A little walk down memory lane – how Portfolio Kanban was discovered
In 2009 Stefan Roock an me found out that it might be useful to use Kanban on the Portfolio level. We simply stumbled across the opportunity. Initially, we actually wanted to coordinate releases w/ a Kanban like system, but we felt it didn’t help much on this level without a more strategic view. Coordinating here was already too late, too much downstream. Reactive. Then one day, we thought „wouldn’t it be great if we just knew which project would be the next one to start at all times on a global level?“. What we could do on a local,project level – pull the best next feature, we wanted to be able to do on a global portfolio level: Pull the next, most important and valuable project – without guessing or being killed in discussions.. This was more a social than a „technical“ problem: none of the stakeholders in isolation could help. As, of course, in bilateral meeting each stakeholder had to think his project is most important.
Of course, at the time, we weren’t allowed to use to use Portfolio Boards (PB) in my the enterprise environment right away. After some experiments there, I introduced the concept at LKBE 2010 in Antwerp to a still new, fresh and young Kanban crowd. A fear that drove me during the preparation of the talk was, that it could possibly not be considered Kanban. While what I demonstrated was a WIP constrained pull system, it didn’t look like it. The WIP limits were not made explicit on my board and being explicit about WIP limits was the talk of the town in the early days. Of course, some stood up and asked the nagging questions, without hackling me though. David Anderson got up and understood the concept right away and thus defended its construction, that was much more abstract than Kanban on a ticket or feature level. Bliss and relief set in. As Arne and Henning were into translating the epic book Kanban by David into German, David asked me if I could write up what I presented in Antwerp and provide it as an additional chapter in the books’ German translation. He gave me the very important hint to riff along on the topic of social capital – how right he was. But read on. So we ended up with a nice description of the basic setup of a Kanban system for portfolio decisions.
Since then a lot has happened. I didn’t focus on the topic very much and kind of lost control and impact on it. On the positive side, the topic has grown and scaled quickly. Even to the point that SAFe is using the concept on the portfolio level. (Make of that what you want. I have no opinion.) On the negative side, the concept is still badly documented. Worse yet, it seems that everyone is having a different concept of portfolio in mind. Also, everyone seems to have a different purpose for PB in mind. Most tend to see a demand for monitoring a portfolio or program or department with the board. I don’t. These are different boards, different animals.
Also, since then, I have seen a lot of PBs to be implemented and degenerate quickly. The very reason was in almost all cases, that some secret, undocumented ingredients were not part of the system. So, here I give some hints on some of the most basic rather than the most advanced constraints of a PB. In fact so basic, that they are never really mentioned?
There is no English translation of my chapter in David’s German book available. So all I can point you to from my side is this deeper, longer article on infoQ, nicely edited by Ben Linders at the time. It goes deep and covers innovation, Horizons 1-3, design thinking and much more. Sadly, it also misses out all the basics and is more advanced fodder.
What I’m writing today is an explanation of some more undocumented features or constraints of a portfolio board system. Many things are on my mind, but I try to keep it short and simple from here on.
Introduction – What is a Portfolio Board?
A Portfolio Board is a WIP limited pull system for decisions on the portfolio level. I don’t call it Portfolio Kanban anymore without being forced to. The audience that is being addressed and helped by the concept does not care about Kanban too much, or worse is rather distracted by the overloading of the Kanban concept in production and knowledge work. Please don’t feel offended by this, it is a necessity in my environment.
What do I mean by portfolio? By portfolio I mean the whole of all high level projects, products, services and activities that are needed by a company to support and implement their strategy. In this sense, a portfolio board helps to operationalize the strategy of a company. I come to think of a portfolio board and all its extension as a great tool for strategy deployment. In fact, the board is the visualization of a companies’ strategy. If not, it does not do its job.
Basically, how you manage a portfolio is by creating and managing options. In a healthy portfolio , the majority of options should be of the nature of having strong to unlimited (while possibly unknown) upside (the sky is the limit) but very limited downside (or cost). Managing your companies’ portfolio this way, many things must go wrong for you to fail completely. Still it is a hard and collaborative effort. To support this, we have – voila – the board and its surrounding system.
While I am talking about collaboration, what we do with the board is to offer a system so that people in the company can quickly pull decisions from upstream that are needed to meaningfully work in implementing the strategy. The whole system is built to be able to quickly decide in a tactical way across the company. This way, we have a tool to align all necessary actions across the company along out chosen strategy.
In fact, one of the biggest issues when installing a Portfolio Board is that a strategy is not yet made explicit. (If no strategy is yet explicit, read one of the very last sentences of this post for a hint.)
Now to make things more clear, here is a picture of a Portfolio Board. While I am talking in abstract and complex terms, the board itself is easy! See:
Before we go on, a little explanation of what a PB and its purpose is not.
The purpose of Portfolio Boards
PB boards are not meant to monitor complex lifecycle processes of projects, products or services or activities in your company. If you have that desire it is perfectly ok, but this can’t be done with a PB. It is also not a program management tool. You can also do that, but it is not a PB. You can do many things that look similar, but PB is and feels totally different.
PB is not about process control it is about gaining strategic alignment and operating the strategy across your company by supporting all your teams through providing them timely decisions in a pull manner.
PB is also not about capacity planning. Another requirement that is often vital for companies and enterprises, but which again is not the of a PB. Sorry. I totally and absolutely support and understand all these needs and you can and should do all this. It is simply dangerous to do it by integrating all these aspects it into the PB system. Your PB will totally be watered down by mixing in these aspects. Don’t do this.
Main course – Some secret ingredients
Now, finally, here come the main ingredients that I observed in those cases were the board does its work:
- The board needs to be discussed in a very high frequency cadence. My preferred cadence is in fact weekly. I mean it. Get all the people you need for portfolio decisions into one room on a weekly basis. It seems hard. But it is a must. Read on to understand why. As an extreme, I could accept bi-weekly as the heartbeat of the board. Anything beyond that means you should not even start doing PB. I am really, absolutely dead serious about this. Normally at this point I get to hear that people would really like to do this and absolutely understand the necessity, but it wouldn’t work like that right here in this company. It’s what I hear about most great breakthroughs: „we love it, but it’s not possible here!“ That’s ok with me, but then you don’t do the hard work. It’s your life so you don’t have to excuse yourself to me. Just don’t tell it didn’t work afterwards, ok?
- As I just came up with the ridiculous request to have all VIPs of your company that are required for portfolio decisions in a room every week in front of a stupid techie board, I make it my obligation to waste no time of these people and keep the meeting totally, absolutely short. To me that means something along 15-20 minutes. I hear you: That’s also impossible in your case. Again: That’s ok to me, but then better don’t start. It makes no sense to discuss things longer than this. But it makes all sense in the world to discuss things that shortly and provide all information necessary in a very brief, concise way.
- Don’t mix in monitoring aspects. The 15 minute meeting is a decisions only meeting. It is not a monitoring or status update meeting. Things will simply be decided. And only issues that are pulled. This is why it can work in 15 minutes. It is a no junk, no confusion, no bullshit, no politics 15 minutes meeting to decide the most necessary portfolio issues. Project A or Project B? Product increment F vs. service initiative K? We have a nice 14 days product development, R&D or XYZ slot available in 3 weeks time. What to do with it? Again: I hear you. You have all these people in the room. They crave for monitoring, status updates etc. Talk to them while having a coffee tomorrow evening, get better at being transparent about your status anyways, so that they don’t crave for all the information they should have at their hands anyway to. BUT.PLEASE.DON’T.ABUSE.THIS.MEETING. (Sorry, I did not want to scream at you!) This meeting is about making decisions in a WIP limited pull system for portfolio decisions, remember? (Sorry, I did not want to sound like repeating myself over and over again.) And again, tell me it’s not possible at your place. Then.don’t.start. It’s fine. Work on getting there. I know how tempting it is to use the same meeting to get great messages across, to report on how great people are, how things were harder than expected and so on. It just is not the right place. believe me! It will be great to have a decision rather than a mixed reporting ‘ decision meeting.
- Only discuss things that are to be discussed, where teams need the decision or status updates. Do not discuss the whole board. do not discuss all initiatives on it.
- Whoever facilities the meting needs a strong mandate from the CEO to do so. He will need to ask CEO questions. If he does that without an explicit mandate, this meeting will kill him. seriously. Get a mandate for CEO questions when facilitating the meeting. It might even be a good idea to let the CEO facilitate the meeting himself. It would be a sign. I understand that this might not be possible in your case. Then get this mandate if you are the daredevil to facilitate this meeting. (It is fun, believe me!)
- Advanced level only, but important: Separate the horizons from each other – Horizon one options will kill Horizon 2 and 3 options when discussed in the same meeting! (Yes more complexity …). Read the infoq article for more on this.
Some basic checks
Here are some basic checks on smells I often observe in implementations that are not working or quickly degenerate:
- If your portfolio level has many columns, it reeks of process control rather than big strategic decisions. See the above example for suggested simplicity.
- If the board has no lanes (or rows) it seems that you did not map strategic topics to the board. They would make good lanes.
- If the lanes are called after teams, you seem to manage teams, not the portfolio respective the strategic topics.
- If your board carries explicit WIP limits, you may feel nice and a true fellow of Kanban, but managers will give you weird looks. It does not work this way. What is important is that the system is implicitly WIP constrained. Discuss if anything more is possible or not, but don’t stick numbers to the board. Portfolio initiatives are too divers in size and complexity, especially across different phases to stick to explicit WIP limits. Also, socially in these circles this looks and feels like process ideology rather than artful portfolio decision making.
I make all these points to make clear how PB is intended to be and how it will work. If you are not there yet, work on getting there. But don’t start the thing without having support for all the basic issues. You will need to have strong backup on these points. There will be hard days, hard decisions. If there is no trust in you on these issues, you and the board will not survive hard discussions . The board will be in the way of people under stress. Remember that most methods or tools get one shot. So better be prepared. You will only have one shot for starting this. If it degenerates too quickly, the method, not your local approach will be blamed. There will be slide back and no second chance. Don’t miss this opportunity and be prepared.
Once you have prepared your environment and have the basic ingredients in place, wonderful things will happen. The first few meetings might be a little awkward. But after a little while, you will realize the relief of a very focused decision meeting. What will happen is that all your VIPs will, by discussing the decisions at hand, implicitly align on one interpretation of the strategy and its deployment. And as they are important multipliers, your company will be able to I live one interpretation of a strategy, if the messages from the meeting will be sent across the company and will be operated through all your other boards on department and team levels. You will have a perfectly aligned system of boards on all levels, all triggered through the strategic and tactical a decisions on a portfolio level.
A little hint and observation: You can start your board without having an explicit vision and strategy. This will very soon be a point of discussion. Without vision strategic corner stones, most decisions will be based on opinions. These opinions will not be resolved without the string foundation of a strategy. The value of any initiative depends strongly on the context of the strategy and its common interpretation. A PB in a void will lead to people screaming for a strategy. Be prepared for the moment! (Like: call me then )
Another pragmatic hint: people will tell you and you will think that it will be hard to populate the board in the beginning. People and you will think there will be huge discussions about the prioritization of the initiatives on the first board. It will be easier than you think. People will behave when placed at one table. That’s one basic ingredient of the whole thing: having all people in one room having to decide will lead to people deciding. It’s different than taking to all parties in one on one situations. The hard part about the first board will be the realization that way too much is going on and that the first decisions will be about which initiatives to kill.
Please send me feedback where things are still unclear! I will integrate updates on this post based on it. Take this as a kid that still needs to grow up. Thanks!
Image all rights w/ Wikipedia