Jake Kendall

Tapping the Network to Fight Poverty: How targeting influential leaders can increase financial innovation uptake

Whether you measure it by Facebook “likes,” Twitter followers, or just by how crowded your kitchen is at dinnertime, it’s clear that some people have larger social networks than others. And a few individuals invariably emerge as central points of contact for their larger social groups.

Innovative research suggests that activating these influential network “hubs” can affect the diffusion of information about a new product – a potentially important finding for those who are trying to introduce financial innovations at the BoP.

The research, published by Banerjee, Duflo, Chandrasekhar, and Jackson in the journal Science, takes a very interesting data set comprised of social network maps of 43 rural villages in Karnataka. The data also documents the pattern of adoption of microfinance loans by individuals in the villages, after an initial intervention by the microfinance lender who met with specific leaders in the village to introduce the product.

For the researchers, the leaders serve as the “injection points” of the new innovation (credit) into the network. The data allows the researchers to track how the subsequent adoption by others flows through the social network of the village. From these observations, they create a network model of how information flows that allows them to measure how influential someone is in spreading the innovation to others (people with higher scores are more “central” to the flow of information through the network.) They then show that these measures could, in theory, be used to target marketing to improve uptake. In fact, the data indicates that if they were to use the targeting measure to select only leaders with the 90th percentile of “centrality,” they would improve microfinance uptake by over 10 percent over the average current uptake of only 18.5 percent.

How is this relevant to our work in mobile financial services?

One of the things the researchers point out is that both the measures of centrality they came up with are “information intensive” (this is researcher-speak for “expensive.”) One approach requires mapping of the full network by conducting face-to-face surveys of everyone in the village. Researchers must then also conduct an experiment by injecting innovation to estimate the parameters of the model before finally going back to calculate the “centrality” measure – thus it would be infeasible as a targeting tool. Another approach requires just the map of the social network from which they estimate the measure without the experiment. This is more feasible, but still it’s quite expensive to go around asking everyone who their friends and colleagues are in order to map the network. If you are going to do that, why not just market the product to everyone?

This is disappointing, because it seems that there is significant potential to drive uptake of financial products more efficiently by exploiting knowledge of the social graph, but it’s too expensive to map the network. Hmm, if only there were a way to get the map for free… Well, maybe there is. Mobile network operators have a natural social network map embedded in their database already. The call and SMS patterns they have form a partial social network map and are created for free by people as they talk or text on their phones to their friends, relatives and associates.

I spoke to Abhijit Banerjee (one of the study’s authors) in Boston several weeks ago, and he was quite keen to try out the same techniques using mobile data. Partnering with a mobile operator would allow a much more sophisticated approach, permitting experimentation with tactics and incentives to maximally activate the viral adoption cycle.

The hard part is getting an MNO to try an innovative but untested new approach like this. If any of you mobile operators out there feel you want to pursue this kind of work, partner with some of the top economists and network mappers in the world, and are willing to allow a little openness in sharing the results after the fact, let me know!

Jake Kendall is a program officer in the Financial Services for the Poor initiative at the Bill & Melinda Gates Foundation.

financial inclusion, microfinance, research