Our client is a leading Telecom company in India with a subscriber base of over 180 million. The company is a GSM operator offering 2G, 3G and 4G mobile services and is the third largest in the geography in which it operates.
This is a story about how the Ellicium Gazelle platform made it possible for our client to get near real time data of the mobile experience of its subscribers; something which was not possible using its current IT systems.
Our Telecom client has subscribers spread across India. Due to the huge geographic spread of its cellular service and customers, there is a wide difference in terms of the mobile service which the subscribers experience at different places and at different times. There was no mechanism of gathering reliable data related to Signal Strength, Data Speed, App Usage or Browsing Behaviour and analysing it. This severely restricted their ability to resolve network issues, improve service, predict churn or design effective marketing campaigns based on mobile usage.
Solution: Ellicium Gazelle framework for streaming data was used to achieve the given objectives
Ellicium worked with the IT Strategy team of this Telecom client to understand the data which could be captured on the mobile devices, the various data formats and their frequency of capture. The various visualization and analytics requirements which were needed were also discussed. Various options for data visualization and geospatial analysis were considered and evaluated on parameters such as look-and-feel, ease of use, accuracy, cost, response time and integration with Gazelle. In conjunction with the client, it was decided that Google Maps would be the best fit for their requirements.
The configurable data ingestion mechanism of Gazelle and the pre-built data validation accelerators made it very easy to pre-process the data and ingest it into the Hadoop Data Lake. Further, a detailed design of the data aggregation mechanism and analytics for the data was defined, designed and agreed with the client IT team. The customized analytics was built using the Hadoop ecosystem tools such as Impala, Hive and Python scripts.
Subsequently the entire Gazelle solution was customized using the Agile methodology and deployed on their 5 node Test cluster within a period of 12 weeks. It was also tested for accuracy and robustness for a continuous run of 2 weeks using several hundreds of Gigs of data from subscriber handsets.
A detailed Production Implementation roadmap was also delivered. This used Ellicium’s Hadoop Cluster Sizing Tool for the Production cluster which gave all details needed for the initial sizing of the Hadoop cluster. Further details of the incremental manner in which the cluster would need to be scaled up, depending on the number of subscribers whose data would be captured and analysed, was also provided.
As a direct result of being able to measure the actual Customer Experience of their mobile service, the Telecom company could achieve something which they were never able to do earlier:
- Visualize the Customer Experience data related to Signal Strength, Data Speed, App Usage or Browsing Behaviour at an individual and aggregated level; in less than 15 minutes of the data landing on the Gazelle server.
- Analyse all the above parameters on Google Maps and do geospatial analysis, right from an aggregated pan-India level to a drilled down level of 200m X 200m area.
- Work on Network Infrastructure Optimization by correlating data from Subscriber handsets and Cell-Sites data.
- Design targeted Marketing Campaigns by segmentation using data related to App Usage, Browsing History and Crowding analysis.
Overall, the Ellicium Gazelle Platform has revealed Customer Insights which were never available to the Telecom client earlier. This has allowed the client to stay ahead of its competitors in the very competitive Telecom space and thereby remain amongst the top-3 leading providers.