Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Interviews David Anderson on Adopting Kanban, Modern Management Framework and Enterprise Services Planning

David Anderson on Adopting Kanban, Modern Management Framework and Enterprise Services Planning


2. Great to have you here. The first question that I would like to ask you is that you said that media, internet and SaaS firms are among the first firms to adapt Kanban. Can you explain why?

This five-year old information, 2009, the first companies we really saw adopt are companies like Yahoo, media firms like the BBC, the Lonely Planet, Globo TV in South America, NBC Universal and SaaS firms like Constant Contact, Ultimate Software. Why? Because the world they live in doesn't move in nice little time boxes, doesn't move in one month boxes of time of two weeks or one week and real world events unfold whenever they do. They need to be responsive and they need to be able to adjust to that. So the decoupling that you get with Kanban of decoupling the planning activity and doing that in some regular cadence from however long it takes to do the work to the delivery cycle, that flexibility allows them to adapt to that.

So Globo TV, for example, Rio De Janeiro, they used a Kanban system to manage post-production on daytime soap operas. So they take all the raw video footage they've made with the actors and then they modify. They may be editing out mistakes, maybe a boom microphone that's come into the shot and they digitally erased that or they changed the weather conditions if it's an outside shot and perhaps they made the film on different days, then they digitally changed the weather to make it all look consistent when they edit it together. And from the time when they shoot the video to when it goes live, they have about 11 days of lead time to make any changes and edit the footage together.

So meanwhile, information is arriving in the market, information about how the public are responding to the storylines and the soap opera and the characters. Maybe a character is unpopular, so they have to kill the character off so they rewrite the storyline and they shoot some new footage very late in the 11-day cycle. They change the plot. They add in new stuff. So the world they live in as informational rival is very frequent. So they do the planning twice a day. And then video editing on a section of video may take a couple of days but their release cycle is every day because the soap opera goes live in the afternoons every day. So they've got planning every other day, lead times may be two days, three days, four days, and release cycles of every day. It's not nice little two weeks we planned it, we did two weeks of work, we release it. Kanban provides that flexibility and media companies particularly desperately needed that and that's I think the reason we got that adoption.


3. You also mentioned today that Kanban results turned out to be repeatable. That's something you discovered along the journey of implementing Kanban in organizations. Can you give some examples of that?

I think I mentioned that in a couple of weeks. The first ever Kanban implementation with Microsoft was an internal IT application maintenance group and it was really the mechanics of a Kanban system, the pull system mechanics that we implemented. That gave them some flexibility of responsiveness but other things like predictability of delivery, shorter lead times because of managing the queues of work, and the field commitment. We're not committing on something until we're absolutely sure we want it and we want it soon.

And some other people were able to take the guidance I had written on that, the stories I had told and the way that I described that and understand it and just copy it. A guy called Eric Landes did that with internet website development with Bosch, for example. So he just took the recipe I described and he copied it. And in that sense it was repeatable. Other people could use what I had written and repeat the results and produce I think in similar outcomes, perhaps 200%, 300% improvement in the delivery rate of the work or a shortened cycle times for doing these upgrades and new features on the internet.

And then perhaps a few years later, again maybe 2009, the bigger thing that we now call the Kanban method, the use of Kanban systems, visual boards, a number of feedback loop meetings like operations reviews, other people started to copy all of that wider guidance and produced similar results that we'd seen at Corbis in 2007, cultural changes in the company, changes in the behavior of the wider business and the interaction of different business units and changes in the behavior in the executive management to the point where IPC Media which is one of these first companies that replicated those cultural and outcome business benefits. The business director was actually the one who was on the stage with the Lean Conference in London in 2009 telling the Kanban story.

So that had shown that this much bigger thing that I've done at Corbis, other people could understand it and they could learn from it and copy the results. And the key there was that I didn't need to be present to make it happen and that's what we mean by repeatable.


4. You also talked about the work that Chris Matts and Olav Maassen have been doing in the area of Real Options, and they also published a book on commitment recently describing the way that they are working with this. How does Real Options fit into Kanban?

I've known Chris and Olav for quite a long time at least since 2005. And Chris has been pushing this idea of Real Option theory for at least that long. Kanban is a very natural fit for that style of thinking, the idea that you defer commitment as late as possible and the Kanban systems make that commitment very explicit. The replenishment action of pulling something into the Kanban system to flow it through is that act of commitment. And we've also learned from experimental work really that was done in France and Belgium four or five years ago that you can take the principles of that downstream service delivery Kanban and you can apply it upstream of the commitment point. Therefore, you can use the Kanban system for managing the development of options and discarding them at different stages of information discovery, and therefore an upstream Kanban allows you to manage embedded options. So that's a really natural fit that Kanban is a great way of implementing Real Option theory in action.


5. You mentioned also the modern management framework today as a way of collecting practices in the way of how organizations can be managed which is familiar to Kanban where you've been doing this. So why did you start this initiative? What does it need for a new way of management?

So we’ve been teaching Kanban, educating people on it, holding conferences, helping people understand it, implement it in their organizations for seven years now. And the thing about you implement a Kanban system so it has this very explicit commitment point and you might be saying we have two slots free, two Kanbans available today. Pick the next two things you want me to work on for delivery at some time in the future. Imagine that's Globo TV and you're a video editor and you have the capacity to start working on some news of editing. We’re going to ask their producers and the director, what do you want done next? And they have all set of options available to choose from.

Well, the mechanics of the Kanban system are not everything everyone needs to know. They need to understand things like how do we prioritize things, how do we make choices between different options, what are the risks involved, when is the right time to start something. And if we can't start at the right time, do we need to start early in order to manage the risk associated with it or is it better economically to delay it until later, and when will it be done by and if it's a batch of things, how long will that batch take and what will it cost? All of that stuff people need advice and guidance on.

And then they might start asking the questions like do we have the right options? Are we generating the right things? Are our customers happy? Are we delivering in a way that they want us to? Are they happy with the service delivery? Are we fast enough that we have good enough quality? And we've developed teaching material for all of that stuff. So it's a lot more than just Kanban systems or card walls, Kanban walls. But the world of Kanban can be interpreted in different ways.

In Japan, if you wrote it in hiragana, it means Kanban system like we started with Toyota. But if you write it in kanji, a regular Japanese person will interpret that to mean something like a billboard. But if you write it in the same Chinese characters and show it to a Chinese person, they interpret that as a verb and it means a group of people standing around looking at a board. So Kanban in Chinese is like the cartoon on the cover of my book. And for non-Chinese or Japanese speakers who are reading it in English, they are thinking, well, what does that mean? Is it a Kanban system? Is it a Kanban board? This bigger thing David wrote about in his book. There's ambiguity about it.

Now imagine you are a senior executive and you want to roll all of this out in your company. If you send someone on Kanban training, does that mean they're learning to put sticky notes on a whiteboard? We needed a container for this wider body of knowledge and we've chosen to call that container the modern management framework. Why modern and why framework would be the next questions. Well, modern has a specific meaning in the field of the arts. Modern implies that the form of art changed because there was a technology change.

So real world examples, historically, the arrival of photography changed painting. Expressionism evolved because representative art was no longer required. Painters were not required to paint things representationally. Photographers displaced them. So expressions start to emerge. Electronic music in the 1960, '70s, changed the way the music is written, the way it's played, what you can do with it. Digital sampling in the early 1980s completely changed music. So a discontinuous innovation technologically causes a change in the form. That is the concept of modern. So 1980s electronic music is modern music. But by 2014, if you are still making music that same way, it's not modern anymore, right? Modern has this period.

So what I believe we're doing is modern because there's innovation in there. We’re using sticky notes in a way that they were not intended for. We're using whiteboards in a novel way. We're using software tracking. We’re using new scientific theories and models of understanding how humans interact and how the complexity of the world is and applying that to the field of managing business. So modern management.

And why framework? Well, framework implies that something is incomplete and you saw me present a whole series of things evolved over a 7 to 10-year period of time and we don't think we're finished. We expect more innovation next year and the year after and the year after. So we expect that framework to keep getting bigger, hence the name modern management framework.


6. You also talked about the fitness for purpose today. Why did you start to implement this? What is the need for this?

So the concept of fit for purpose really came out of an evolution of talking about Kanban as a way of driving continuous improvement in organizations and that you had your existing process and you Kanbanized it and that caused it to evolve, created some stress on it and it made it change. The people involved made it change. And in evolutionary biology, you might ask the question, if you mutate a species to a different version, how do you determine whether the original species or the mutation survives and thrives? And while it's more complicated and I am going to explain right now, at a very simplistic level it's whether the mutation is fitter for its environment than its predecessor.

So if we have a business process, maybe it's video editing soap operas and you might be in a world where it takes you two weeks to do that editing and if the scripts are not changing rapidly, two weeks is long enough. But what if something happens in the external environment? Perhaps the owner of the company changes and the owner likes to poke his nose into the script and the characters because his wife watches the show. And all of a sudden you get all this late breaking information. The company's owner wants this character killed off and you don't have two weeks to respond anymore. So you have to reinvent the way that you do the editing process. And if you've installed an evolutionary capability in your organization, you can respond to those changing external conditions, and the reinvention of process is much more graceful than it would be if you were sort of dumb to the outside environment and what the outside world needs.

So fitness for purpose is about asking what business are you in and you are delivering some kind of service whether it's software or IT operations, other IT services, at least I hope for an InfoQ audience, but it could be something non-IT related. If you deliver pizza, what makes pizza delivery fit for purpose? How long did it take to deliver? What was the quality of the pizza when it was delivered? What was the predictability of the delivery time? And did the restaurant conform to regulatory standards for food hygiene in the preparation? On the other hand, did the pizza delivery boy break any traffic laws while he delivered your pizza? Maybe you don't care.

So we're teaching people to look at the outside environment and use what the customer cares about as the metrics that drive the process improvement on the inside. And we no longer talk about continuous improvement; we talk about making something fit for purpose.


7. One last question. You talked today about enterprise services planning, something new that you are working on right now. Can you elaborate what it is?

So there's a lot of material in the modern management framework for which there isn't any easy tooling. We do a lot of forecasting. Kanban is famous for no estimates and to some extent for no prioritization and no planning. We must have some solution to that and of course we do. We teach that stuff in class. And it's not terribly complex but it's complex enough that you might want a spreadsheet that could do it for you or some sort of calculator program to do it. And often people are asking that while we might teach them, how do you forecast when will the product be finished and how much did it cost and how many people did you need and what should the wet limit be on the Kanban board, all that stuff.

We can teach that in class but right now there's no software that can tell you that stuff. And if you've got this large scale implementation, imagine an ecosystem of interdependent services in the company. Perhaps you are a big system engineering company, telco equipment manufacturer, internet company or you make electronic devices like mobile phones. There's a lot of moving parts and different business units all over the world now interconnected.

Well, how do you know how many people you need in each place? How much capacity do you allocate to anyone project? At what time? For how long? Which classes of service are you offering for pulling the work through the different systems? What is an acceptable lead time? How predictable does it need to be? What should the quality level be? That becomes quite a complex problem even at two or three interdependent systems. You need software to solve that problem.

So in order to be able to answer these hard questions that big bosses have wanted for so long, we believe that we understand how to do that now but it's not a manual process. It needs software. And I've chosen to call that software Enterprise Services Planning. And there's a natural mapping to the evolution of Kanban in manufacturing. Kanban starts at Toyota in 1947. They didn't call it Kanban. In fact, during the 1950s it was called something like Toyota's Manufacturing Planning. They only called it Kanban in 1964 when they were applying for the Deming Prize and the auditors asked them, "So what is the system called?" And they had to scratch their head and think, "What should we call it?" So the same year in 1964, MRP, Manufacturing Requirements Planning, was launched and that was in response to Toyota obviously doing something clever. We should find a way to teach all other manufacturers to do it.

So MRP of course later in the market, MRP is basically a good thing and it's useful and the litmus test for that is ask a manufacturer, would you take your MRP system away? And they will all say no. And software companies that got into MRP of course wanted to expand their market, and now we’ve got this Enterprise Resource Planning which has a terribly bad reputation but these companies make a lot of money.

Actually, I think ERP solved their own problem and what we need to worry about with modern knowledge work of businesses, creative services, businesses where people think for a living, create stuff out of their heads, we need to think of that as a service delivery business and think what is the customer looking for and hence ESP. So Kanban and the technology sector 11 years later hopefully becomes ESP in the way that Kanban and manufacturing became MRP 17 years after its original introduction.

So I think that's the future for Kanban and the future for the tools market. The smarter folks listening to this should know a little bit about marketing or realize what I am doing is I am resegmenting the market. We have this Kanban tools market and really it’s a visual management market and some of those products are free and many of you will be users of those free visual Kanban tools. And really that's fine, we can have a visual market and maybe it’s a very low cost, low price market.

And I want a separate category of thing called Enterprise Services Planning and it clearly communicates -- this is tooling to run your business and it's worth a lot of money. Therefore, it’s reasonable that it should cost a reasonable price. And I actually think that's going to help the manufacturers of these software products greatly, and I expect that we'll be collaborating with the companies that choose to be in that ESP market. We'll be collaborating with them over the next year.

Ben: Thank you, David, for the interview.

My pleasure. Thank you.

Dec 24, 2014