There are some new services which can be very clearly good — a treatment for a virus, for instance, or another sort of medical innovation. But different improvements have worth that’s extra unsure, resembling an unproven know-how. In her newest analysis paper, Wharton administration professor Valentina Assenova examines the position of social networks, each on-line and offline, in the unfold of these advanced improvements. Her paper is titled, “Modeling the Diffusion of Complex Innovations as a Process of Opinion Formation Through Social Networks.” She joined Knowledge@Wharton to talk about her findings about which sorts of improvements unfold extra rapidly than others in several networks, the position of influencers, and what which means for entrepreneurs.
An edited transcript of the dialog follows.
Knowledge@Wharton: What was the inspiration for this analysis?
Valentina Assenova: The inspiration for this analysis was taking a look at the unfold of microfinance. Microfinance is one of these improvements that isn’t clearly good or unhealthy, and there’s a lot of blended proof round whether or not or not it’s really helpful for ladies, whether or not it improves welfare and so forth. But it was one thing that basically acquired me intrigued about the position of public opinions and of social networks — in the sense of individuals who you speak to for recommendation, for assist in making a choice — and the way some of these advanced improvements unfold.
Knowledge@Wharton: Tell us a bit of extra about what you imply by advanced improvements?
Assenova: Complex innovation is basically an innovation that has rather a lot of uncertainty round its worth for a possible adopter. When we take into consideration a fancy innovation, we’d take into consideration an unproven know-how. This is a know-how that no person but is aware of whether or not it will likely be roughly helpful in relation to current options.
Given the uncertainty round this know-how, sometimes there’s a want for some form of social validation for different individuals to undertake it. This is contrasted to one thing like a easy innovation, like penicillin or a medical innovation that could be very clearly good. With just some adopters, there may be not as a lot of a necessity for social validation for one thing like that to unfold in a inhabitants.
“There are certain features of network structure that are more conducive to the diffusion of a complex innovation than others.”
Knowledge@Wharton: What is included on this form of social validation or social community?
Assenova: When we take into consideration social networks in social science and in administration, we have a tendency to assume of patterns of interplay amongst individuals. A social community may very well be individuals in the workplace that you simply work together with day by day. Those interactions will be represented as a graph, as a community, or it may very well be a community of individuals that you simply speak to on Twitter and that you simply observe. It is basically who’s influencing you in your opinions and in your beliefs about the worth of sure concepts or sure applied sciences.
Knowledge@Wharton: How did you take a look at this?
Assenova: The paper is a theoretical mannequin that appears at studying on these networks utilizing a DeGroot naive studying mannequin. I start by taking a look at random networks, then I transfer on to testing it. In the case of the unfold of microfinance in India, the magnificence of utilizing these very simplified … random graphs is that we all know precisely what is occurring on the community. We can take a look at very cleanly what the mechanisms are. In this case, the key mechanism is affect from the opinions of different individuals.
Knowledge@Wharton: What did you discover once you checked out these fashions?
Assenova: I discovered a pair of fascinating issues. The first is that there are particular options of community construction which can be extra conducive to the diffusion of a fancy innovation than others. Remember that we don’t know what the worth of these improvements are, and in idea we’d by no means know. What individuals are going off of is the opinions of different individuals about the worth of these improvements. There are a pair of completely different options of this mannequin. One of them is that we are able to manipulate the construction of the community itself in phrases of the density — how related individuals are to different individuals — and in phrases of the asymmetry of the relationship — whether or not I’m extra probably to affect you than you to affect me.
What I discovered in utilizing these fashions is that networks with each excessive density and excessive asymmetry are optimum for diffusing advanced improvements when the boundaries to adoption are low. That principally means now we have a case of a reasonably easy-to-adopt innovation. It’s quite simple. It doesn’t have rather a lot of added steps to what it takes to undertake it.
Conversely, when now we have an innovation that could be very advanced, the place the boundaries to adoption are very excessive, it’s really the reverse. Here we’d assume of a reasonably advanced know-how that requires rather a lot of extra enter and data about how to use it. So, low density and low asymmetry networks are the ones which can be almost certainly to diffuse these improvements.
Knowledge@Wharton: What does a excessive density, excessive asymmetry group versus a low density, low asymmetry group appear like in the actual world?
Assenova: We can take an workplace for example. If we take a look at an workplace the place individuals are sitting and speaking to one another, and we had been to map out the community of their communications, a really excessive density and a really excessive asymmetry community would appear like one the place everyone in that workplace is speaking to everyone else. They are interconnected. Moreover, there may be one particular person or two individuals in that workplace who’re dominating that dialog. Their opinions are much more influential than anybody else in that group.
By distinction, a low density and a low asymmetry group would appear like one the place only a few individuals in the workplace are speaking to few others. Communications are rather more selective, and a few of these variations in whose opinion issues are much less pronounced. I’m simply as probably to hear to your opinion as I’m to one of my different colleagues in the workplace, and there isn’t a single one who is dominating the dialog.
“There is this interplay between my own values and my own thresholds, and the opinions of other people that I am influenced by.”
Knowledge@Wharton: In my life, I’m probably to be half of each sorts of teams. How does that come into play after we are taking a look at the unfold of innovation throughout completely different elements of life?
Assenova: That is a incredible query, and it’s a matter of a latest stream of analysis that I’ve been taking a look at, which is multiplex networks. In actuality, individuals are embedded in a couple of form of community. You have your mates, your co-workers, your loved ones, and people networks might look very completely different. We are simply starting to discover how diffusion dynamics would possibly matter in these varieties of networks. The preliminary findings present that there are differing kinds of multiplexity which can be roughly conducive to diffusion, and one key issue that issues is how broadly you’re spanning these completely different networks and the way unconnected they’re. That could be a fairly large predictor of whether or not you, as a key one who is a spanner throughout these networks, are in a position to affect different individuals inside them.
Knowledge@Wharton: How does it assist any individual who’s making an attempt to unfold an innovation to perceive the networks and the way they affect individuals?
Assenova: That is a good query. When it comes to improvements with unknown worth, like new applied sciences, what are the greatest communities to goal and who’re the greatest individuals to goal for getting the phrase out and actually selling adoption? I believe this analysis has a pair of key implications for an entrepreneur or practitioner. One of them is to actually take into consideration the boundaries to adoption for this know-how. Are these boundaries comparatively excessive or pretty low?
If any individual is creating a reasonably easy app that they consider could be extensively invaluable and relevant, then focusing on a excessive density and excessive asymmetry community the place they’re figuring out the key influencer could be the manner to unfold it, and that will be the manner to promote its very speedy diffusion.
Conversely, if any individual is creating a biomedical innovation that’s tough to perceive and tough to type an opinion about, then one would possibly need to goal a community that has decrease density and decrease asymmetry, resembling a community amongst skilled physicians who is perhaps educated to use this know-how after which speak to one another and rationally consider what the prices and the advantages of this know-how are. In very sensible phrases, it simply means it offers entrepreneurs a manner of deciding on the proper viewers and the proper communities for diffusing their applied sciences.
Knowledge@Wharton: You have to perceive who your clients are or who they may very well be, appropriate?
Assenova: Exactly. And who’re the key individuals who will likely be forming an opinion about the worth of what you’re proposing, and what’s the greatest manner of creating social contagion inside the networks of those that they contact and that they convey with. I believe that may be a key takeaway in understanding what it means for a selected know-how in a selected group that any individual is making an attempt to goal.
Knowledge@Wharton: I’d assume understanding how they’re going to talk about the innovation can be essential. Are they going to do it by way of on-line social networks? Is it going to be phrase of mouth? Is it going to be one thing else?
Assenova: This mannequin is agnostic as to the mode of implementation. Certainly in these varieties of networks, the assumption is that the affect that individuals are getting is primarily by way of the evaluations of different individuals, and that issues. It’s not the form of know-how the place it’s straightforward to simply go on a web site, examine it and make sense of it.
This is basically the place the complexity comes from. It is that this want for social validation by way of different individuals in a community to perceive that know-how. But definitely when it comes to these evaluations, making an attempt to perceive how the construction of the networks shapes the formation of these opinions and whether or not individuals are forming a consensus that one thing is basically invaluable or not, is a crucial factor of making it profitable.
“The fact that you can diffuse an innovation more quickly and more broadly in a sparser network … calls into question the conventional wisdom around choosing key influencers.”
Knowledge@Wharton: Were you stunned by any of the findings?
Assenova: I’d say I used to be a bit stunned by the second discovering of actually taking a look at how the thresholds to adoption average the advantages of density and asymmetry as a result of I believe the fashionable conception in the literature is that density is all the time or comparatively good. Density and asymmetry could be conducive to diffusion. When we discuss key influencers — that’s precisely what this literature is referring to. These are the people who find themselves very, very related in a community and may instantly affect tons of different individuals by posting one thing on Twitter or voicing an opinion. What is stunning is that these sorts of networks and people varieties of individuals are not universally helpful for diffusion. It relies on the boundaries to adoption associated to particular know-how.
The truth that you could diffuse an innovation extra rapidly and extra broadly in a sparser community is stunning and fascinating, and it calls into query the standard knowledge round selecting key influencers and selecting high-density networks.
Knowledge@Wharton: What is it about that top barrier to entry that created that consequence, the place they did higher in networks that didn’t have rather a lot of asymmetry and weren’t very dense?
Assenova: I believe there are two components to it, and there are two components that I take a look at in the mannequin. One of them is the truth that individuals have a selected threshold for adoption. For instance, I’m keen to swap to the new iPhone whether it is X many occasions higher alongside the dimensions that I worth in relation to the current know-how. People have this threshold about the worth they would want to have the option to get out of this for it to be worthwhile switching from the current resolution. The second factor of that’s associated to threshold to adoption. There is clearly a consensus worth that’s forming inside a bunch of those that I do know, so if I worth your opinion fairly a bit and also you’ve switched to the new know-how and you’re altering my opinion about it, I’m extra probably to swap over.
There is that this interaction between my very own values and my very own thresholds, and the opinions of different those that I’m influenced by. It’s that interaction that basically does have an effect on what individuals find yourself doing and the way broadly an innovation spreads inside a inhabitants.
Knowledge@Wharton: What is subsequent for this analysis?
Assenova: There are a pair of completely different initiatives that I’ve been engaged on. One is taking a look at the position of multiplexity in networks and the position of group leaders or opinion leaders in these networks and the way they span these networks over time. I’ve a second paper that may be a follow-up to the research, actually parsing how multiplexity, this overlap in several networks that individuals have, impacts the diffusion of these improvements. I’ve one other paper taking a look at how enjoying a number of roles inside these communities — say, being each a developer and a consumer on a platform — would possibly have an effect on the diffusion of some of these improvements.