Home Events

Abstract

Remember healthcare.gov?  The Boeing 737Max anti-stall system? Traditional governance methods focusing on heavyweight documentation deliverables and earned value metrics are obviously not working.  In this session we will summarize a governance and oversight framework we put together for large-scale agile software development for the US government.  The framework is designed to be agnostic as to the choice of Agile scaling framework, and to apply to efforts in the tens to hundreds of millions of dollars.   Oversight groups often do not have the technical skills nor the bandwidth to take a deep dive and truly understand the risks associated with large scale efforts.  And all too often the status is consistently reported as green—until it is suddenly revealed that the project is 6 months late.  How can we fight our implacable enemies: complexity and entropy? Fortunately, there are some relatively simple principles and techniques we can use to proactively detect the danger signs—before it is too late.  We will review key aspects including measuring the right things, avoiding common failure modes, adopting better practices, and keeping track.  We will see how DevSecOps enables us to enforce a “broken windows” strategy.  We will explore how to use counterbalancing metrics to encourage good technical practices and prevent gaming the system.  If you are working at scale, this session is for you.

Learning Outcomes

• Understand qualitative differences between small-team agile and agile at scale: how not just more but different practices are required
• Know how to see warning signs via leading indicators to know when a large effort is getting into trouble, so we don’t wait until it is too late
• Understand which basic metrics, KPIs, and health indicators should be monitored for large-scale agile
• Explore the differences between legacy-heavy and “cold-start” environments such as the federal government and how they require different techniques for forecasting and estimation
• Review the NoEstimates approach, Kanban metrics, and how they can help for large-scale teams—especially scrum teams

Nov 20 at 8:45 AM-12:00 PM

“I didn’t know you could do that with Kanban!” This interactive workshop explores Kanban boards that can be used in diverse industries and at multiple levels of the organization. We will explore several interesting examples, including Kanban boards for managing non-IT knowledge work, facilitating multi-team coordination at scale, and even planning and communicating corporate strategy.

The workshop begins by exploring the ramifications of some alternative board designs, which will help participants learn to construct their own Kanban boards and improve the design of existing boards. Part two of the session challenges the attendees to construct a series of Kanban boards at progressively higher levels of abstraction. We will start out at the team level (“flight level” one), then scale up to multi-team (“flight level” two), and finally the portfolio level (“flight level” three). Using painter’s tape and sticky notes of all shapes, sizes and colors, we will gain an appreciation of how thoughtfully designed boards can help unlock true business agility.

Learning Outcomes

• Understand that kanban boards can be useful for coordinating efforts of and managing dependencies between multiple scrum teams
• Think about using kanban to manage strategic efforts; understand the importance of limiting WIP at each level of abstraction (team, value steam, portfolio)
• Systems thinking: Understand how well-performing teams can still be seen as non-responsive by business if they are allocated among multiple efforts
• Systems thinking: Understand how a better performing team can actually reduce effectiveness in a large project by causing dependencies/inventory to pile up for other, slower teams
• Understand some simple and effective ways agile methods can be applied throughout an organization, without requiring disruptive moves like adopting the Spotify model, or firing all middle managers
• Understand concept of flight levels: level 1/team, level 2/value stream, level 3/strategy

Learn Gherkin, the business language specially designed for automating acceptance tests and behavior driven design

BDD is a new, exciting approach to developing software that has been shown to reduce rework and increase customer satisfaction. While other testing tools focus primarily on “are we building the thing right?”, BDD tools attack the problem of software directly at its source: “are we building the right thing?” By retaining all the benefits of automated unit testing, while extending them upstream to cover requirements, we cut the Gordian knot of risk and complexity to unleash hyper-productivity.

Why is BDD so effective?

• As a form of Test driven design, BDD helps produce frugal, effective and testable software.
• As a development tool, BDD frameworks like SpecFlow provide many convenience functions and are pre-integrated with powerful libraries like Nunit and selenium to make writing tests a snap.
• As a collaboration tool, BDD helps ensure the “three amigos” (tester, analyst and developer) sync up – ahead of time.
• As a facilitation technique, BDD enables product owners to efficiently provide the team with concrete examples that clarify the true intent of a user story and define the boundaries.
• As a reporting tool, BDD captures functional coverage, mapping features to their acceptance criteria to their test results, in an attractive hierarchical presentation.

Want functional documentation? How about documentation that is guaranteed to be correct, because every feature maps to its test results? Witness the holy grail of traceability – executable specifications.

We will start with a brief “show and tell” with several examples of reports generated from BDD tools, to provide context and to immediately highlight the bottom line business value that makes an investment in BDD so worthwhile.

We will spend a few minutes talking about the pre-requisites, so attendees have an idea of what type of investment they are signing their teams up for. We will see that in return for a modest amount of investment in tools and training, very significant benefits can be realized, and the benefits compound over time.

This workshop then dives right in to Gherkin, the structured English language technique used to capture BDD specifications. We will spend the better part of the session learning the tricks and techniques that make for robust and maintainable gherkin specifications. We will review and critique lots of examples, both good and bad.

Finally, we will spend a few minutes reviewing how to organize BDD tests for best results while generating reports, and to ensure that they stay neat and tidy even as you build more and more of them over time.