Professional Documents
Culture Documents
These values are at the core of why agile works and continues to be used on projects with high uncertainty today.
- Sprint Planning
- Sprint Development
- Sprint Retro & Review
A Sprint is a timebox, or a period that is used to contain the time allowed for work to be completed. It can be anywhere from two
weeks to a month (although shorter is becoming more popular among advanced practitioners).
Sprint Planning starts with the Product Owner selecting the work to be done from the Product's Backlog. The Product Backlog is the
list of work that is prioritized by its importance, either for ongoing improvement or the completion of a new product. The Team
reviews the stories and then selects what work they will be able to complete during the sprint. This process is facilitated by the Scrum
Master who does not participate in the doing of work, but instead is focuses on enabling the team to move quickly with good
processes and best practices. The final set of stories should form a cohesive "product increment" that the team can demonstrate by
the end of the sprint.
Sprint Development begins with the daily standup. Daily standups are self-reporting of the team on what work they will get done that
day. This is usually done around a Kanban board, or other big visual information radiator (BVIR). The team opens only a few stories at
a time and work story-by-story to analyze, build, and test the work. The result by the end of the Sprint is a shippable increment that
can be demonstrated to the product owner. Meetings are facilitated by the Scrum Master, and the Product Owner determines if a
Story is complete to meet the stakeholder needs.
- Input: Sprint Backlog of stories the team commits to complete by the end of the Sprint
- Process: Daily reporting and execution against a few stories at a time: designing, building, testing and closing
- Output: A shippable product increment that can be demonstrated
Sprint Retro and Review are ceremonies to gain feedback and drive continuous improvement into the team. The first step is the
Sprint Review, where the Product Owner demonstrates the product increment from the Sprint to Stakeholders. This is an opportunity
to gain stakeholder buy-in and feedback, so the team knows its on track with the product direction. The Product Owner can also get
feedback on what should be in the next Sprint. The Sprint Retrospective or "Retro" is the second ceremony used to close a sprint. The
Retro involves the team going into a room to evaluate how the sprint went, and identify opportunities for improvement in the next
sprint. The best Sprint Retros are run as games to facilitate input from the whole team and quickly identify improvements.
All aspects of the Iron Triangle are constraints and costs to the organization. More schedule means a delay of project benefits and tie-
up of capital. More budget means more dollars or capital invested. More scope means a larger product to support or maintain for the
organization. These are all forms of cost and constrain how the work can be accomplished when they are fixed.
The three types of project management are Agile, Traditional, and Lean.
The goals and requirements of each method are essential for understanding the place of each method in the project manager's
arsenal:
- Agile - goal is speed (deliver early versions fast), and requires trust to minimize scope for fast value delivery
- Traditional - goal is efficiency (best price), and requires efficiency to deliver lowest cost on time and budget
- Lean - goal is to innovate (solve problems), and requires expertise to minimize time of delivery
False comparisons across types of projects abound. Many times the objections one hears about using Agile is that it's missing critical
elements, such as design, testing, or documentation. These are all wrong. In fact, every project must have the following to be
successful:
- Charter
- Plan
- Documentation
- Design
- Testing
Remember that we vary scope to target just what the customer needs, so we don't waste time or money in the process. That's the
power of varying scope. It's fast and limits waste by reducing the work to a minimum viable product (MVP) that meets the project
objectives (in the charter). To do this, every Agile project needs:
- Shared Vision Robust to Change (can vary scope and stay on target)
- Whole Teams (customer + cross-functional team)
- Incremental Delivery (learn by doing and using small "sprints")
- Continuous Integration & Testing (teams test increments to ensure they work)
Scrum, SAFe, or Disciplined Agile are all frameworks that help define roles and processes to scale and implement the methodology of
Agile. They provide a shared language. But the method remains the same.
This is unheard of speed in innovation and delivery. Today similar innovation would take many years, if not a decade. Kelly Johnson
did this using principles that align closely with Agile:
- Goal: select projects to reduce energy costs and use of "brown power"
- Process: evaluate and select the best projects delivering the highest "bang for the buck"
- Project: build a decision support tool quickly to enable support for selection
The scope of this project was to build decisions support systems for projects to identify and select $500M in energy investments. This
project was executed iteratively over four years for about five million dollars. The team makeup included:
- 2 cross-functional teams
- 8 contract personnel from Booz Allen Hamilton (BAH)
- 5 customer personnel from the Navy
Outcomes included a fifty dollars per dollar return on investment (ROI: 50). That means the Navy gained $50M per year because of
this project and its decisions support systems it developed. This was achieved through iteratively identifying and building the scope
needed in multiple releases:
- $20M were gained per year in savings, by building a quality management tool for projects
- $30M were gained per year in benefits, by building systems to better select projects
- Sustainability was improved by modeling where the next best projects would be with 95% accuracy
- This enabled BAH to win $10M per year in new contracts at the Navy for renewable energy management
Condor Cluster - result of large amounts of reuse and modular architectures (Agile Engineering Example)
NASA's Faster Better Cheaper Initiative - reduced scope and size of spacecraft (Lean/Agile Release Designs)
- Improving Quality Decreases Costs - lowers costly defects, customer support, and recalls
- Continuous Improvement - for the systems and people in the systems
- Pride of Workmanship - the primary driver of knowledge workers and source of quality is joy in good work
- Plan-Do-Check-Act (PDCA) - this cycle allows for testing a complex system that can't be modeled easily
One of the famous beliefs was that the Knowledge Worker is different from the Manual Laborer, because the Knowledge Worker
knows more about the work than their boss does.
Proof that TQM Works: Edward Deming turned around Ford Motor Company in 1986 from billions of dollars in losses to its first
profits in years using the TQM approach.
TPS: The Toyota Production System (TPS) was developed by Taichii Ohno that was the first true Lean system. Focus was on reducing
waste, based on lessons from TQM. The focus was on reducing the wastes in a system:
Proof that TPS Works: Toyota is a top three (3) manufacturer of cars, with a 70% employee satisfaction rating - that's more than
double the satisfaction rating of employees in the USA, which stands at 30%.
TOC: The Theory of Constraints (TOC) was developed by Eli Goldratt. It emphasizes that the system is always governed by a
bottleneck and there is a competition between local optimization and system (global) optimization. The theory states that
Throughput of the system should be the focus of managers, not "Cost Centers" that drive local optimization. His ideas are captured in
the famous book The Goal which is read widely and cited as critical to the revolution of management in the 1980s.
Proof that TOC Works: TOC was used by the BP Oil Spill Cleanup initiative to save over $200M and rapidly deploy 10,000 boats after
the Gulf Oil Spill to skim and clean oil from the surface.
Only 10% of software projects were successful in the 1970s and using waterfall we still see that half of those projects fail today.
By 1980s the Waterfall Method was being used by DoD (and continued until 1996), which resulted in the Ninety-Ninety Rule:
"The first 90 percent of coding accounts for the first 90 percent of development time,
The last 10 percent of coding accounts for the other 90 percent of development time" - Tom Cargill, Bell Labs
Important reference: Waterfall Model Probably the Most Costly Mistake in the World.
- Consolidated Up-Front Planning - RAD and DSDM did not refine, but XP does
- Iterate on Designs - design, build, test, refine
- Timeboxes - to ensure continuous, on-time delivery
- User Stories - to describe requirements (introduced as standards in XP)
- Test-Driven Development (TDD) - enables exploration of designs and refinement before release
This led to what was called a culture of "Freedom and Responsibility." By empowering employees and holding them responsible, and
through incredible innovations in cloud and deployment technologies, Netflix was able to achieve unheard of responsiveness to
customers. As a result, Netflix continues to be a market driver and the standard in movie streaming experiences.
What all of these ideas have at their core is an understanding of one core belief: Speed Wins!
When comparing the efficiency of managing the volume of material and delivery of content, or the velocity of delivering changes to
those systems at scale - Speed always wins because at scale you can't expect efficient, perfect solutions. Things will always break at
scale and you'll need to respond quickly to fix and correct those issues in your system. Therefore, all best efforts for perfect designs
up-front and efficiency models in delivery lose, and only speed (for responsiveness to customers and infrastructure issues) can drive
sustainability into the system.
https://www.youtube.com/watch?v=wyWI3gLpB8o
The California Social Services group needed an application for case management of child welfare services. These services were used
by 200,000 social workers. The services provided support for children suffering from abuse and neglect. Over 500,000 cases were
managed annually by the CA Social Services group.
For almost ten years the organization had been stuck doing paper exercises on the design of an application, never getting out of the
requirements and design phase.
18F was brought in and used a number of Agile techniques to drive the process forward:
- Modular Contracting
- Agile Development
- User-Centered Design
- Open-source practices
The result was the release of a functioning system within one year of kickoff. Ten years of paper compared to one year to deploy a
working system - now that's Agile! The project is still in its early stages, but the results have been:
- Traditional
- Agile
- Lean
Traditional Project
Often the Traditional Process is controlled further with Stage Gates, where stakeholders agree the project is ready to move from one
Waterfall Development Stage to the next. This can help give executives clear points of approving or disapproving the work; as well as
inform stakeholders of the high-level progress of development.
- Issue - could be an idea, major problem, or series of problems the client or owner foresees for the business
- Work Concept - work is formed as either a series of small challenges or one big challenge
o Issue Backlog - for support contracts, there needs to be total or sample backlog of work to support
o Technical Challenge - for technical solutions, there is a challenge set with clear performance objectives
- Bid & Proposal - (see Traditional above, it's the same; although contract types will differ)
- Work Definition
o Dispose Issues - issues are classified in urgency and impact to define the priority and who should respond
o Define Work Breakdown - team evaluates and decomposes the Technical Challenge into a work breakdown
structure or "WBS."
- Continuous Delivery
o Value streams - support for solving lots of small problems goes through a series of predefined steps to ensure
quality and drive customer satisfaction the issue is addressed.
This is often supported using a defined workflow
The issues are routed based on priority and impact to different team members
Only the client or customer or their representative can accept it as done
o Incremental Delivery - a team will incrementally work through the WBS to solve a major problem
Often use a Kanban board to solve highest priority/most uncertain work
Delivery is continuous and incremental to explore solution sets that COULD work
- Operations - solutions are deployed into operations where customers receive benefits
o Example issue solution - solving someone being locked out of their system
o Example Technical Challenge solution - deploying a new upgraded machine for manufacturing a car
- Disposal - (see Traditional above, it's the same)
2.2 Who Uses Agile?
Understanding who uses Agile also requires understanding the methods employed to control a project's constraints. This brings us
back to the Iron Triangle for the Triple Cost Constraint:
Controlling Scope
- Traditional
o Work Breakdown Structure (WBS) - controls work by concretely defining its components
Often has three levels: Product, Major Features, Feature Components
Used to define what will and will not be in a project
o Change Control Board (CCB) - controls changes to the WBS by committee review
Includes all major stakeholders
Must be organized and often slows changes to a project
- Lean
o Tickets - identify work items and their priority for response (urgency and impact)
o Requests - these are informal or semi-formal requests that could be tickets
o Notes
Both tickets and requests go into a queue for work, and are executed through a value stream
Value streams are steps to complete work (e.g. define, analyze, build, test).
- Agile
o Product Backlogs - the list of work to be done for the entire project. It's an ordered list of work increments.
o Sprint Backlog - the work that will get done during the sprint.
Note that Backlogs are used for Tickets in Lean and Stories in Agile. You can also have what's often called the "Kanban Sandwich"
where Lean processes are used to set Sprint Goals, Agile is used to manage a Product and Sprint Backlog, and then work during a
sprint is managed in a Lean process.
Controlling Schedule
- Traditional
o Estimated Tasks and Schedules - work is estimated and modeled for precedence
o Program Evaluation and Review Technique (PERT) - adds stochastic modeling of task completion
o Critical Path Method (CPM) - uses deterministic modeling to identify critical tasks for on-time delivery
o Notes
Determining the critical series of tasks helps to focus managers in traditional on the important tasks
This does not necessarily align with business importance
Schedule and scope are fixed; however, scope modeling comes before scheduling to define estimates
and dependencies in the work
A schedule is considered the primary tool in Traditional Management for controlling delivery
- Lean
o Kanban & Queues - work is managed in a list and executed based on priority
o Service Agreements - sets the priority of work by defining what is critical, major, or minor
o Notes
Together the Kanban & Queue techniques, along with the Service Agreements, allow for Lean projects
to adjust when delivery will occur for each work item.
This is intended since schedule is varied in Lean projects
- Agile
o Timeboxes - a set period of time in which the most important work is done first
o Releases and Roadmaps - sets goals for major features to be release together
o Notes
Timeboxes are used at all levels of the project to set deadlines
Sprints are given a fixed time to drive improvement
Any work not done in the timebox goes back into the backlogs
Releases and Roadmaps set objects for multiple sprints that can be met at varying quality levels
This allows for the most important work to achieve an objective to get done first
This aligns the business importance with what work actually gets done on time
Controlling Budget
- Traditional
o Earned Value Management (EVM) - compares current performance to the plan
Planned Value (PV) - shows the cost over time expected to complete the work on schedule
Earned Value (EV) - shows how much work is completed to date
Actual Cost (AC) - shows the cost so far to complete the work
o Cost Centers
Evaluates the differences in performance by cost center
Cost Performance Index (CPI) is the factor EV / AC, where above 1.0 means good performance
Schedule Performance Index (SPI) is the factor (EV / PV), where above 1.0 means good performance
This allows you to estimate the costs or savings expected for on-time delivery of total scope
- Lean
o Service and Severity Levels - sets the level at which the company reaps benefits from the solutions
Service Levels set the Goal
Severity Levels set the Impact of meeting or not the goal for different problem types
o Key Performance Indicators (KPIs) - evaluate performance against goals for set time periods
If the KPI is meeting or exceeding the Service Level Goal, then you're making money
KPIs will often vary over time, so it's important to look at trending
- Agile
o Return on Investment (ROI) - the net income as a ratio to total investment
Positive ROIs should be expected after the first or second release of a product
Allows for selecting and refining the backlogs
o Burndown Charts - shows progress in achieving the backlog over time
Used for projects that haven't yet released the project, or cannot easily estimate ROI
Projects often start with a set of stories and story points estimated
The expectation is a linear burndown - meaning a linear decrease in total remaining work to be done
Teams often are slow at the beginning and speed up over time, or hit snags that stall backlog burn
Goals Predictable (Low Cost) Speed (Maximize ROI) Innovation (Problem Solve)
Traditional
- Typically Large
- Many connections between stakeholders
- Impact many stakeholders
o Many departments
o Many technologies and concerns
- Good Examples are
o Large building construction
o Military platform acquisition
o Government civil works projects
- Planning is done using a Master Schedule
- Goal is to be predictable and efficient (low cost)
Agile
Lean
- Typically Small
- Good examples
o Building a new solar panel
o Selling or closing a deal
o How we manage ourselves
To Do: Get the dog from the vet
Doing: Pick up kids from school
Done: Bought the presents for the kids' party
- Planning Uses Value streams and Lists
o Setting up a process (e.g. To-Do, Doing, Done)
o Establishing a backlog of work to go through that process
- Goal is to be responsive and innovate (problem solve)
A key question we should ask after considering is why do these groups form around size? Why are the Traditional projects large? The
Agile projects medium in size? The Lean projects small?
Asking these questions drives us to the next lessons on key concerns: Customers and Engineering.
2.4 Comparing Methods of Costumer Management
Traditional Agile Lean
Traditional
Agile
Hopefully these new details are opening up one clear set of constraints on Project Management Methods: The Customer. Scheduling
and managing resources using traditional means, with large meetings and formal communication can really make sense at scale; but
also when you have many differing concerns in requirements (stakeholder groups), or an unavailable customer on large, complex
work. Agile works well when the customer prioritizes your project and there can be a single, decision-making representative. Lean
makes sense when the customer needs immediate attention, but at uncertain intervals. Lean also makes sense when the work and
complexity is very small (although the expertise level may be very high).
But is it always up to the customer? What about the product? What kind of engineering is required or resultant based on these
methods to maximize their potential returns?
Traditional
Agile
Understanding the impact that the chosen PM Method has on Engineering is essential for any project manager. Depending on your
domain of engineering different types of engineering may be possible or nearly impossible. Can one incrementally build a house? It
wouldn't be allowed by the zoning boards, but it would also be dangerous!
At the same time, you only need a single general contractor and a small team to get those repairs done before putting a house on the
market. You may have a fixed budget and need the best ROI to get that house at maximum attractiveness so buyers offer you the
best bang for your investment buck.
And when it comes time to sell, can we understand the inefficiency of always re-writing the typical contract used for closing a house?
Why redesign a whole new contract when you can just update and modify a template? The legal profession lives on templates.
Every aspect of work has its natural fit, but knowing that Agile delivers the maximum profit - can we make our projects more Agile-
ready? What does that require from the customer and the design?
Hopefully this has you thinking about scope, schedule, and budget in new lights and you've enjoyed learning about who uses Agile
today - will you tomorrow?
Week 3: How to Scrum And Be Agile?
- Project Objectives - what the sponsors and/or customers expect from this project
- Stakeholders - who "has a stake" from sponsors to customers and why. Includes technical, security, business, and
operational stakeholders
- Constraints - what must the project also do or do not in accomplishing the objective, such as standards, interfaces, and
dependencies
- Risks - what are major risks (internal and external), this includes business, technical, political, social, environmental
- Definition of Done - the agreement among stakeholders of how work is closed by the Product Owner and supporting
stakeholders
Once the Team Charter is in-place, the team should assemble based on the skills needed to do the work:
- Product Owner - person responsible for managing the Product Backlog; can only be one person;
- Scrum Master - person responsible for facilitating and keeping the team on track; leads Sprint Planning and Retrospectives
- Development Team Member - person on the team who takes responsibility for the team's success in completing the Sprint
objectives
These are the general types of team members on a Scrum Team; however, we know that many times we need variations based on
our organization. The following are typical variations
- Types
o Business Representative - from the business line using the product, or subject matter expert (SME) for customer
o Technical Expert - from the systems engineering or technology group of the customer, has experience with
building for business lines
o Sponsor - the executive, director, or product manager for a business line (internal or external); focuses on
marketing and feature analysis
o Business Owner - the owner of the business selling the product; often a combination of other types of
representatives
- Level of Availability
o Dedicated - always on the team and available to the team; focusing on the backlog refinement whenever not
supporting other team members
o On-Call - available when needed at all times; however, may have limited time outside of team support for backlog
refinement (e.g. other duties)
o Matrixed - works on multiple products or projects, balancing time based on their own direction or the direction of
departmental goals
o Minimal - available only for Sprint ceremonies and minimal other times
o Absent - available with long lead times for set consultation hours; cannot predict if/when they will be available
- Types of facilitation
o Facilitator - facilitates planning meetings, daily scrums, coordination with stakeholders (requirements, etc.), and
retrospectives
o Project Manager - works as facilitator, as well as managing human resources; and is responsible for reporting and
project success
o Junior Project Manager - works as facilitator and is responsible for reporting and project success for the Scrum
team
o Business Analyst - works as facilitator and also provides supports for Product Owner and Development Team with
requirements elaboration
- Types of Availability
o Dedicated - soles works on the Scrum team
o Split - works across multiple Scrum teams (can be the same or different projects or product lines)
o Rotating Team Member - can be a rotating member of the Development Team who acts as the Scrum Master for
a Sprint
o Matrixed - can be part of a department with responsibilities to the department, Program, Program Management
Office, etc.
o Minimal / Absent - can be only available for running ceremonies and on-call for other facilitation
- Types of representation
o Generalizing Specialist - a team member that can perform any role as needed, but is trained in a certain skill set
(usually development)
o Developer (Hard/Software) - a technical team member who focuses primarily on building the product to
specifications
o Business Analyst / Tester - an analyst who defines and checks the work being developed by the team meets the
Product Owner's intent
o Technical Writer - an analyst who provides support for other team members in capturing notes, metadata, and
communicating work
o Architect - an experienced team member in a technical or business domains that serves as subject matter expert
(SME)
o Support Team - a team member that provides enabling technology, such as work tracking software, builds,
deployments, machining, etc.
- Types of Availability
o Dedicated - soles works on the Scrum team
o Split - works across multiple Scrum teams (can be the same or different projects or product lines)
o Matrixed - can be part of a department with responsibilities to the department, other projects or product lines, or
Centers of Excellence
o On-Call - is available as needed, when needed by the team to provide surge support or expert guidance
o Absent - available with long lead times for set consultation hours; cannot predict if/when they will be available
There are many advantages to having fully dedicated team members. Most often the objections for fully dedicating team members
arise when considering the Product Owner and the Scrum Master. It appears as if these team members have a lot of "downtime"
during the actual Spring Execution phase. This is hardly the case:
- Product Owners need to use time between ceremonies to liaison with Stakeholders
- PO need to research market changes (Political, Environmental, Social, Technological) to validate product features
- PO need to work with team members and provide quick feedback to keep the Development Team running as fast as
possible
- Scrum Masters should to facilitate meetings, especially with stakeholders - this can take a long time to research and
prepare!
- SM should to help guide requirements, development, and testing based on best practices
- SM should to help escalate issues outside the team; and address real-world obstacles that get in the team's way
- SM should always be analyzing the data of the team's performance to identify continuous improvement opportunities
- SM is often having to "prep" the stakeholders and retrain where necessary the Product Owner and Development Team
How does this compare with your organization? Do you have "professional facilitators" who can ensure productive meetings? What
about dedicated product owners to answer questions and make calls so development keeps on-track?
Consider the difference an approach like this might make, and the challenges in implementing it.
3.2 Three-Part User Story
There are three parts to a User Story:
- Value Statement
- Assumptions
- Acceptance Criteria
These work together to for the complete "User Story." It is NOT just the Value Statement.
Value Statement
The proper means of creating a Value Statement is up for debate, but usually the statements are as follows:
The "Who" is the user, either directly using or consuming outputs from the product the Development Team is building. It's important
that if this user is not clearly defined already in the Project Charter then they are added by the Product Owner, with consensus by the
Stakeholders. This ensures there's a clear understanding when say the term "End User" is put into the [Who] part of the Value
Statement. Usually "End User" is too vague. It should be a descriptive title such as "Research Librarian" or "Fighter Co-Pilot."
The "What" should be the MOST important aspect of the component being built. Often because product features serve multiple users
and needs or wants of users the User Stories can seem to compete with one another. When multiple types of value or outputs are
created by a new product feature or component, make sure that the one that is put in the Value Statement is the priority. Priorities
matter because a product that does everything for everyone in the end serves no one well.
- Most importantly the "What" or "Functionality Desired" should be stated in only business terms. Examples:
o I want to Login Using My User Name and Password --- WRONG!
o I want to access my account --- RIGHT!
The "Why It's Important" is a critical failing of most User Stories. This is true for experienced and beginning Scrum teams. The easiest
trap is to simply restate the "What" in another way. Such as "I want to Login to access my account." As stated above, the proper
"What" in this statement is "access my account." So what's the "Why?" It could be:
- To check notifications
- To assign work
- To check the rankings of my fantasy football team
- The key is that it's important and the MOST important "Why" of the feature or function. Others that are less important
belong in the Assumptions.
Finally it's important to note that all User Stories should be modular because they capture business functionality. If written correctly
these shouldn't ever be in conflict because of technical dependencies.
User Stories are owned by the Product Owner when in the Product Backlog, and then once they are moved into the Sprint Backlog
they are owned by the Development Team. The Development Team can update these stories with comments, but cannot change the
Value Statement, Assumption, or Acceptance Criteria without the Product Owner and Development Team Members' consent.
All User Stories are governed by the Project's Definition of Done. In this way the Definition of Done is the hidden "Fourth Part" of any
user story, although it is focused on process for closing a story. The Definition of Done captures all the standard requirements for
completing work defined in a User Story, such as:
- Standard approvals,
- Reviews by stakeholders
- Prototyping (if required)
- Documentation (for sustainability, reporting, etc.)
- Design constraints (for compliance, integration, etc.)
In this way the Definition of Done helps to modularize the statements within an User Story; but it also provides a clear set of
expectations around quality control and sustainability. These benefits are further elaborated in part in this course, and in greater
depth throughout the Agile Project Management Certificate program.
However, to ensure the authenticity of those objectives are met the following must happen:
The "timebox" for sprint planning is usually half a day for a two-week sprint, or a full day for a four week sprint. The same applies for
Team Retro and Reviews. This allows for one day out of two weeks (1:9) ratio to be spent on planning and review.
- Great User Story writing by the Product Owner (and/or Development Team in variations of Scrum)
- Planning Games
Great User-Story writing means that the Scrum Master and Product Owner need to work together to ensure User Story quality. This is
especially important when the team has a new Product Owner, who hasn't worked on a Scrum team or THIS Scrum team. Story
writing is extremely cultural.
The use of Planning Games is essential because it helps to facilitate and elicit creativity from the team. Often this is when many teams
will ask, "why not just talk about the stories?" The key reasons for NOT using unfacilitated discussion or "open discussion" is that it's
NOT open:
For this reason we use games in Scrum. The most common is the Planning Poker Game.
- Planning Poker Game. Key source for these rules is: http://agileinaflash.blogspot.com/2009/07/planning-poker-r.html
o Agree on a point scale
o Team briefly discusses the User Story
o Everyone picks a card in silence
o Team members reveal the card
o If an outlier exists (more than one step from the mode), then discuss
o (Optional) After two rounds take an average, and validate with a "yes/no" team vote
Many times there are disagreements about how to develop a point scale. There are valid reasons, especially when Story Points are
misused by team members or management outside the team. Here are some key points to consider:
Please note that many types of "Scaling Frameworks" disagree, with basis, on whether User Story points should be an absolute
metric, or at least shared across multiple teams. For Scrum and it's variations this is not important. Story points are solely to help the
team know how much work can be accomplished and if the team is getting faster Sprint after Sprint.
We continue to explore these concepts of how to use scaling of points and measurements of team performance in the Agile for
Control course in the Agile Project Management Curriculum. As well as briefly in next week's lessons on Scaling Frameworks.
By following these guidelines your planning will be faster, more inclusive, and more accurate than any group discussion or interview
process could hope to achieve.
- Daily Stand Ups - daily face to face communication so there's no "scheduling" of coordination
- Whole Teams - Everyone knows what the work is (we planned it together), and works on it together
- Team Ownership - Multiple team members working on the same User Story
- Limit WIP - limit the Work-In-Progress (WIP) to achieve faster, predictable development with focused time
When we look at these principles we see a lot of the Agile Manifesto Values restated within them.
Whole Teams
Team Ownership
- Teams know what work needs to get done, each supporting as needed
- Sometimes that means three Development Team Members work on one User Story
o First gathers user interface (UI) and process requirements from stakeholders, and builds UI
o Second begins working on technical solution for middle (logic layer) and backend (data management)
o Third designs and builds tests based on Acceptance Criteria (automated tests)
o Scrum Master might facilitate meetings with stakeholders or UI rapid prototyping
o Product Owner will validate the final product
- If one User Story begins to fall behind, then pair programming or "war rooms" can be established to solve the problem
o Team members know the work, and can jump in to help
o No individual ownership removes issues of "pride of authorship"
o Team reduces risk of failing to deliver critical functionality
- This "Team Ownership" ensures progress on the highest priority work, which also carries the most value for stakeholders
Limit WIP
Note that many of these techniques are borrowed from Lean approaches; especially the Kanban board for limiting WIP. This is
because once Stories are planned the team should operate like a well-oiled machine cranking through work. The goals of continuous
improvement and reducing waste definitely apply.
However, one key aspect of Scrum that differs and remains Agile is the fact that the Product Owner is on the team, and they can
accept work incrementally. Also with the full transparency of the Sprint Backlog showing what work the team must accomplish by the
end of the Sprint, the Development team can decide to manage their time based on those Sprint Requirements. These Stories also
combine to form a meaningful, shippable increment.
For these reasons, Agile offers sustainability, purpose, mastery, and autonomy to the Development Team not offered in pure Lean
approaches. These elements have been proven to be the most motivating for teams of knowledge workers.
More on these ideas of motivation and team facilitation is presented throughout the Agile Leadership Principles course in the Agile
Project Management Professional Certificate.
- Sprint Review: the Product Owner presents the completed, potentially shippable increment to the stakeholders.
- Sprint Retro: the Sprint Team collaboratively inspects the sprint and looks for ways to build on or change for the better.
The last point is often missed. People when running Reviews think that the Sprint Review is another lecture opportunity. "Look at all
the great work we're doing....in PowerPoint.." That's NOT what the customer or stakeholders want to see. They want to see the
product!
How to run a great Sprint Review:
Sprint Reviews are also great opportunities to take what can be a very insular process of "Sprinting" to get work done and bring in
those most impacted by the work.
Then, with a wonderful Sprint Review and lots of positive feedback, it's time to take that positive energy into the Sprint
Retrospective.
Sprint Retrospectives are much like Sprint Planning, they require practice. However, the benefits are enormous for teams that
perform "Retros:"
Just like in the Sprint Planning, the best way to execute the Sprint Retrospective is to play a game. This has become known as the
"Retro Game:"
The heart and soul of an Agile team can be made stronger through great Retrospectives and Reviews. However, at first this practice is
going to feel awkward for those in less open speaking cultures. That's why we need the structure and the games to organize input.
That's also why the Scrum Master needs to be a great facilitator.
This process should be in its core a team building exercise. That's why there's one more step often missed, but sorely needed:
The whole team gets out of the building (or area) and does something social together.
If there's anything to remember is that we are all people on this team. It's all about Individuals and Interactions OVER processes and
tools. As you engage in the end of Sprint ceremonies the momentum and potential fatigue from a hard-fought sprint looms over.
Don't let that inertia win. Shake it off and run great, personal Retros and Reviews that treat people like people.
In short order you'll have a high-performing team that likes the way the work and the people they get to work with every day.
Week 4: What Scrum Framework Fits Best?
This means that you've made a great investment learning Scrum, because no matter what you're likely going to need these skills in
one out of two projects you conduct professionally. In fact, if we look at the primary reasons why projects fail we see the likelihood
that Agile will continue to grow.
Here are the top reasons why projects fail according to PMI's Pulse of the Profession:
When you review these aspects, who do they align to the Agile Manifesto? How do they align to the benefits we've discussed in
Week 3 on how Scrum works?
The reality is that Agile addresses the primary challenges facing projects today directly: objectives, requirements, and
communication.
Understanding that Agile continues to grow, we also see in Verizon One's State of Agile report that most projects use Scrum:
This is because Scrum is simple, works, and is great for small teams. But as we've reviewed, there's no greater "organization" in the
Scrum model. So what about Scale? How do these organizations that use Scrum scale their Agile approaches?
There are four major types of scaling methods used today, according to Verizon One's survey:
This accounts for the major majority of methods used. Often we see that internally created methods tend to look like Hybrid
methods. In the next sections we'll explore SAFe, DAD, and LeSS. Each has the following items in common:
All differ on how to mange "Support Teams" and how to make an organization Agile.
The simplest means of understanding Agile at Scale is to look at the two methods originally used to Scale Agile:
Scrum of Scrums - teams coordinate work through sending representatives to have Daily Stand Ups
across teams.
Hybrid Methodology - teams are coordinated by using Predictive or Traditional controls, such as stage
gates, to manage delivery
Scrum of Scrums:
Scrum of Scrums Stand Ups: Teams send a representative, although usually this is a Product Owner or
Scrum Master
Originally this was proposed by Jeff Sutherland, one of the founders as Scrum and signatories of the
Agile Manifesto in 2001
"Agile Can Scale: Inventing and reinventing SCRUM in Five Companies" - Jeff
Sutherland, Cutter IT Journal, 2001
Provides story of using Scrum of Scrums at IDX, where weekly product line Scrums
and Monthly Management Scrums occurred.
Scrum of Scrums offers a very simple, and yet elegant means of scaling Scrum. In fact, the practice has been used by many very large
scale organizations to achieve organizational agility.
In the HBR Article, Agile at Scale (2018), there are examples of Scrum of Scrums spanning hundreds of people in less than one hour:
Saab's aeronautics business has over 100 agiel teams for its Gripen fighter jet ($43M project)
By the end of the the Stand Ups using Scrum of Scrum approaches, the executive team knows the
critical issues it needs to address
That's over 500 people fully reporting using word-of-mouth in less than 90 minutes
This model is having a resurgence now among many Agile practitioners. Especially as many organizations are expanding their
understanding of Agile and using Hybrid means or SAFe to transition into becoming an Agile organization.
The Hybrid model that is most commonly used by organizations is rather simple:
Stage Gates are leveraged especially for Requirements, Design, and Operations
(Deployment)
This offers a chance for the Traditional / Predictive leadership to approve the next
"Stage" of the project
Agile (Scrum) methods are used to rapidly and iteratively develop products in each stage
Development is often still iterative and incremental for speed and learning
This effectively looks like iterations between stage-gates; and is often more successful in organizations that resist Agile. Since
organizational resistance to Agile and misunderstanding of Agile are the primary reasons for Agile failing (according to Verizon One),
this makes sense for many new entrants learning Agile for the first time.
When looking at the major reasons that Agile fails, we see the following:
These statistics show why the Hybrid model is so popular among about half the practitioners of Agile in more traditional
organizations. For this reasons "Internal Agile" is a great means of proving the effectiveness of the method when there's no upper
management support. However, the only means of truly achieving agility is to have leadership support and good Agile training.
So how do we convince the leaders who are disbelievers in Agile? Show them the data:
And in terms of impact, the benefits can be four to five times the productivity and value output; as seen in the Ambysoft survey in
2013.
Now that you understand the Scrum in the World of Agile, and you know that world is here (!):
The other lessons in this week will try to help you answer the first question. To answer questions 2 and 3, we highly recommend you
sign up for the remaining courses in the Agile Project Management Certificate. Just like in this course, if you Verify you'll get tools,
additional resources, and insights that go beyond these lectures and notes to enable the change in your projects.
Sources:
Agile At Scale, DarrellK Rigby, Jeff Sutherland, and Andy Noble, Harvard Business
Review: https://hbr.org/2018/05/agile-at-scale
Agile Can Scale: Inventing and Reinventing Scrum in Five Companies, Jeff Sutherland, Cutter IT Journal:
http://static1.1.sqspcdn.com/static/f/447037/6486358/1270929593650/Sutherland+200111+proof.pdf
?token=NxTveQbRt0U%2FEaViggp0AauZLY8%3D
4.2 Exploring the Scaled Agile Framework (SAFe)
SAFe introduces many new roles to the Agile framework beyond the Scrum's three main roles of Product Owner, Scrum Master, and
Development Team member. These roles considered essential to manage the integration and flow of products across many Agile
teams running concurrently.
o System Teams - those that manage delivery and integration of products produced by individual Scrum teams
o Architecture Teams - manages and promotes the shared architecture framework across teams
o Product Manager - leads the Product Owners as the primary person in charge of targeting features and EPICs
o Release Train Engineer - leads the Scrum Masters on each of the Scrum teams, and conducts the large team or
team ceremonies
By adding these essential teams, and others when needed, many teams can work together. These teams help make up what is
termed the "Agile Release Train" or ART.
ARTs are how many agile teams work together on a single product or part of the business. For instance if there is a Financial company
that wants to develop a new mobile banking application for loans, then all Agile teams working on that application might be on the
same Agile Release Train. There then might be a separate Agile Release Train or "ART" for developing internal accounting software.
SAFe aligns ARTs to the business Value Stream. By modeling the business as a Lean process (remember Week 2), the organization can
then continuous improve how it delivers value to customers using the Plan-Do-Check-Act (PDCA) cycle. ARTs are the teams that build
and deploy changes to each step in the business value stream.
Agile Release Trains (ARTs) align to one or more similar parts of the Business Value Stream
ARTs are limited to up to 120 people, keeping on the lowside of Dunbar's number
ARTs work together through the Sprint process, with shared ceremonies at the Release boundaries
This introduces the need to coordinate very large groups through the typical Sprint processes of Planning, Development, Review and
Retrospectives. One of the most prominent aspects of SAFe is what's called the "Big Room Training" and "Big Room Planning" during
Release Planning. Each Release is called a "Program Increment" and usually takes four to six sprints.
All Agile Teams get in a room (can be up to 200 people, when accounting for stakeholders and Systems
Teams)
The Release Train Engineer organizes and coordinates the Planning Day
The Product Manager provides a shared vision, set of features, and priority for the next Release
Product Owners and Scrum Masters each play their role to execute Planning
Points are considered absolute (at least at first) to compare across teams with one point equal to one
person for one day
Teams commit to complete PI Objectives, instead of stories (which belong to the team)
Teams identify their dependencies across each other during this planning
Everyone gives a "vote of confidence" on whether they can meet the objectives, and keeps going until
the whole team puts up "5 out of 5."
Program Increment Inspect and Adapt (IA):
Humanizes management
Retrospectives are run briefly to identify the most important problem to solve
Problems are then addressed using workshops that include Business Owners, with clear outcomes and
support by leadership
o Take an economic view - Instead of just responding to customer wishes, work is evaluated in terms of cost of
delay (CoD).
o Plan on cadence, release on demand - All teams must plan together, but they can release whenever work is
ready.
o Base milestones on objective evaluation of working systems - The work is only considered done when it is fully
demoed at the system level
o Visualize and limit WIP, reduce batch sizes, and manage queue lengths - leverage the Lean principles of limiting
WIP and managing queues with small batches helps prevent turning independent teams back into departmental-
like groups
Another issue that SAFe addresses at scale is the need to continuously explore, develop, and deploy new solutions. This is embodied
in their ideals of "Continuous Everything;" which promotes the movement of potential work, work-in-progress, and done work
through Value streams.
SAFe has three four levels of implementation to help answer these ideas:
o Essential SAFe - basic SAFe with only Business Owners managing as executives often on a single Agile release train
(ART)
o Portfolio SAFe - includes a portfolio management function to align funding across teams or trains
o Large-Solution SAFe - Introduces the concepts of having Suppliers that integrate delivery along with multiple
ARTs on a Solution Train
o Full SAFe - Includes a Portfolio Management function above the Large-Solution when managing across Solution
Trains and other ARTs
To learn a little more about SAFe, check out these great videos as well that can help guide you through the
If you plan to learn more on the Scaled Agile Framework (SAFe) we're sure you'll enjoy the ride!
All materials and information are courtesy of Scaled Agile, Inc.'s general release at www.scaledagileframework.com/videos-and-
presentations/
4.2 Exploring Disciplined Agile Delivery (DAD)
Disciplined Agile Delivery starts from one single premise: being Agile doesn't permit teams to be undisciplined.
Disciplined Agile Delivery (DAD) grew from origins in Extreme Programming (XP), Scrum, and Lean. As it's original forms were being
codified it was clear that Agile was being used by many teams to avoid good practices of sustainable development. "We don't do that,
we're Agile" was a common phrase that frustrated many in the Enterprise. DAD offers a compromise that many people today find
extremely valuable in fitting into the organization.
People first
Learning oriented
Agile
Hybrid
Goal-driven
Delivery focused
Enterprise aware
Scalable
In order to achieve these goals, the Disciplined Agile Delivery process uses a hybrid framework with stages that align with the
Traditional Stage-Gate model, from concept to retirement (disposal). It's focus in generally on very large solutions that are required
for organizations or developed within large organizations that roll out a product line.
The key difference between this model and the typical Scrum Model that starts when a team is initiated with funding is as that DAD
uses a startup phase called "Inception." During this phase many important things happen to help scale:
During this phase the teams often work in functional and cross-functional groups. This startup phase allows for a shared
understanding going into the solution development or "Construction Phase."
Disciplined Agile also believes in many more roles than Scrum does:
Primary Roles:
Stakeholder - these are the same as in Scrum - anyone who is impacted by the
solution being built (owners, support, customers, etc.)
Team Lead - servant leader that coaches and helps organize delivery, and often
considered an "Agile Project Manager"
Product Owner - provides the "voice of the customer" as either the representative,
actual customer, or business line expert
Architecture Owner - can be simple as the "senior developer" or an architect; with
the goal of reducing technical debt risk at scale
Secondary Roles:
Specialist - may be the specialist in a certain technology or tool that's used in the
solution
Domain Expert - provides detailed domain expertise on critical topics for parts or
details of complex solutions
Disciplined Agile Delivery works by always having the Primary Roles (although sometimes the Architecture Owner and Team Lead are
the same person). Then it adds the Secondary roles as needed.
In order to scale across teams, DAD uses the "team of teams" model, building on the "Scrum of Scrums" concept invented by Jeff
Sutherland.
DAD Team Leads meet separately to coordinate delivery, as the Product Delivery Team
DAD Architecture Leads meet separately to coordinate architecture and remove dependencies, as the
Architecture Team
DAD Product Owners meet separately to coordinate planning, as the Product Management Team
These teams of teams models are then coordinated, as needed, by an overall Program Manager role.
From this basic scaling model, Disciplined Agile then extends these concepts to the organizational level. How can we mature the
organization into a "Learning" organization. In fact, the ideas are that the Agile model is a startup model for Disciplined Agile that can
be used to eventually create a lean-agile organization that continuously performs the "stages" of development as needed.
Concepts that are built in are the idea of scaling the "Supporting Cast." Those in the secondary role can become their own Agile
teams that produce products used by the Disciplined Agile teams to delivery new product. Support teams include everything that
could be considered "Development-Operations" or DevOps:
IT Operations
Customer Support
Security
Data Management
Release Management
These teams are then further scaled up into the total Product Management realm, deemed the "Disciplined IT" realm including
Governance and Reuse of products (COTS, GOTS, FOSS), as well as Enterprise Architecture, People Managemenet, and Portfolio
Management.
Then Disciplined Agile goes one step higher, suggesting that the entire Enterprise can behave in an Agile manner. Every part of the
organization can be its own Agile team or team of teams. This applies to Sales, Marketing, Legal, and Finance, as well as other
organizational areas.
To keep the entire organization aware of its own structure, there is a need for "Organizational Assets" and the "Knowledge Base."
This means that the Organization becomes its own market for consuming products both internally and externally. By managing a
central portal to access the key information and tools needed to run the organization, each team can run without being directive to
the others.
This is the actualization of the ideas of a adaptive learning organization. The stages of which are below (shorted from longer forms
you can find at http://www.disciplinedagiledelivery.com/dae/):
Do these ideas make sense to you? What kind of organization do you work in? You can learn more about the concepts and ideas of
Disciplined Agile Delivery here: http://www.disciplinedagiledelivery.com
Books referenced:
Disciplined Agile Delivery: A Practitioner's Guide to Agile Software Delivery in the Enterprise, Scott
Ambler and Mark Lines, 2012.
o LeSS Huge - used for up to thousands of teams delivering products together in single increments
This way the organization can avoid unnecessary "dotted lines" that take away power and control from the Scrum Teams which in
turn zaps productivity and pride of workmanship.
According to the philosophy behind LeSS, Scrum works as it is designed for one team. And it shouldn't be changed for one team. No
matter the complexity of the product, it should remain that there are three roles:
And in the LeSS these roles persist and remain the only roles for up to eight (8) teams. It's important to note that LeSS suggests that
teams be aligned to features to maximize their ability to "specialize" on the architecture of a product and work independently.
Beyond eight (8) teams only one additional role is added to support the Product Owner. This role is the Area Product Owner, and they
are needed because of the pressure on the Product Owner to manage such a large backlog is too great:
Area Product Owner (APO) - provides a buffer of work definition and act as intermediaries to help manage an "area" of
requirements
Product Owner (PO) can have up to ten (10) APOs working with them to manage requirements, forming
the Product Owner Team
Cannot override the prioritization of the product backlog items - this still belongs to the PO
At this point there is enough "scaling" that it can handle global operations of product development. Sites can have multiple teams,
and areas can include anywhere from five to ten teams. When multiplied this allows a single Product Owner to manage up to a
thousand or more team members.
The answer is using Large Solutions to De-Scale the organization. Organizations looking to create smaller, more independent teams
to Scrum will drive towards an model that performs "Fake De-Scaling." These ideas are as follows:
o We create "independent" teams with internal and external markets, everyone has customers and are consumers
of Agile teams
o For larger products, "independent Teams" need a portfolio manager to ensure alignment of efforts, because
otherwise they'll diverge
o Portfolio Management leads to integration and dependencies because there are real-word physical dependencies
in products
o Integration and dependencies drive the need for rules so teams can stay independent
o These rules re-introduce the Program Management Office (PMO) to control coordination
o Rules remove ownership and power from the Scrum teams, and quickly lead to additional complex teams to
manage them
So what does the organization then look like for the LeSS company?
For those that there will be an enabling leader who is the "Head of Product." This person will provide coordination and support using
the Servant Leadership model of management. The responsibilities include managing:
o Site Locations - ensuring the teams are supported with facilities, people, etc.
o "Undone Departments" - those departments that are still functionally aligned or separate from Scrum Teams
o Support Teams - these teams provide a lean service of support to the large teams (such as delivery or DevOps)
o Product Owner Team - led by the PO in charge of the product, this team develops and prioritizes the backlog
o Competence & Coaching - the functions of training and coaching to ensure continuous improvement
Higher levels of management are intended to provide similar types of "servant leadership" support for the organization. The goal
being to enable teams to work according to the LeSS principles:
o Large-Scale Scrum is Scrum - LeSS is just dealing with the unavoidable scale of some products, but remains scrum
o Empirical process control - processes should be subordinate to the product and market needs
o Transparency - required to drive fear out and improvement into the workplace
o More with Less - gain better teams, ownership, and reduced waste with less process and control
o Whole-products - focus on the product in delivery, not just the part of the product, so efforts are aligned
o Customer-centric - by focusing on the customer, everyone aligns to the priorities and needs while reducing waste
of other distractions
o Systems thinking - understand that the system is the focus, not the local performance of a team, so waste is
reduced and flow is optimized
o Lean thinking - continuous improvement and the Go See approach to management drive a respect for workers
and continuous improvement
o Queuing theory - understanding that long queues increase cycle time, reduce feedback speed, increase variability,
and result in multi-tasking
Sources:
o De-Scaling Organizations: https://less.works/blog/2015/05/08/less-scaling-descaling-organizations-with-less.html
o How too many rules at work keep you from getting things
done: https://www.ted.com/talks/yves_morieux_how_too_many_rules_at_work_keep_you_from_getting_things
_done#t-983376
o Requirements Areas: https://less.works/less/less-huge/requirement-areas.html
o More-with-Less: https://less.works/less/principles/more-with-less.html