S: (by 10:30AM PDT) Documented
Daily Scrums (onshore and
offshore)
• User Stories Complete
• Final sprint review with PO
and architects
Day 10 (T+9d)
• S: Scrum of Scrums (RTE,
Proxy RTE, Scrum Masters)
(7:30AM)
• S: Retrospective
• S: Sprint Review/Demo
• S: Recorded Sprint
Review/Demo
S = Full onshore and offshore Together
P = Partial onshore and offshore team together
> Offset teams by one day
8. Time
Titans
(Vietnam)
Mirosoft-2
(Tel Aviv)
Microsoft-3
(New Jersey)
Android 3
(Tel Aviv)
Aviato (El
Segundo +
Tel Aviv)
Miracle
Workers
(Tel Aviv)
Roku 1
(Tampa)
Roku 2
(Tampa)El
Segund
o (PDT)
Tampa/
NJ
(EDT)
Tel
Aviv
Vietnam
7:30 -- 18:30 -- -- -- -- -- SU -- -- --
7:00 10:00 18:00 20:00
SOS SOS SOS SOS SOS SOS SOS SOS
7:30 10:30 18:30 20:30
Retro (T) Retro (T) Retro (T) Retro (T) Retro (T) Retro (T) Retro (T) Retro (T)
8:00 11:00 19:00 21:00
Demo
(T10)
Demo (T10) Demo (T10) Demo (T10) Demo (T10) Demo (T10) Demo (T10) Demo (T10)
SU: Daily Standup
SOS: Scrum of Scrums
Retro: Retrospective
Demo: (Recorded)
Note: This slide covers meetings attended by teams in different locations.
Scrum Ceremonies: Plan and publish time of meetings that accommodate everyone
9. Principle: Build incrementally with fast, integrated learning cycles
Distributed Team(s) operate in
the same eco-system as any other
team.
They use common set of tools
across all teams
They Build and Test incrementally
ALM: Jira, AgileCraft
Collaboration: HipChat, Confluence
Code Quality/ Unit Test: Junit, SonarQube
Test Automation: LISA, Jmeter, Selenium
Build Tools and Repo : Maven, Nexus
Test management: Qmetry
CI/CD Orchestration: Jenkins
Monitoring and Log: AppDynamics, Splunk
Security: Fortify
NOTE : Tools are very specific to this case study
10. • Distributed Agile Teams’ success is all about giving & getting feedback
across stakeholders in the ART value stream – in timely manner!
• Distributed Agile doesn't come in one box.. It’s a journey with goals!
• Distributed Agile is more than tooling .. It involves improvising our ways of
working, process, and culture!
10
Summary
Animation/ Interactivity Notes:
Voice Over Script:
Principle #1 – Take an Economic View - SAFe is about delivering value to the customer on a frequent basis. Thus Taking An Economic View means all activities should be driven by business value. This is carried out in many areas of the enterprise, including:
Portfolio Level, where value streams are funded and agile release trains (ARTs) are assigned according to the enterprise’s strategic themes
Value Stream/Program Levels, where capabilities/features are sized and prioritized according to business value (opportunity, time-to-market, effort)
Program Increment or PI provides a mechanism for the entire ART to deliver features, resulting in a solution that can be delivered to the customer
Architectural Runway, where technical enablers are designed/implemented just-in-time -- as required by the upcoming PI -- thereby avoiding non-essential research/development efforts
Team Level, where user stories are taken into a sprint according to the features in the current PI. Hence development work is aligned with feature benefits.
A good example of this is when Product Management identifies its minimum-viable product (MVP) for each program increment (PI). Thus the teams are continually delivering working code for essential features, resulting in a solution that could be released to the customer if the business chooses to do so. In contrast, focusing on features that aren’t part of the MVP (minimum-viable product) may not be taking an economic view since it could involve non-essential work (e.g., nice-to-haves or even gold plating).
A – Potentially Shippable Product – a functional solution, built from the code produced by the teams in each of their sprints. It may be missing some key business features (e.g. trick play) or technical enablers (e.g. DRM), however the solution could be shipped via the release team for customer feedback should the business choose to do so.
B – Shippable Product – a functional solution that meets the MVP requirements and can be shipped via the release team.
C – Final Product – a functional solution that meets all the requirements specified by product management for the PI’s.
Animation/ Interactivity Notes:
Voice Over Script:
Principle #1 – Take an Economic View - SAFe is about delivering value to the customer on a frequent basis. Thus Taking An Economic View means all activities should be driven by business value. This is carried out in many areas of the enterprise, including:
Portfolio Level, where value streams are funded and agile release trains (ARTs) are assigned according to the enterprise’s strategic themes
Value Stream/Program Levels, where capabilities/features are sized and prioritized according to business value (opportunity, time-to-market, effort)
Program Increment or PI provides a mechanism for the entire ART to deliver features, resulting in a solution that can be delivered to the customer
Architectural Runway, where technical enablers are designed/implemented just-in-time -- as required by the upcoming PI -- thereby avoiding non-essential research/development efforts
Team Level, where user stories are taken into a sprint according to the features in the current PI. Hence development work is aligned with feature benefits.
A good example of this is when Product Management identifies its minimum-viable product (MVP) for each program increment (PI). Thus the teams are continually delivering working code for essential features, resulting in a solution that could be released to the customer if the business chooses to do so. In contrast, focusing on features that aren’t part of the MVP (minimum-viable product) may not be taking an economic view since it could involve non-essential work (e.g., nice-to-haves or even gold plating).
A – Potentially Shippable Product – a functional solution, built from the code produced by the teams in each of their sprints. It may be missing some key business features (e.g. trick play) or technical enablers (e.g. DRM), however the solution could be shipped via the release team for customer feedback should the business choose to do so.
B – Shippable Product – a functional solution that meets the MVP requirements and can be shipped via the release team.
C – Final Product – a functional solution that meets all the requirements specified by product management for the PI’s.
Animation/ Interactivity Notes:
Voice Over Script:
Principle #1 – Take an Economic View - SAFe is about delivering value to the customer on a frequent basis. Thus Taking An Economic View means all activities should be driven by business value. This is carried out in many areas of the enterprise, including:
Portfolio Level, where value streams are funded and agile release trains (ARTs) are assigned according to the enterprise’s strategic themes
Value Stream/Program Levels, where capabilities/features are sized and prioritized according to business value (opportunity, time-to-market, effort)
Program Increment or PI provides a mechanism for the entire ART to deliver features, resulting in a solution that can be delivered to the customer
Architectural Runway, where technical enablers are designed/implemented just-in-time -- as required by the upcoming PI -- thereby avoiding non-essential research/development efforts
Team Level, where user stories are taken into a sprint according to the features in the current PI. Hence development work is aligned with feature benefits.
A good example of this is when Product Management identifies its minimum-viable product (MVP) for each program increment (PI). Thus the teams are continually delivering working code for essential features, resulting in a solution that could be released to the customer if the business chooses to do so. In contrast, focusing on features that aren’t part of the MVP (minimum-viable product) may not be taking an economic view since it could involve non-essential work (e.g., nice-to-haves or even gold plating).
A – Potentially Shippable Product – a functional solution, built from the code produced by the teams in each of their sprints. It may be missing some key business features (e.g. trick play) or technical enablers (e.g. DRM), however the solution could be shipped via the release team for customer feedback should the business choose to do so.
B – Shippable Product – a functional solution that meets the MVP requirements and can be shipped via the release team.
C – Final Product – a functional solution that meets all the requirements specified by product management for the PI’s.
In order to ensure the PI planning event is productive for all teams, DFW has implemented a Pre-PI planning process that starts 6 weeks before the Planned PI event.
The process of Pre-PI planning at DFW starts with establishing Goals/Objectives for the PI with supporting features and relevant acceptance criteria. This is typically aligned upon by product management and architectural leadership as based on product objectives there is typically engineering groundwork that needs to happen first to enable those goals. This is what is called “Architecture” and “Enabler” features in the timeline which are specified concurrently to the above business goals.
Since there are a lot of suppliers involved with the DFW architecture, supplier input is needed ahead of time with specifying what is needed from them from a PI perspective. This typically happens 5 weeks before the PI event.
Once the capabilities have been refined and agreed upon by management, these are shared with all the Scrum teams and RTEs for allocation of work on respective trains. Based upon these requirements Scrum teams do an initial SWAG of the features sizes ensuring that they can be accomplished within a PI and then stack rank them from a priority and logical sequence perspective.
While the previous activity is happening the PO/TPO/Architect for the scrum teams starts to build out all the user stories with detailed acceptance criteria to support required Features for the PI. This activity lasts for 3 weeks and runs through to the PI planning event. After the first week, the suppliers (both internal and external) need to start to review their plans with affected Scrum teams so that both sides know what to expect for the upcoming 10 week increment. This activity culminates with a supplier plan review and readout to key leads on the trains before the PI planning event.
Note for the above user story build out and refinement activity by Scrum teams, this activity may involve multiple grooming sessions with all members of the Scrum teams and giving feedback to the PO/TPO/Architect on immature stories via a rating process and helping with gaps in acceptance criteria and/or proper sequence of activities.
Once the above activities are complete the trains are now ready for the PI planning event.
Animation/ Interactivity Notes:
Voice Over Script:
Principle #1 – Take an Economic View - SAFe is about delivering value to the customer on a frequent basis. Thus Taking An Economic View means all activities should be driven by business value. This is carried out in many areas of the enterprise, including:
Portfolio Level, where value streams are funded and agile release trains (ARTs) are assigned according to the enterprise’s strategic themes
Value Stream/Program Levels, where capabilities/features are sized and prioritized according to business value (opportunity, time-to-market, effort)
Program Increment or PI provides a mechanism for the entire ART to deliver features, resulting in a solution that can be delivered to the customer
Architectural Runway, where technical enablers are designed/implemented just-in-time -- as required by the upcoming PI -- thereby avoiding non-essential research/development efforts
Team Level, where user stories are taken into a sprint according to the features in the current PI. Hence development work is aligned with feature benefits.
A good example of this is when Product Management identifies its minimum-viable product (MVP) for each program increment (PI). Thus the teams are continually delivering working code for essential features, resulting in a solution that could be released to the customer if the business chooses to do so. In contrast, focusing on features that aren’t part of the MVP (minimum-viable product) may not be taking an economic view since it could involve non-essential work (e.g., nice-to-haves or even gold plating).
A – Potentially Shippable Product – a functional solution, built from the code produced by the teams in each of their sprints. It may be missing some key business features (e.g. trick play) or technical enablers (e.g. DRM), however the solution could be shipped via the release team for customer feedback should the business choose to do so.
B – Shippable Product – a functional solution that meets the MVP requirements and can be shipped via the release team.
C – Final Product – a functional solution that meets all the requirements specified by product management for the PI’s.