Portfolio management with SAFe – Tips & Tricks

I had the opportunity to sit in on a Denver Tech Center Scaled Agile Framework Meetup sponsored by my company, Travelport, where the topic of lean portfolio management was discussed. The event was facilitated by Michael Casey and Jake Case – two excellent Agile Big Picture Enterprise coaches assigned to Travelport that I have personally had the pleasure of working with in the course of my role as a Product Director performing portfolio management.

Michael Casey kicked off the presentation with a discussion of the global nature of SAFe while introducing some of the topics around portfolio management – portfolio entities, portfolio kanban, roles, planning and forecasting. Jake took a poll of the room around efficient decision making – only one person in the audience said their company does this well! SAFe can help with that at the enterprise level and in portfolio management.

If we think about why we do portfolio management and planning in the first place, these are some of the things we hope to gain:

• Planning for scope across the organizational value streams
• Capacity matching
• Balance centralized initiatives against localized input (Balance of enhancements vs defects vs enablers against capacity)
• Sound economic prioritization
• Bridge/connection to the Enterprise

What entities at the portfolio level must exist? There’s a structure to how you capture great ideas – using the epic hypothesis statement (For, who, the, is a, that, unlike, our solution). It’s a succinct way to capture the key points of your idea and relate it to the value stream of a corporation (and how we make money). This should live at the portfolio level. One anti-pattern to watch out for is the fact that three aspects of the epic hypothesis statement are often neglected – business outcome hypothesis, leading indicators, and NFRs. That is – what is the hypothesis that you expect to achieve or explore from making this investment, and what is the minimum amount of money you actually need to prove or disprove that hypothesis.

We then split into teams and worked through a group exercise around leading indicators, as applied to a mythical wristband instant payment system at Disneyworld. Leading indicators are the data we would use to see early as to whether we are on the right track or not to meet our objectives. Many of the groups came up with leading indicators around doing a pilot or a prototype or focus group. This is a great way to get some data more quickly without a huge investment, because a leading indicator is all about accruing meaningful information earlier than we otherwise would (fast feedback). Some other ideas brainstormed by Michael and Jake around how we can evaluate against a goal.:

• Instant Net Promoter Score information
• A/B testing on purchases
• Throughput rates
• NFRs are also incredibly important – so think about the “ilities” upfront at the portfolio level. Without these, you are not well-armed to supply the teams and trains with the context they need to succeed.

Our team included a healthy mix of Travelport and non-Travelport people, including some members of the Search to Sell product team that I work with.

There is also a template called a Lean Business Case that is essentially an entire business case on one piece of paper front and back. This is a tool used in lean portfolio management. There is a significant amount of CRUFT – correct, read, updated, followed, trusted – that must be involved in the Lean Business Case so that it can be managed and used on an ongoing basis by the organization. Business outcome hypothesis, leading indicators, and NFRs are part of it, but so are areas to think about sponsors, users and markets affected, Minimum Viable Product (MVP) features (what we need to build to test the hypothesis), additional potential features, and impact on products/programs/services/sales/distribution/deployment. The other thing covered is outcomes and types of return – is it market share, presence, revenue – as well as anticipated business impact in for of revenue, ROI, or other applicable financial metric. This forces us to close the loop with measurements.

Michael and Jake next introduced the concept of WSJF (weighted shortest job first, also pronounced “wizjif”). This is a concept used in practice by agile teams to compare features against each based on cost of delay and job size, to determine how to prioritize a backlog and which features to build first. WSJFs help us get to a lower effort, higher-yield MVPs sooner that is best for us from an economics perspective. It’s also about avoiding the big hairy complicated feature that produces little value – it’s really about time value of money for prioritizing work. For WSJFs, here are some rules of thumb:

• it’s not about precision
• done repeatedly
• revisit on cadence even at the portfolio level (every Program Increment)
• don’t consider sunk costs (size based on what’s left to create the value, focus on what remains)

Another pro-tip: implement WIP (Work in Process) limits at the portfolio level. You shouldn’t have multiple things happening at once because it’s always better to finish one thing and move on to the next as a way to increase throughput. This is a basic practice of Agile with an uppercase A.

Epic tracking – another tough topic to deal with in a short amount of time. What’s been done, what’s being done and what’s left to be done, including tracking progress against the plan (did it take more capacity than expected). In sizing epics to get an idea of how big something is – there are threebasic techniques:

• socialize the concept between a preliminary (SWAG)
• refined (larger than this and smaller than that relative triangulation)
• final estimate (break the epic down into the components, estimate each component and then aggregate

Another useful technique in portfolio management is something called innovation accounting, which was coined by Eric Ries in “The Lean Startup”. To make better economic decisions, we use leading indicators in order to measure the MVP and hold ourselves accountable to evaluation of previously agreed metrics, while ignoring sunk costs.

Michael next introduced the cost of pivot or preserver – when do we know? There are three basic principles to keep in mind when answering this question

• Has the hypothesis been proven? It isn’t a negative to stop work after the hypothesis didn’t prove out – this is actually a good outcome! We want to fail fast if we are going to fail.
• What were the KPIs?
• What specific items around innovation accounting have been followed and what have we learned?

The last concept of the day was around core principles for managing your Kanban board (Kanban is really just a way to manage work that has not yet been done):

• Visualize the workflow
• Limit WIP
• Manage flow
• Make processes more efficient
• Feedback loops
• Improve collaboratively

And don’t forget you will be putting all of your potential epics together in your Kanban board, including regular business epics and enabler epics (categorized around exploration, architectural, infrastructure, and compliance), and prioritizing them all against one another. If you are doing correctly you can the use lean portfolio management as a way to manage epics across release trains and capacity to prioritize work and increase velocity.

So, that’s a wrap for lean portfolio management. It was a very interesting and interactive session and much thanks to Michael, Jake, and Travelport for setting it up!