modeling diffusion patterns


I had one of my first inklings that quantitative sociology was for me after reading David Strang’s and Sarah Soule’s (1999) Annual Review piece on diffusion. The idea that you could track or describe the spread of an innovation with a statistical approach struck me as very cool.

My interest in diffusion patterns has waned a little over time. Now I’m much more interested in examining variation in influence and its consequences. However, this article grabbed my attention in the way that many innovative marketing papers do. As described in this press release, Wharton scholars Christopher Van den Bulte and Yogesh Joshi attempt to model how diffusion patterns differ depending on the extent to which the product is mainstream or niche-oriented. Mainstream products (like movies with big-name stars headlining) exhibit a familiar sharp increase in users and then a steady decline thereafter. “Sleepers,” on the other hand, tend to have a typical bell growth curve – slow-takeoff that builds towards a gradual increase followed by a peak and a gradual decline.

Van den Bulte and Joshi offer a third alternative – the two-segment model. In this model, a population is divided into segments of influentials and imitators (following Riesman’s notion of inner-directed and other-directed individuals). The influentials discover new products and set the standard for their quality, which is then adopted by imitators. In this case, the diffusion pattern exhibits a dip (as influentials determine the standard) followed by a gradual increase and then a steady decline. The model is useful, according to the authors, because it allows managers to forecast growth in certain kinds of products. Indie-films, for example, should be targetted to a specific aficionado niche, where it will either be accepted or rejected, before attempting to market the film more widely to your typical indie scene.

The paper consists mostly of formal modeling that might be applied in various ways with the sophisticated heterogeneous diffusion model techniques developed by David Strang and Nancy Tuma.

Written by brayden king

September 7, 2006 at 2:45 pm

%d bloggers like this: