KM and Monte Carlo simulations

The attached paper, "Using Monte Carlo simulations to predict outcomes of KM interventions" is a brief summary of the results of some exploratory research I have been doing in this area.

Monte Carlo simulations are a very useful way to explore non-deterministic systems using a combination of deterministic calculations and random numbers. In particular, the outcomes of my experimentation suggest that many across-the-board KM interventions are are risk of being dropped due to lack of results.

These untargeted interventions inevitably try to achieve smaller improvements across the board, typically in situations where there is a low probability of failures occurring. The Monte Carlo simulations demonstrate that for a substantial number of organisations, even successful attempts to reduce knowledge failure risk by, say, 10% may have a negligible effect on bottom-line savings over a sustained period of time.

KM and Monte Carlo simulations.pdf130.17 KB
Did you know...

Our expertise in complex systems analysis, combined with a deep understanding of technology and modern, agile management and leadership techniques makes knowquestion uniquely positioned to find strategic solutions to your tough problems. Contact us today.


Joe Firestone (not verified) — Wed, 29/04/2009 - 13:54

Hi Stephen,

Congratulations on a great post and paper. I've blogged about it here:



Mike P. (not verified) — Wed, 29/04/2009 - 23:00

There's another interesting KM simulation that looks at how well different knowledge strategies enable an organization to adapt (or not) the won the 'Best Paper' in the KM Track award at ICIS. The authors put it online for free download here:

Patrick Murphy (not verified) — Thu, 30/04/2009 - 00:46

As it happens, the more successful a KM program is, the more difficult it is to measure the value of KM discretely. See

Mark Gould (not verified) — Fri, 01/05/2009 - 21:25

I am not sure if I have understood the paper fully, with regard to compound knowledge failure. Am I right in interpreting the condition "risks of compound knowledge failure are negligible or non-existent" to mean that interventions are less likely to be successful when the knowledge failure is compounded by other factors or circumstances? That is, KM interventions should target failures where there is little or no compound effect?


Stephen Bounds — Sat, 02/05/2009 - 22:22

Mike, Patrick: Thanks to your references to other papers and presentations, these look very interesting.

Patrick: I agree that it's difficult to measure KM's value as an independent set of activities. Joe Firestone's view (which I agree with) is that KM should target knowledge processing activities which almost by definition are one step removed from operational behaviors. Thus, KM interventions will always only have an indirect impact on a business's success.

This, by the way, is one of the benefits of using the cost of knowledge failures as a benchmark. The assumption is that knowledge failures will always happen; the issue is what can be done to reduce their occurrence and/or if the costs of recovering from these failures can be reduced.

Stephen Bounds — Sat, 02/05/2009 - 22:43

Mark: No, I'm not trying to say that areas where compounding knowledge failures are occurring shouldn't be targeted. These areas may well be very fruitful areas to target.

What I am cautioning against is the implementation of a general "knowledge sharing" program which will generally only have a marginal reduction in the risk of knowledge failure. Even a reduction in risk of 20% or 25% is not sufficient to produce a high probability of substantial cost savings for the business.

This makes such a KM program a low-value proposition for any business's executive managemet. Let's say that the proposal is a Sharepoint rollout, a classic general KM play:

15% of all IT projects fail (Standish 2004 CHAOS report)
50% of all KM projects fail (Akhavan, Journal of KM practice, May 2005)
~ 30% of the time, a Sharepoint implementation that is not targeting specific improvements and thus effectively leaving results to random chance will show negligible benefit (my paper)

Now admittedly, many of these factors would overlap so we're not talking a 95% failure proposition. But nonetheless, looking at all these risks, you could hardly blame management for looking for another way to spend their money.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <small> <blockquote> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <br>
  • Lines and paragraphs break automatically.

More information about formatting options