May 7, 2006

...Learn TDD with Codemanship

Nonlinear Management - Games & Strategies

In previous posts, I've suggested models for project planning that build in uncertainties - which, in real projects, are a result of complexity. These models are effectively dice games that illustrate the nature of Agile planning but simplify the whole process by replacing requirements with desired dice numbers, and replacing software development - or whatever it is that the project is doing - with throws of the dice.

The throws of the dice represent effort: cracks of the whip, stabs at a solution, or whatever you want to call them.

I've also introduced two other simplified models to help us understand the Agile development process. One is the evolutionary model, where the end product adapts to a continuously changing fitness landscape through what are essentially random mutations, and the other model is of a problem tree where the solution to any problem can present multiple sub-problems that will also require solutions.

When you combine these models, they create an essential, stripped-down model that helps us explore strategies for successful innovation in the face of fixed constraints and considerable uncertainty.

Scaling the model up is not so hard, since it is now just a simple dice game with simple rules that a child could follow. The underlying premise of Nonlinear Management is that the strategies that work in the simple model will also work in the real world, because the uncertainty introduced by the dice masks a whole universe if complexity behind a simple probability.

Managers responsible for governance of multiple projects - and I don't use that word lightly - therefore need not concern themselves with the details if they have a grip on the probabilities and on effective nonlinear strategies. At the highest level of governance, the game of chance is most pronounced, and the smartest managers use strategies more usually associated with the stock market or betting on the horses.

The best way to manage unmanageable complexity is to encapsulate it with manageable uncertainty, and leave the details to the teams themselves. Allowing teams to self-organise to achieve testable goals and solve clearly communicated problems, while ensuring the conditions that will maximise their chances of success, is one half of effective Nonlinear Management. The other half is knowing a lame horse when you see one, and knowing when to bet and when to cut your losses.

On August 10th in London, I'll be explaining these and other models, and exploring a variety of strategies for Nonlinear Management in two day workshops. You can find out more and reserve you places by visiting our events page.
Posted 15 years, 7 months ago on May 7, 2006