Understanding Diffusion

John W Lewis's picture
Chat Date: 
Thu, Jul 13, 2017

Today I met a leading researcher in a field of medical research, whose field is the development of new research methods (i.e. a meta- or possibly meta-meta- field!) These methods are often very different from methods in current use. So he is particularly interested in the adoption of these new methods by researchers in his field.

During a presentation, he remarked that adoption of a particular method had initially been fairly steady, but had then flattened off, and that this was contrary to his expectations of exponential growth.

This seems, to me at least, to be similar to the assumptions of "hockey stick" growth often attributed to venture capitalists funding innovative developments in any field.

This prediction seems to be based on the assumption that the population of adopters is homogeneous. If the innovation is equally attractive to everyone, then the uptake is expected to depend only on the number of people who have heard about it, subject to going through a process of trialling and accepting or rejecting it. Based on this model one would broadly expect an initially exponential growth rate as more people inform yet more people, followed by an inflection and a levelling off as adoption by the finite population saturates: an 'S' curve.

However, for decades, we have known that adoption is more complex than this. The best known work in this area was done by Everett Rogers and published in his book "The Diffusion of Innovations" in the 1960s. This led to numerous other people building on that work, including Geoffey Moore's "Crossing the Chasm". As regards the adoption of individual innovations by populations, Jackie Fenn and Mark Raskino's "Mastering the Hypecycle" provides a model which includes the "Trough of Disillusionment" which has some equivalence with the "chasm", and has long been used by Gartner to publish information on the stage of adoption of multiple innovations.

Accordingly to these types of models, populations are not homogeneous. People belong to different categories of adopter. And their category is not universal, but they belong to different categories for different types of innovation. Also knowledge about innovations is not static, adoption by people in different categories yields different kinds of information about the innovation.

An understanding of these models contributes to a doctrine which fundamentally alters our expectations of the diffusion of an innovation through a population. Not only does this understanding include the different criteria for adoption applied by people in different categories. It also involves understanding the different types of communication which occurs between people in these categories.

The better we understand the fundamental behaviours in a field, the more likely it is that we can develop an overall approach which is likely to be effective.


Let's discuss this during #innochat on Twitter on Thursday July 13th, starting at 12noon Eastern time. Let's discuss the benefits of understanding more about about how innovation actually happens, as the basis for forming expectations of likely behaviour in terms of adoption (by individuals or organizations) and diffusion (by populations).


All donations go to help defray hosting, maintenance and domain name costs. Thanks!