PI Planning—Plan to Discover the Next PI – Improving SAFe Journey

Note: This is the third post in the Practice Makes Permanent series. Read the first post here and the second post here.

Is your SAFe® implementation slowing? Has the energy and enthusiasm faded and it feels like just one more process change? Maybe you’re just not seeing the value you expected? That’s OK because PI Planning presents significant opportunities for your relentless improvement of the SAFe journey.

If you’ve read the previous posts in this series, you know that the right kind of practice doesn’t make perfect, it makes permanent. And you know that using creative tension will help you find the right outlook to discover new opportunities and foster relentless improvement. The reality is that many Agile Release Trains (ARTs) have a distorted view of PI Planning—they see it as a readout of the already created plan or a time to direct teams toward a plan rather than the discovery session it’s intended to be. This makes it a great place to get your implementation back on track. PI Planning is an opportunity to discover the plan for the next PI. Discovery is exciting, it’s engaging, and it’s about learning.

Through my years of teaching and implementing SAFe, I’ve identified key practices to help organizations successfully discover a plan for the upcoming PI. My goal in providing these tips and techniques is to give you practical ideas on how to revolutionize your PI Planning and improve the quality of your plans.

Consider these key practices to make PI Planning a discovery session:

  • Breadth vs. depth planning
  • Good objectives start early
  • Iteration plans from PI Planning are “what if?” scenarios
  • Raise the levels evenly
  • Preconceived is pre-committed—limit pre-PI Planning

Breadth vs. Depth Planning

During PI Planning, the Agile teams are instrumental in converting the ART vision and roadmap into team-committed objectives while discovering the risks, dependencies, and delivery goals for the PI.

The anti-pattern

A very common anti-pattern in PI Planning is when teams focus on one iteration at a time, attempting to create a solid plan for iteration one, followed by a deep dive in iteration two, and so on. This is dangerous because we’re not seeing the big picture of the whole PI. And often we get to the draft plan review, and the teams don’t have a high-level, end-to-end plan, which is critical for the management (adjustment) review and problem-solving activity.

Additionally, because teams have gone iteration by iteration, concentrating on creating and loading stories, they haven’t collaborated sufficiently on with other teams on dependencies. This can potentially lead to radical plan changes during team breakouts on day two.

Implement breadth vs. depth

This is where breadth vs depth comes in. Encourage your teams to think across the whole PI and create a very high-level plan within the first 60–90 minutes of the first breakout session. This will include initial starter objectives, the discovery of some initial dependencies on other teams, high-level goals for each iteration, and some initial stories (perhaps slotted into an iteration). These activities give teams a broad, overall approach.

Now, they can use the remaining time in the breakout to improve the depth by:

  • Discovering more stories
  • Identifying more dependencies
  • Refining objectives
  • Identifying and (perhaps) mitigating more risks

This breadth vs. depth strategy ensures that there’s always a draft plan to review and helps teams align on the approach and effort distribution.

Here’s a quick recap of the key principles of breadth vs. depth:

  • Start with a high-level, end-to-end plan that’s full of holes.
  • As time permits, go back and fill in those gaps.
  • Focus on “based on what we currently know, this is our plan. However, we know we have more to learn to fill in some gaps.”
  • Raise the water level evenly. Create some story placeholders, which will generate some objectives, which will raise some dependencies, which will identify more stories, which will update objectives, and so on.

Apply SAFe® Principle #3: Assume variability; preserve options. Start with minimal constraints and a high-level approach and add details as they emerge.

Good Objectives Start Early

Team objectives are one of the key outputs of the planning effort. Objectives are a team’s opportunity to say:

“Hey, ART leadership, we saw the vision, we saw the roadmap from the top x features. Here’s our contribution to the success of the ART for this PI.”

This is a powerful opportunity to engage SAFe Principle #8 and SAFe Principle #9 and is a critical component to ensuring alignment.

Many teams struggle with understanding objectives because they’re not used to being asked, “What can you contribute to our success?” Instead, they’re used to being told “this is what needs to be done.” I like to explain objectives with an analogy.

Team objectives

Imagine you’ve just read a really powerful, thought-changing book (such as Donald Reinertsen’s Principles of Product Development Flow). Most likely you’ve read the book chapter by chapter. You want to share the power of that book with your friend, and you’ll probably share highlights and key areas and concepts from your perspective. Now, another friend reads the same book and shares their key takeaways from their perspective. While there may be similarities in key concepts (for example, Economic Sequencing) each perspective may pick up on different areas and ideas that were key to them but that maybe didn’t stand out to you.

The chapters in the book are the features. All the teams read the features and come away with their key areas and contributions. These are the team objectives. Some teams may have shared contributions across the ART (Program Objectives), but many will have specific contributions unique to their own team (Team Objectives).

Start early, iterate often

So, how do you achieve these powerful objectives within PI Planning? That’s where the phrase “good objectives start early” comes in. Objectives should start as early thoughts and concepts and grow in clarity and alignment as the breakout continues.

These are the key principles to successfully writing objectives:

  • Start writing objectives early in the first breakout.
  • Don’t wait until you have ‘perfect’ objectives; perfect is the enemy of good.
  • As you learn more about the plan during the breakout conversations, potential objectives will emerge. Write these down, no matter how incomplete they are.
  • As you learn more about the objective through further planning-fueled learning, update the objectives.
  • Don’t try to start with SMART objectives, work toward them.

Iteration Plans from PI Planning Are What-If Scenarios

A common issue that limits the self-organization aspect of SAFe Agile teams is the mistaken belief that teams are creating iteration plans that must be locked in and committed to. This is a highly critical point: teams do not commit to iteration plans in PI Planning. They’re committing to the objectives they discover, and to the Program Board, which identifies when they believe they can deliver on the features and what dependencies are needed to deliver. When teams have the false belief that they’re going to be held accountable for plans for four iterations in sequence, they become nervous, realizing they can’t really know exactly what will be needed for each iteration. They believe they’re being held to a mini-waterfall approach. Actually, the opposite is true.

We ask teams to create iteration plans so that they can discover the objectives they want to commit to, discover the significant dependencies needed to deliver the features they pull in and discover the risks that may threaten the plan. The plan they create is one of many possible ways they can deliver, and many of the details of the actual plan will surface as they execute that plan. In my experience, the specific iteration plans the teams create are about 60-percent accurate. As long as we have the significant dependencies and risks identified, that level of accuracy is good enough to get started.

Key principles around the what-if scenario approach:

  • Teams do not commit to iteration plans. They commit to objectives and the Program Board’s dependencies and Feature deliveries.
  • Teams create iteration plans to unearth dependencies, discover objectives, and learn how to deliver to the business.
  • The iteration plans will almost certainly change as we iterate through the Program Increment. Understanding that we are only identifying one of many possible scenarios to deliver on these objectives helps the teams focus on what’s important.

Raise the Levels Evenly

Closely associated with the breadth vs. depth concept is ensuring you have a good balance of focus on each component of the Draft Plan Review. Following SAFe Principle #4, we want to shorten our Plan-Do-Check-Adjust (PDCA) cycles to increase the pace of learning. If you think of each key area of focus in the team breakout, we have:

  • Creating, sizing, prioritizing user stories and enabler stories
  • Identifying and improving team objectives
  • Identifying and attempting to mitigate risks
  • Discovering, discussing, and gaining commitment to cross-team dependencies
  • Updating the Program Board as we discover new feature delivery and dependency aspects in the emerging plan

Consider each of these areas as buckets to fill. We don’t want to fill one bucket to the top and then go to the next. Instead, we want to evenly bring up the level on all buckets together. That means we want the teams to generate some stories, which can lead to updated objectives, which can lead to new risks identified or mitigated, leading to new commitments with other teams, and leading to updates to the Program Board. This order is not rigid, there’s nothing wrong with discovering a dependency, adding some stories to cover this dependency, and then updating Program Board. The idea is to make sure that we are keeping the levels (amount of recorded information) approximately even across all the buckets.

safe for business

These are the key principles to raising the levels evenly:

  • Don’t try to create a full iteration-by-iteration plan too early.
  • Use the energy and knowledge from adding to one bucket to add to the next bucket.
  • At regular intervals, step back and review the levels in each bucket. Are these relatively evenly filled? If not, do we need to revert to focus on a particular bucket?
  • At key points, review the entire plan we have created so far. Do we see major gaps? Dangerous assumptions? Note: a great time to do this is about five minutes before the next Scrum of Scrums team breakout. It provides a really clear view of the current progress that will help us identify any systemic issues in our planning process.

Preconceived Is Precommitted—Limit Pre-PI Planning

Before we start PI Planning, we need the right level of readiness. But too much can lose the discovery aspect of PI Planning. Finding that balance is a learning journey, but there are key elements to balance.

Too much pre-PI Planning:

  • Takes away from the current PI effort. The focus should be on this PI’s objectives, not pre-planning for the next.
  • Locks into a plan that is not well-informed. PI Planning is about learning, not perfecting. The best objectives come from shared learning during PI Planning.
  • Damages the team-of-teams culture in the ART.

Too little pre-PI Planning:

  • Leaves the teams anxious
  • Can waste time in PI Planning trying to answer foundational questions

The right balance of PI Planning:

  • Allows teams to ask intelligent questions during the morning briefings. The test I apply is to verify: are the questions probing and refining (good) or are they high level and more about initial alignment (not so good).
  • Ensures we have the right people in the room. For this PI’s features, we may need other shared services and assistance. Knowing this in advance helps us extend invites to the right people.
  • Sets the stage for PDCA-based learning cycles during PI Planning. When teams come into PI Planning thinking they already have the right plan, it leads to a fixed mindset for the next PI, which blocks the true discovery needed. We want the team of teams (ART) to iterate through these PDCA cycles together as they discover the plan.

I use a very simple approach—called the five-sticky rule—to help teams understand a good starting point for the level of pre-planning needed. Each team is encouraged to bring five sticky notes into PI Planning that represent potential stories. This requires the team to do enough discovery around the features, vision, and other elements needed in the next PI but keeps them from creating a deep-dive plan that loses the discovery aspect. Please note that the five-sticky rule is more of a guideline than a rule and can be adjusted as needed.

These are the key principles to finding a balance of pre-planning:

  • Prepare to create the plan, don’t pre-create the plan
  • Look for intelligent, informed questions during the briefings
  • Beware of a fixed mindset view of the plan coming into PI Planning

PI Planning is one of the Ten Critical Success Factors to achieve transformational progress with SAFe. By applying the above ideas, you can make it the dynamic, engaging, energetic event it’s intended to be. So, go make your PI Planning a discovery session!

Check back soon for the next post in the series.

About Dwayne Stroman

Dwayne is an Enterprise Transformation Coach and Trainer and SAFe® Program Consultant Trainer

Dwayne is an Enterprise Transformation Coach and Trainer and SAFe® Program Consultant Trainer (SPCT) with more than 20 years of experience. He is ultra-passionate about helping large organizations learn how to build the right products and deliver optimal value through learning and customer validation. Dwayne uses his SPCT role to help several Fortune 100 companies, as well as many growing companies in finance, retail, healthcare, and logistics, realize the benefits of a Lean-Agile mindset. Connect with Dwayne on LinkedIn.

Creating Your PI Backlog Content

Glenn Smith and Darren Wilmshurst with Radtac, a Scaled Agile Partner, co-wrote this blog post. 

At the conclusion of Program Increment (PI) Planning, we’re always reminded of something one of our colleagues always says. There’s much to celebrate because we’ve created a set of committed plans. But we first have to complete a retrospective of the PI Planning event (cue groans from everyone in the room) and we “start preparing tomorrow” for the next PI (more groans).

Moreover, the critical path for any PI Planning is the creation of the content, suitably refined and prioritized. Without this, we can’t do any planning! But what does this look like in practice? 

This blog post is aimed at coaches who need to think about the content preparation for the next PI. By that we mean SAFe® Program Consultants (SPCs) supporting the Agile Release Train (ART) and Release Train Engineers (RTEs). But more importantly, Product Management (PM) and System Architects (SA) need to create, refine, prioritize, and socialize the content supported by Product Owners (POs) and Business Owners (BOs). We will explore each of these roles in turn during the course of this post. 

The traditional siloed hierarchy of organizations can engender a ‘this isn’t my job’ attitude. Yet many people and roles need to work together to create a compelling backlog that delivers economic benefits and satisfies your customers.

The visual model below is a high-level view of the intensity of the preparation activity for each of these roles. It isn’t meant to represent the number of hours. That is, high intensity does not mean 100 percent of your time, we just expect more time spent on preparation while recognizing that there will be other things to be done.

PI Backlog Content
Preparation intensity for specific roles.

You will also notice that there is a significant spike in preparation during the Innovation and Planning (IP) Sprint for PM, BOs, POs, and the Teams. This is when PI Planning happens.

Product Management and System Architect

PM and the SA will follow a similar pattern to each other, as their roles are two sides of the same coin—one focused on the outward market and the other technically oriented. They are going to be collaborating and working closely to make sure their respected needs are met and the right balance of the work is correctly scheduled.

Crafting backlog items for an ART, whether they are Business Features or Enabler Features, follow a pattern of Creating, Refining, Prioritising, and Socialising. While overly simplistic, each step could follow the first four iterations of a PI. In the first half of the PI, expect PM and the SA to be looking to the future. This will include looking at upcoming Epics, decomposing existing Epics, and the ART roadmap and associated Architecture Runway.

A common pattern is to see poorly defined Features with weak benefit hypothesis statements and acceptance criteria. It shouldn’t be overlooked how much effort this takes to do well. This is because the work involved isn’t just writing them down in your Agile Lifecycle Management tooling, but working with BOs, talking to a wider stakeholder cohort, including customers, and reviewing market trends. Improving their own understanding of the value proposition and scope enables people on the ART to more easily deliver against it. Through the PI, their effort tapers as other cohorts take the backlog content and prepare for PI Planning.

Business Owners

BOs are key stakeholders and critical friends of the ART. As such, they gradually experience an increasing demand on their time to support creating backlog content throughout the PI—with the most involvement happening during PI Planning. As a cohort, BOs are available when needed by the likes of PM, and actively participate in the System Demos every iteration. Here, they not only get to see the progress of delivery but give feedback to help PM and the POs inspect and adapt the backlog.

We recommend that prioritization be a ‘little and often’ affair. And as it is a team sport, BOs must attend these sessions (these are the little spikes on the BO line in the model).

Product Owners

In a scaled environment, POs serve both the team and the train. In the initial periods of the PI, as you might expect, the PO has both a team execution focus and needs to support PM with Feature creation and refinement. As the content starts to get in better shape for the upcoming PI Planning, PO involvement increases, but with a shift in focus to Feature elaboration and decomposition into drafting user stories to later socialize with the team.

The Team

Through most of the PI, the team is execution-focused, although on hand for those ad hoc, short whiteboard sessions with PM, SAs, and POs. Larger demands on the team’s time should be scheduled like any other demand on their time—after all, work is work! This will be done through the use of exploration enablers in a future PI, or spikes and innovation activities that occur during the IP iteration. Either way, the outcome is gaining knowledge that reduces the uncertainty of future work.

The team’s involvement, however, peaks during the IP iteration when the POs socialize the upcoming backlog of work—the Features and the draft stories they have created. It is during the preparation for PI Planning that the team takes time to understand what is needed and answer questions that need “I’ll look in the code” to find out.

Release Train Engineer and Scrum Master

Hey wait, you didn’t forget about the RTE and Scrum Master (SM), did you? Surely they are just facilitators, we hear you say, what do they have to do with backlog items? But let’s think about this. As facilitators at the train or team level, they are key advocates for the improvement of flow and value delivery. Therefore, it is not unreasonable to expect them to create improvement items that require effort from the teams during the PI. And we know that anything that requires effort from the teams should be planned accordingly.

The items that the RTE and SM will bring to the table for inclusion will likely come from team retrospectives, the Inspect and Adapt problem-solving workshop, or from insight gained from activities like the SAFe® DevOps course.

Creating Content During PI Planning

During each PI Planning session, PM presents the vision, which highlights the proposed features of the solution, along with any relevant upcoming Milestones. While some may feel that at this point in the proceedings the content creation is over for PM, there is actually still work to do. During the planning, there will likely be scope negotiations and prioritization calls needed as the teams get deeper into understanding and scheduling in their breakout sessions.

Similarly, the BOs have a role in adaptive content creation too. Beyond providing the business context in the briefings, they will work with the team to discuss the needed outcomes from the work. And they’ll support PM and the SAs in adapting the scope from what was originally crafted—because tradeoffs need to be made during planning. Discussions with the teams during the assignment of Business Value could influence what gets produced in the upcoming PI too.

While the POs and the Teams need to sequence and plan their stories to maximize economic results, there will almost certainly be variability of scope that will need to be accommodated as new information emerges. This will involve further elaboration, negotiation, planning, and reworking of the content during PI Planning.

In addition, the model shouldn’t be followed religiously, but used to identify who, when, and by how much focus the different roles on the train need to spend to make this happen. While putting an emphasis on the quality of the backlog items is going to help your ART, it alone won’t fix your delivery problems but will act as a key enabler in doing so. 

It is important to give a government health warning at this stage: context is king! While we have given our view on the preparation activities and the intensity, your context will provide a different reflection. In fact, when creating this post, we both had a slightly different approach for prioritization based on our respective experiences. Neither is right or wrong but a reflection on the clients that we have worked with. So please treat the model we have created as a ‘mental model’ and something you can use with your trains to frame a discussion. 

The pattern, while broadly accurate, will change in some situations, particularly if you are preparing for a training launch and this is your first PI. Here, the cadence may be condensed and more focused, but this will be guided by the quality of the backlog content you already have.

A final thought and back to our colleague who says that “PI Planning starts tomorrow.” So does PI Execution. There’s no point in making a team committed to the plans that you have created and then not executing on them. Otherwise, what was the point of PI Planning in the first place?

If we’ve piqued your interest, check out this post about changing a feature in the middle of the PI. It’s a question we always get asked when we teach the Implementing SAFe® class.

About Glenn

Glenn Smith is SAFe program Consultant Trainer (SPCT)

Glenn Smith is SAFe Program Consultant Trainer (SPCT), SPC, and RTE working for Radtac as a consultant and trainer out of the UK. He is a techie at heart, now with a people and process focus supporting organizations globally to improve how they operate in a Lean-Agile way. You will find him regularly talking at conferences and writing about his experiences to share his knowledge.

Share:

Back to: All Blog Posts

Next: We’re Giving More Than a Donation for Pride Month