Churn prediction for telecommunication companies

Contageous churn Our algorithms are based on the occurrences of contageous churn in mobile networks. It is shown that when customers churn, the neighbours in the call network have a higher probability to churn in the future. This might happen only several weeks after the original churn event.

Our application, SNA Engine, integrates into your infrastructure and retrieves at scheduled intervals (for example overnight) the call data records and stores the predicted churners in any existing CRM system.

Identify brand influencers among customers for retail companies

Viral marketing is used more and more, also in the retail sector. It is based on the Word Of Mouth (WOM) process, where good (and bad) experiences with product and services are shared with friends, colleagues and relatives. And in turn, those people will share the same message with their social network.

One way to conduct is viral marketing campaign is to identify a segmented customer population and target them with a marketing message which might include free product samples or promotion offers. To reduce the campaign cost, the target population needs to be selected carefully.

Marketing campaigns could target customers with a high Customer Lifetime Value (CLV), which is often defined as the expected profit from sales during the lifetime of the relationship between the customer and the company. But this measurement ignores the influence customers have on the buying behaviour of others.   » read more

Fraud detection for insurance companies

Often fraudsters do not work in isolation, therefore the best way to fight a criminal network is to take it down as an organization and our analytics software helps organizations in doing this by providing insight into the big picture. Our solution, SNA Engine, contains routines for fraud prediction from insurance claims. We go beyond the individual claims and analyse networks to detect fraud.

Our SNA software gathers data from multiple sources and links individuals who share information, are involved in claims or engage in transactions with each other. All this information is analyzed using a predictive model which provides a fraud probability score. This can easily be integrated with existing case management software as an additional rule.

A Personal Data Framework based on a Big Data Architecture for Privacy-Preserving Analytics

With the explosion of big data analytics the last years, more customers or citizens are concerned about the usage of the personal data. Most companies and organizations are not transparent on how they use data on their customers.

This creates mistrust with the customers which is even enlarged when the data stored on the servers of the companies are hacked and made public. These trends, e.g. usage of (big) data to analyze customer behavior and the increasing hacking attempts that violates the privacy of their customers has made the issue of personal data more compelling.

  » read more

Churn Prediction

White paper: Social Network Analysis & Decrease Churn in Telecom

Social-3 applied its SNA solution to the Telecom Industry based on call data. This white paper provides detailed insights.

Marketing Influencers

Find Fashion Influencers

This example report identifies influential fashion bloggers with a good follower base, in the cities Florence, Milan, and Rome.

Fraud Prediction

Presentation on Fraud Prediction (Spanish)

Presentation by Social-3 given on fraud prediction using social network analysis for insurance companies. The seminar (Fight Against Insurance Fraud) was organized by ICEA.

Copyright © 2011-2016 Social-3. All Rights Reserved.