Tuesday 21 October 2014


Why knowledge transfer through discussion is 14 times more effective than writing


Knowledge can be transferred in two ways - by Connecting people so that they can discuss, and Collecting knowledge in written (explicit) form so others can find and read it (see blog posts on Connect and Collect). 

Connecting people is far less efficient than Collecting while being far more effective - but how much more effective?

We can never be sure about the effectiveness of knowledge transfer without some good empirical studies, but there are 2 pointers towards the relative effectiveness of these two methods. These pointers are as follows;

The often repeated (and sometimes challenged) quote that “We Learn . .
  • 10% of what we read 
  • 20% of what we hear 
  • 30% of what we see 
  • 50% of what we see and hear 
  • 70% of what we discuss 
  • 80% of what we experience 
  • 95% of what we teach others.”

David Snowden's principle that

  • We always know more than we can say, and 
  • We will always say more than we can write down

Let's make two assumptions here, firstly that the percentages in the first list are correct, and secondly that we equate the "more than" in Snowden's principle to "twice as much as" (OK, this is entirely arbitrary, but I want to see what the consequences are).

With these assumptions, the effectiveness of the Connect route is as follows
  • I know (100%)
  • I say (50%) 
  • You learn through discussion (50% x 70% = 35%)
The effectiveness of transmission of knowledge through Connecting is therefore 35%.

The effectiveness of the Collect route is as follows
  • I know (100%)
  • I write (50% x 50% = 25%)
  • You learn through reading (25% x 10% = 2.5%)
The effectiveness of transmission of knowledge through Connecting is therefore 2.5%.

Therefore transferring knowledge through Collecting is 14 times less effective than transferring knowledge through Connecting people.

If we change the proportions in Snowden's principle then we change this conclusion. If for example 
we always know 3 times more than we can say, and we will always say 3 times more than we can write down, Collecting becomes 21 times less effective, and so on.

I know all these figures are arbitrary and inexact, but what we are looking at here is some sort of estimate of relative efficiencies.

Note that this does not mean that Collecting knowledge has no place in Knowledge Management - quite the opposite. Despite being very ineffective, it is very efficient. Knowledge has only to be documented once, to be re-used one thousand times. Efficiency can trump effectiveness. However we can conclude the following
  • Because of these relative efficiencies, Knowledge should shared in explicit form (the Collect route) only when it is relatively simple and when it can be codified with minimum loss of context. 
  • Where knowledge is more complex or more contextual, it should be shared through discussion (the Connect route) - for example through conversational processes such as Peer Assist
  • Where efficiency is more important than effectiveness (i.e. broadcasting relatively straightforward knowledge to a large number of users), the Collect route is ideal.
  • The Collect route is also necessary when a Learner (a recipient for the knowledge) cannot be immediately identified, so no Connection is possible (see "speaking to the unknown user").
  • Even then, it is worth "keeping the names with the knowledge" so that readers who need to know more detail can call the originator of the knowledge and have a discussion. 

4 comments:

Unknown said...

I am learning Rubic's cube, and found even learning from video is also very difficult, sometimes more difficult than a well written document. While learning from others is much easier, because the teacher would identify your problem and offer very precise information to help to understand, and when I tried to teach my son to do that, I found letting him watching video is useless, or too difficult. The only way to teach him is to teach by myself, very slowly, very precisely.

10 times of difference is quite fair in the case of learning playing Rubic's cube.

Anonymous said...

Same counts for presenting. When you, as a public speaker, limit your interaction with the audience to standing in front, delivering your talk, and doing a Q&A at the end, you’re going to miss every opportunity to socialize with them. So, break up your monologue from time to time and join the people who have listened to you for coffee breaks and networking drinks.
Read also my blog post about "The importance of fika" http://b2bstorytelling.wordpress.com/2014/10/09/the-importance-of-breaking-for-coffee/

Valdis Krebs said...

Daft and Lengel discovered this back in the 1980s...
http://en.m.wikipedia.org/wiki/Media_richness_theory

Nick Milton said...

Thanks Valdis

Did Daft and Lengel put any quantitative figures on media richness, or did they keep it qualitative?

Blog Archive