Orange is one of the largest operators of mobile and internet services with sales of €39 billion and 154,000 employees worldwide. Present in 29 countries, the company has 248 million customers worldwide, of which over 75% are mobile subscribers. Previously known as France Telecom, Orange is not only France’s largest telecom provider but has also consistently been awarded France’s top voice and data provider for “Mobile Network Service Quality.”
Understanding Subscriber Usage
Wireless markets worldwide are maturing rapidly; wireless penetration is well over 85% in France. Most operators are battling to keep their customers from switching and at the same time encouraging their subscribers to buy more services. Understanding subscriber usage can capture more revenue through tiered pricing or specific packages for social media, video streaming or map applications.
Up until recently, most of the innovation has been at the network level, with little to no visibility into subscriber application usage. In order to obtain these insights, Orange needed to analyze millions of source and destination IP addresses with subscriber data to capture individual data usage. Given the massive amount of data to be ingested and processed, Orange recognized this problem was an ideal use case for Big Data technology but didn’t have the technical expertise or tools to create a solution.
Analyzing Subscriber Activity Patterns
Orange Silicon Valley, the technology and strategy division of Orange, selected Zaloni to architect and build a Hadoop solution. Zaloni was tapped because the company is a recognized industry leader in implementing big data lakes. A data lake is a repository containing a vast amount of raw data, in native formats that allows different users to analyze and manipulate that data for multiple applications.
First, Zaloni created a data lake to capture terabytes of Call Details Records (CDRs) created by the wireless network. Second, Zaloni, correlated network data (IPFIX, DPI, Bulk Stats) with CDRs to create meaningful, actionable subscriber insights and presented this through a dashboard. Underlying this dashboard is a Hadoop solution that enables near real-time processing, that groups customer into clusters based on usage patterns (voice, data, SMS).
Gaining Real-Time Insights Through Dashboards
When the customer groups are overlaid with physical maps, Orange can answer key questions such as:
- Which clusters are using the most data and where are they geographically located?
- What services are popular, by cluster and device (handset, tablet, hotspot, manufacturer)?
- Which OTT packages will have the best take rate?
For the first time, marketers are able to graphically visualize subscriber usage as it occurs. Orange can modify its pricing plans and offer data packages tailored to its subscriber base.
In addition, Orange can take the analysis a step further by leveraging machine learning to identify the most profitable customer segments. Engaging Zaloni to create a Hadoop solution for subscriber analysis is helping Orange retain its status as #1 provider of mobile services in France.