Customer Churn and Upgrade

An advanced Machine Learning and big data solution that predicts the customers’ propensity to churn or upgrade, leveraged by plethora of Customer Segmentation metrics built from diverse data streams.

For Tier 1 operator in Hyper price-competitive market.


Which of my current 2G or 3G customers are most likely to lap up 4G subscription?

Which of my current customers are most likely to churn to my competitor?


Created 200+ customer features encompassing network experience, device experience, care experience, retail experience, customer usage and travel patterns.

Creation and analysis of micro segments of customer based on the features created.

Run ensemble modelling techniques for figuring out the highly probable customers to be targeted by the marketing campaigns.


30% improvement in correct identification of prospective churners.

70%-win backs achieved.

Competitor Network Throughput Prediction

Prediction of throughput for wireless networks of competitors (on a granular geolocation level) using crowd source data, powered by ensemble Machine Learning models, to drive various network strategies.

For Tier 1 operator in North America.


How can we measure the throughput of our competitors’ network without extensive drive / download tests?

We would like to compare our network throughput performance vis-à-vis competitors on a granular geo location over a period of time.


Data throughput modelling using crowdsourced data (only a few RF parameters).

Identification of grids and areas with poor performance as compared to competitor.


Model accuracy of greater than 85%.

Highlight of important variables impacting poor throughput.

Broadband Planning

A GIS based advanced analytics solution to enable business teams to identify prospective high return areas for broadband deployment.

For Tier 1 operator in APAC.


How to figure out potential areas for broadband deployment without carrying out field surveys?


Data collection (from internal and external sources) and analysis of key components influencing successful broad band deployments such as customer data usage level, customer affluence, presence of competitors etc.

GEO spatial analysis to build User metrics and characteristics at location level.

Overlay the perspective areas for deployment with the current fiber deployment to calculate / estimate ROI of the deployment.


Key variables identified to create a composite 'area potential score'.

Standard methodology for prioritization nationwide leading to best “bang for buck” deployment.