The Case Study


Uncovering The Influencers in Social Network

Kcore Analytics Influencer Algorithms help to optimize targeting influencers in marketing campaigns.

We have teamed up with Grandata, an Argentinian Corporation that integrates financial and telecommunication data at the country level, to perform a marketing campaign targeting the influencers identified by our algorithms. 

Our main finding is that people location in the social network, measured by our state of the art Collective Influence Algorithm and Optimization Theory is a powerful predictive indicator for their economic status. Said in very simple terms, we discovered that people located in positions of high collective influence found by our algorithms also have high economic levels. This result comes from the unprecedented combined analysis of two large-scale datasets including telecommunication and financial data of 110 million people in Mexico, and the development of cutting edge network theory.


Our work unequivocally shows that information about financial status is encoded in the network of social ties, and can be easily decoded by our influencer algorithms.

Using our powerful social predictor of economic well-being, we carried out a real-life large-scale social marketing campaign by targeting the identified influencers in the whole social network of Mexico. The results of our marketing campaign quantify the pragmatic benefits of targeting our influencers. We found a remarkable five-fold increase in marketing success rate obtained by targeting the individuals identified by our collective influencer algorithm as compared to other measures.


Specifically, we used the mobile communication network of all residents in Mexico resulting from the aggregation of calls and SMS exchanged between users over a period of 3 months from December 2015 to February 2016. The resulting social network contains 110 million people and 351 million links. The campaign was conducted on a total of 656,944 people who were targeted by a SMS message offering a consumer product according to their ranking in our list of top influencers in the social network. We also sent messages to a control group of 48.000 nodes, chosen randomly. To evaluate the campaign, we measure the response rate, i.e. the number of recipients who requested the product divided by the number of targeted people as a function of the influencers ranking. In the control group, the response rate to the messages was 0.331%. On the contrary, groups targeted according to our influencer algorithm showed a sharp increase in their response rate, with a sound three-fold gain to 1% in the rate of response of the top influencers identified by our algorithms when compared to the random case.


© Copyright 2018 Kcore Analytics - All Rights Reserved

Show Buttons
Hide Buttons