Yesterday I had the privilege of sitting on a panel on the subject of portfolio management at Lean Kanban France. Our facilitator was Thomas Lissajoux and my co-panelists were Ian Carroll, Chris Young and David Joyce.
Thomas did us the great service of asking for some notes ahead of time. I found his headings really helpful, and I’ve decided to share my notes. And here they are…
How would you define ‘portfolio management’?
I see portfolio management as managing at that kind of level where maintaining the right balancing between competing projects is as big a challenge as any individual project.
What are the specifics of a lean/agile approach?
Depressingly, I see too many people talk about portfolio management like it was a sausage factory. The story goes like this: “just pay attention to the quality of ingredients going into the machine and success is guaranteed”. I doubt that this metaphor works very well even for sausages!
A lean approach means several things:
1) Paying attention to the balance of demand and capability (supply), managing WIP relative to capacity and relative to lead time aspirations
2) Paying appropriate attention to flow, because speed, predictability and timeliness matters (to the extent that there are economic penalties when we get these wrong)
3) Expecting the way we manage the portfolio to keep evolving. That evolution needs to have some positive effects at project and organisational level that customers will recognise and appreciate, and there needs to be effective feedback in the other direction too (success and failures at project level should generate portfolio-level learning).
In which context/clients do you do portfolio management?
I’ve twice been in the role of portfolio manager myself, once with a $XXmillion annual budget, the second time at a smaller scale but with wider organisational impact.
Nowadays as a consultant I find that many problems on the ground point back to problems in portfolio management. Parts of the process get overwhelmed, and the organisational feedback loops aren’t fast or capable enough avoid real pain and cost. I’ve seen extreme cases where the whole portfolio seems to be grinding to a halt (there are reasons why this happens, a topic for another post maybe).
What about the previous situation and its shortcomings?
In many organisations, to get a project started it’s enough merely to sell the idea. The capability of the organisation to deliver and the impact on existing work is not given much consideration.
What key practices/tools/artefacts/meetings/metrics do you use?
- The sheer number of projects (relative to the capability to manage them effectively)
- Their value (with an understanding of the relationship between value and urgency)
- Planned, potential and actual burn rates (making sure we have the capacity to deliver; too often the sum of all promises doesn’t match the sum of available effort)
- Lead times
- Amount of work in certain key states
Tool-wise I’ve had success with
- A3, as a means to test project rationale
- Changes to portfolio reporting (feedback loops)
- Customer validation – for its own sake and as a catalyst for process change
- Regular customer meetings (cadenced)
What about the Kanban principles and practices at portfolio level?
They need to be applied with imagination. Add the phrase “Find ways to…“, perhaps “Find multiple ways to…”.
- Find (multiple) ways to visualise work at portfolio level (transparency).
- Find (multiple) ways to limit work in progress at portfolio level (balance).
- Look at how policies can drive better portfolio performance (transparency again).
- Pay attention to the feedback loops (and again!)
- Keep improving and evolving (collaboration)
Foundational principles (or rather, their underlying values):
1) Understanding: Make sure change is based on genuine understanding of how things work now, taking both internal and external perspectives
2) Agreement: Always have the right people (ie those with a stake in the solution) solving problems at the right level, seeking agreements that will hold in practice
3) Respect: the first two, plus creating the expectation that improvement will benefit people, not just the numbers
4) Leadership: grow the next generation of portfolio managers (growing yourself in the process)
Additionally I would add that the portfolio mindset and Kanban are really complementary. Identify
- different kinds of projects
- different risk profiles
- different stakeholders
- different budgets/appetites
Make these dimensions visible (I mean an in-your-face, Kanban’s-killer-feature kind of visible, not just a filter on a report), aim for a healthy mix of work. With success comes trust in both your ability to deliver and the underlying principles and techniques that make it happen.
What challenges did you face?
Pretty much everywhere, far too many projects and projects that take far too long to deliver.
- In one company, more projects than there were people, not just in IT but in the whole company!
- In another, projects taking multiple years, little thought given to the possibility of incremental delivery
What key advice would you have for ‘change agents’ about to dig into portfolio management questions?
- How is demand managed against capability?
- Are lead times too long? (a very different question to the one of whether projects are late – dates tend to be genuinely critical on no more than 20% of a typical portfolio – though the long projects tend also to be ones most prone to delay)
- Can you put a value on the amount of WIP (I refer to this as $WIP)? High shock value!
- Another great metric once you understand it: What is that WIP costing in delayed business opportunity and delayed feedback?
- Is the typical project process set up just to deliver to spec or to meet an evolving customer need? See recent posts Stand up meeting, thinking tool, leadership routine and Anticipating needs ahead of time.
What are the next steps? How can you improve/scale?
Quantify, visualise, sanity check. Look for imbalances. Look for sources of unpredictability, especially waiting. Look at the relationship between project size and predictability.
And remember that scale comes with addressing coordinating costs and other kinds of friction end-to-end, not from rolling out more process or adding layers on top.