Tuesday, March 29, 2016

On speaking at Mendix World

I've been invited to speak at Mendix World later this year and their press people have asked me some questions for an article. In fact the questions came from Sander van der Meijs, the chief editor of CIO.nl, and they are so interesting that I'm going to respond here as well.

Q 1. You say companies often do not have a strategy. The only strategy they have is copying other companies that seem to be successful. What does this mean in the context of software development? How can organizations break free from this ‘copy cat’ mentality?

On the contrary, companies often have oodles of strategy but it is based upon poor situational awareness (i.e. little understanding of the environment). Without situational awareness then you are forced to rely on verbal reasoning and meme copying others including the use of backward causality without context. In other words, the general mantra becomes "if A does B and A is successful then if we do B then we too will be successful!"

Dilbert has a fabulous cartoon for this.

As Professor Martin says, managers are busy people and there's an army of consultants willing to provide these "solutions" of copying others. In management theory we even write books encouraging others to simply copy the greats of Kodak or Atari. If you want to break free of simply copying others then you have to get to grips with the environment and improve your situational awareness.

Q2. Agile development is often used because of the hype. But in your top predictions for 2016, you suggest that one size does not fit all. What are views on where agile is best applied and how to manage multiple methodologies within one organization?

Agile development was something I used extensively back in the early 2000s. However, it soon became clear that it was not suitable in every circumstance. This echoed the Salaman and Storey Innovation paradox of 2002 that in order be efficient you needed to use one set of principles and in order to innovate you had to replace such erstwhile practices. Though we experimented with dual structures within the organisation, this quickly degenerated and we discovered that we needed three basic sets of methods and techniques to manage the process of evolution. This echoed Robert X. Cringley's description of three organisational types in Accidental Empire, 1993.

In order to do this in practice we needed to understand the environment and how evolved the components of any system are. This is something mapping enabled us to do by 2005 in a more formal way.

So, Agile (as in a focus on the core agile principles) is a perfectly reasonable technique for exploring the uncharted space of the novel and new because you need to reduce the cost of change as change is the norm. By the time the activity becomes industrialised then you need to focus on reducing deviation and hence a different set of approaches is needed e.g. six sigma. What people often fail to realise is there exists a huge chasm between these two extremes and this needs to be managed. In this middle or transitional stage then you need to focus on learning and reducing waste around the activity. Back in 2005, we were struggling to find consistent mechanisms for this but since then Lean (which uses a more formalised Agile approach e.g. Scrum with artefacts such as MVP - minimal viable product) has emerged as the most suitable.

Q3. Another one of your predictions was focusing on user needs, which aligns nicely with your Value Chain Mapping framework. Describe this framework and how it helps organizations align their business priorities with user/customer needs.

In order to map an environment (LEF members / general public info) you need two key aspects - position and movement e.g. position of the pieces on a chess board and where they can move to. In order to determine position you need an anchor. In the case of a chessboard this is the board itself. In the case of a geographic map it's the compass. In the case of business, the original (and so far yet to be surpassed) anchor that I used back in 2005 was the user and their needs. 

From the user need you can then develop a value chain which gives the position of components (in terms of visibility) relative to the user i.e. if you're making a cup of tea then the electricity provider that you're using to provide power for the kettle is not a visible user need but nevertheless it's still an important component far down the value chain. The power is needed for the kettle, the kettle is needed for hot water, the hot water is needed for the tea and tea is what the user needs.

However, position is not enough. You also need to understand how things can move and in business we're talking about how things can evolve. If you don't understand how evolved the component is then someone might start building a brand new form of kettle rather than buying one off the shelf. Back in 2005, I guessed at a useful axis based upon a pattern I noticed in 2004. It wasn't until 2007 that I had collected the 9,000 odd data points to justify the process of evolution (i.e. genesis to custom built to product to commodity) as being moderately sound.

Once you have a map (and it's a relatively simple task which takes someone with reasonable skills a few hours) then it become trivial to identify duplication, bias, poor contract structure, misapplication of methods and poor structural design with a company. The maps also help with collaboration, communication, linking operations to strategy and getting everyone focused on the user need and where their part of the map fits into providing the overall user need.

Over time and practice you can use the map to learn about common economic patterns, context specific forms of gameplay, universal doctrine and anticipate change in much the same you learn to play a game of chess. For example, there's at least two forms of basic disruption - one of which can be anticipated and one of which can't. However the failure to understand the environment means that companies are constantly being disrupted by anticipatable change that can be easily defended against if only they could see it.

Of course, once learned then lessons can re-applied to other contexts as appropriate i.e. in the military world we learned to use flanking and pincer movement across multiple battles. This is all part of the process of determining strategy and it's here you should determine your priority order. There is no general priority list, it's specific to the context and the game at hand.

Q4. You also developed the model with pioneers, settlers and town planners. Where did you get the idea, how did you develop it and how does it work? Are there situations where more or fewer operating modes are required or more suitable?

The premise of using a three party structure - e.g. pioneer, settler and town planner (nb, the LEF post is from 2011 and uses some older terms which I've since tidied up) - was developed by James Duncan and myself back in 2005 in order to solve the problem caused by a dual structure and its origin is Cringley's Accidental Empires, 1993. I've talked on the topic at numerous public events and in publications since 2007. 

Once you map an environment, the first thing to do is to apply "common" doctrine. One of the things about mapping is that it teaches you that some techniques are context specific whilst some are fairly universal. Doctrine are those techniques which are not context specific but universal and can be applied & copied to almost all circumstances. Unfortunately there are vastly more context specific techniques than those which aren't and hence the problem with simply copying others. For example, you might look at surge pricing in Uber and see that as a technique that is helping Uber become successful. That doesn't mean that you can simply copy and apply the technique to funeral parlours and expect the same results.

One of these universal doctrines is to break down the problem into small teams (today you'd think of Amazon two pizza model or Haier's cell based structure). The FIST principles of Lt Col Dan Ward, USAF are also highly applicable. However, once broken down you should note that each team not only has aptitude (e.g. finance, engineering, marketing and security skills for example) but also attitude (i.e. engineering in the uncharted world is not the same as engineering in the industrialised). Hence once you have a cell based structure, you need to add in attitude which is the purpose of pioneers, settlers and town planners. It's a recognition that we have at least three different attitudes and people tend to excel at one of these. 

Some people do well in the high speed and chaotic world of brilliant flashes and failure and can discover the new spaces of future industry. Whilst others are good at turning the half functioning brilliant flash into something people might actually want to use. This is no mean feat and these are the people that make success happen. There are also others who can turn the more matured thing into a highly industrialised form and build the empires of scale. All three types of people are brilliant in different ways. You need all three.

In order to function and to overcome any tendency towards inertia, you also need to introduce a concept of theft to mimic evolution within the structure i.e. settlers steal from the pioneers and turn the brilliant flash into something useful hence forcing pioneers to move on and do more "pioneering". Town planner steal from settlers and build the component services that pioneers use. It creates a virtuous circle which is what we noticed back in 2005 when it was first used.

There are ways to cut out some of the stages e.g. through the use of ecosystem but this requires very high levels of situational awareness and skill which I rarely see outside of areas where competition is at its fiercest and most critical. Most commercial organisations can get by with simply using cell based structures and keeping in mind the different attitudes.

Q5. What do companies need in addition to the model to be successful?

A map. The issues of culture, organisational structure, strategic gameplay and learning are all quite complex and without a map (i.e. position and movement) they became almost impossible to decipher hence our tendency to simply copy others (unaware of what is context specific and what is universal), use magic sequences of success (i.e. top ten things the successful do) and focus on the importance of "Why" whilst ignoring the fact that "Why" is a relative term as in why here over there. 

Q6. Where do you see the model in practice right now? What are some of the potential obstacles organizations should be mindful of when implementing it? And how can they overcome them?

Mapping is used in numerous organisations. There are also organisations using a pioneer, settler and town planner like structure but these are rare and for the vast majority of companies then the rigours of competition are such that a cell based structure is more than ample.

Q7. You’re coming to The Netherlands to speak at Mendix World. Do you see the model adopted in Europe as well, maybe even in The Netherlands?

The mapping work is creative commons share alike and is used by numerous organisations, in fact I wouldn't even know how many because of its open nature. My concern however is not this but that members of the LEF continue to improve situational awareness & gameplay whilst not getting dragged into following paths which are known to cause problems. I have no particular strong interest in whether other companies adopt the mechanisms of mapping or not. In fact, I know some members who would actively discourage others from using it!

On the question of why I'm speaking at Mendix world. I'm coming along to listen, to catch up with good friends and in the hope of serendipity. You never know what you might discover by chance.