Sunday, April 29, 2007

Stop all that chattering! I'm talking ...

I've read some disturbing posts over the last few weeks regarding freedom of speech, the consumer as a producer and the web 2.0 phenomenon.

I though I'd post something on these, and finally get round to replying to Jenny's questions.


As Yochai Benkler and Eric Von Hippel have studied the open source movement and emerged understanding of the "Wealth of Networks" and "Democratizing Innovation", am I understanding correctly lessons from your open source experience for creating sustainable networked organizations include:

  • i. "you cannot efficiently plan out the process of development as it is more akin to research and therefore dynamic".
  • ii. "three axis of technology, people and requirements being relatively unknown"
  • iii. "try, measure and adapt"


My experience comes not from creating open source communities but dealing with static or dynamic processes of building. I'm not sure how applicable these are to your work on networked organisations, but let me explain these processes and how the three points you mention relate to my experience.

ii. "three axis of technology, people and requirements being relatively unknown"

The process of building a software system can loosely be described as involving people, technology and a set of requirements.

If all three of these "axes" are well known or well defined, the process of building can be described as static. Whereas if these "axes" are not well known or defined, the process can be defined as dynamic.

Hence mass copying CD's, is a fairly static process - you know exactly what the requirements are, CD writing technology is well known etc. Conversely developing a new and novel web site can be described as a dynamic process - even the requirements are generally not well known.

iii. "try, measure and adapt"

Unfortunately in the software industry there seem to be two common re-occuring issues:

  • Firstly, despite much being CODB (cost of doing business), there seems to be an over tendency to develop or customise. Fortunately with the rise of SaaS (software as a service), utility computing markets etc - some of this tendency may diminish.
  • Secondly, where things were novel and new - and therefore technology, requirements and individual performance are relatively unknown, there has been an over tendency to attempt to use static planning processes. Concepts such as "software factories" and the scientific approach to management (e.g. Taylorism) have been misapplied in this context.

About a decade ago, when developers started to use more dynamic planning methods to deal with dynamic problems, there was a significant improvement in productivity for the companies they worked for. Today, techniques like SCRUM & XP (also known as agile development) are becoming widely used because they are inherently dynamic and are designed to deal with new and novel software development, unlike static planning systems.

These new methods are based upon the concept of "try, measure, adapt". In the case of test driven development, you write a test, you write some code to try and pass this test, you measure whether this worked, you adapt if it didn't and on and on.

i. "you cannot efficiently plan out the process of development as it is more akin to research and therefore dynamic".

Now "try, measure and adapt" is a valid form of control, but notice there is no specific planning step. This doesn't mean you don't have a plan, it just tends to be fairly minimal.

I'd like to make a joke that "you wouldn't try to Gantt chart a cure for cancer"; unfortunately in todays R&D environments in the UK, for some quixotic reason such static planning methods are being enforced. Lunacy ... no, just wasted energy and effort.

Novel software development is more like a game of football - you never play the same game twice. You have a common goal, an idea of how to attack the game, but fundamentally you try out something, see if it worked and adapt - during and between games.

Every football team dreams of playing a team whose players are following a rigid plan. Could you imagine Gantt charting a football game before the game? Could you imagine a team who followed such a plan? What happens if the ball isn't where you planned it to be?

This illustrates why static planning processes are good for static systems, whilst dynamic processes are good for dynamic systems.

Hence my points :-

  • i. "you cannot efficiently plan out the process of development as it is more akin to research and therefore dynamic".
  • ii. "three axis of technology, people and requirements being relatively unknown"
  • iii. "try, measure and adapt"

Now let me try and link these ideas to the concepts of agile enterprises.

First onto Enterprise 2.0 technology and specifically wiki's. At Fotango, we adopted a wiki some four years ago as the static process around our intranet (this person writes this bit, this person approves etc) had created a information resource which was useless. So we decided to try something new. We put up a blank wiki and before long everyone was contributing something.

Unfortunately, so much information was put onto the wiki that it became overloaded with "noise". So we needed to adapt and try something else - "gardening". By "gardening" I mean a regular pruning of information within the wiki.

In some organisations "gardening" may emerge naturally, in ours it didn't. This is a critical point: you shouldn't plan out in detail the adoption of an Enterprise 2.0 technologies within an organisation because you don't know what behaviours will emerge. Instead you need to "try, measure and adapt".

Note, I say you shouldn't plan out. This doesn't mean you can't. I can always plan out exactly how a football game is going to go, who is going to be where and at what moment in time etc. I'm not going to get the best result if I do - especially if the other team don't follow my plan.

So on to my title ... stop all that chattering! I'm talking ....

Whenever I've been involved in introducing more dynamic processes, I've generally come up against a fairly resistant and incumbent "old guard" who like the "old way".

So we come on to the latest spats about amateur online journalism. The "old guard" of the news world has been very comfortable with the well defined macro level processes of them collecting information, editing and disseminating it. Sometimes, they have been caught out spinning or doctoring information - reinforcing the old adage of "don't always believe what you read in the papers".

The "old guard" also selected who had a voice, it decided upon the criteria of expertise, it chose.

Unfortunately for them the ball has moved, and now we are in a world where anyone can collect, edit and disseminate information.

This "new world" does create a lot more noise. It provides powerful new mechanisms for anyone to be heard. Much of what is behind the attack on "net neutrality" in the US, the involvement of more traditional news organisation into the internet space and the recent spate of articles about the need for curators or editors for the internet appears to be about re-establishing that neat, old view of the world.

What is needed however, are new forms of control that are more adaptable to the reality we find ourselves in. I do want to know what is happening in the world. I do want to trust the source and sometimes I do want to have my say.

However, in this "new world", I get to decide whom I listen to. The only issue is who do I choose?

Unfortunately whilst the mechanisms of dissemination exist, the mechanisms of choice or trust don't. What is needed are reputation-based social networks. A method for searching for information from people that I, my friends, or the general public, trusts.

It may emerge that we choose to trust the same "old guard" as before. If not they'll just have to adapt and try something new. It may emerge that a "new guard" is created through the Stentorocracy as I called it.

You cannot understand everything on the Internet, you cannot make perfect sense of all the noise. In much the same way in economics you cannot make perfect decisions, or be that " rational man" or reach your "pareto-optimality". There is always too much information.

Something needs to separate the noise from that which is useful. Hence a new system, reputation-based social network, is needed to separate the wheat from the chaff.

Of course, there will always be winners and losers. There will be the included and the excluded. Maybe we'll all end up listening to some random 15 yr old blogger because he made some good points.

Maybe not.

However, I don't think an approach of "Stop all that chattering! I'm talking ..." is going to get the "old guard" very far. Especially if they use the same techniques, such as blogs, which they complain about. Hence I'm far from convinced by Andrew Keen

Still he has a voice, he has a right to be heard. But then so does that 15 yr old blogger.

You see, in my simple world it's the idea that's important, not the source.