Everything is fragmented-
The safe-fail approach to KM

I gave a keynote address in May in Perth, Australia, to an audience of technologists looking at the future development of information and communications technologies (ICT), as well as issues of compliance arising from the demands of Sarbanes-Oxley and other regulatory requirements. I started off with one of my favorite quotes, from Cardinal Bellarmine in a letter written to Foscarini, a Carmelite collaborator with Galileo in 1615:

"For to say that, assuming the earth moves and the sun stands still, all the appearances are saved better than with eccentrics and epicycles, is to speak well; there is no danger in this, and it is sufficient for mathematicians. But to want to affirm that the sun really is fixed in the center of the heavens and only revolves around itself (i.e., turns upon its axis) without traveling from east to west, and that the earth is situated in the third sphere and revolves with great speed around the sun, is a very dangerous thing."

The good Cardinal was a very intelligent man, but he was locked into the patterns of past orthodoxy. He can see that a heliocentric view of the universe produces more accurate prediction of the movement of the planets, but he insists that there is no contradiction between this statement and a belief in terracentricity in practice. I argued that every organization in the room had its own Cardinal Bellarmine, and many of them are present in the ICT function in organizations. The governance of ICT is locked into the capabilities of computers in the 1980s, not the 21st century; it focuses on control and linear processes in which the business analysts discover users’ specific needs, which are then converted into projects, either procurement or development. In a mischievous moment, I suggested that user requirements documents were designed by the IT industry so that users would sign up for things that they only partially understood, but for which they could be held accountable later.

Now I accept that approach is valid for major systems—such as the accounting aspects of an enterprise resource planning (ERP) system—where requirements are known and little variation is possible. However, for knowledge management, the requirements are more abstract and uncertain, and in practice users don’t know what they need until they get it, and then they want something different. The capabilities of technology can change requirements as users’ knowledge of potential co-evolves with the delivery of capability. I have experienced the Wikipedia. I know how to use it. But I could not have specified the need for that sort of capability in advance; neither could I have ascertained the need through interviews or by design alone.

In the knowledge economy, applications co-evolve as software components, hardware and communications technology interact with users to create emergent applications and capability. You can’t manage that by the linear process of requirements sign-off. Instead, we need to allow multiple (and sometimes contradictory) software tools to interact with people over time and see what works. Where things work, we amplify them; where they fail, we dampen their impact. We move from a fail-safe design strategy to one of safe-fail experimentation; we do not assume that we can know in advance what the right solution is, but we do understand a process by which that solution can be discovered through action rather than reflection. The problem is that procurement and design methods are based on fail-safe, not safe-fail, as a result of which they are expensive and non-adaptive.

The good news is that the capabilities of the Web 2.0 environment are made for a safe-fail experimental approach to knowledge management system design. More than 10 years ago, when knowledge management started, we had a great idea supported by poor-quality, process-based technology. Ten years later, the idea is still good, if a bit battered from abuse over the years, but we finally have the technology to make it happen. In a modern environment, there is little need for major expenditure in the early days of a knowledge management program; indeed, such an approach is contra-indicated. Anything you can legitimately do on a sustainable basis with communities of practice can be done with freeware such as Ning. Blog software allows rapid take-up and creation of knowledge objects. Search is free, and RSS feeds can be changed frequently as new capability becomes available. It takes minutes to set up a wiki.

I spend my time in a mobile environment with Facebook, Twitter and Skype permanently open, linking and connecting with people in unexpected and unanticipated ways. I can do that as I work for a small company that exists in the Web, rather than within the firewall. We have no Bellarmine imposing the excessive order and structure of corporate procurement. You cannot impose constraints on what needs to be an emergent system.

In my next column, I will discuss a low-cost use of social computing to create a KM program. The podcast of my talk in Perth is available at http://www.cognitive-edge.com/podcastdetails.php?podid=31.