Big Data Analytics: Telecommunications Industry use Case

With the proliferation of always connected data hungry devices ranging from smart phones, tablets, connected cars, internet of things and so on, the need to analyze the data quick and make rapid informed decision has increased many folds in last few years (MapR, n.d.). Some of the areas where big data analytics can help are improving network services and usage, enhancing customer relationships and better security. Key performance indicators mapped to above areas could be:

1. Forecasting the hours of high network utilization and congestion and suggest steps to mitigate

2. Improving customer churn rate by timely identifying their concerns and fix

3. Recognizing delinquent customers and suggest steps to improve recovery

Data available to Telecom Service Provider

Telecommunications service providers collects huge amount of data varying from customer specific like customer complaints, billing, service plans, devices, social media, experience to network specific like call detail records (CDR), network performance, equipment alarms, logs etc. in several granularities like hourly, daily, weekly, monthly etc. Using data coming from these many sources, some of the possible use case areas of big data analytics are:

1. Customer Related

2. Network Related

3. Security

Within customer related, below are potential metrics that can be analyzed:

1. Churn analysis

2. Targeted marketing

3. Optimal subscriber plan

4. Opinion, sentiment and customer review analysis

5. Analysis of clickstream

In the area of network specific use case, below are some potential metrics which can be enhanced using big data analytics:

1. Network performance optimization by analyzing alarms, logs etc.

2. Forecasting of capacity demands and remediation plans

3. Prediction of network equipment failure

4. Creation of service plans based on customer needs

5. Proactive customer service

6. Monetization of GIS location data for advertising, marketing etc.

Security has become an area of paramount importance with the advent of millions of connected devices thus creating more malicious entry points. Big data analysis can help a telecommunication service provider in following areas:

1. Real-time or early fraud detection and prevention

2. Privacy compliance

3. Safe payment processing

Use Case – Data Transfer

In the case of cellular service provider, lots of network information and customer information need to move from the location of network node to central office for monitoring and other analytics. Usual way of moving data was FTP which can be terribly slow in moving huge amounts of data from thousands of field locations to central office. Solution by MapR is to use MapR stream in regional data center first by performing parsing, joining and aggregation functions and create regional dashboards. These regional dashboards are than replicated from regional to central office using many-to-one configuration for global and real time monitoring of performance to improve network performance and customer satisfaction using following metrics:

1. Customer location based service issues

2. Real time optimization of network resources using subscriber usage pattern at an accident location to major sports event

3. Network related metrics like

a. Drop call

b. Coverage issues

c. Slow speed

d. High latency

Use Case – Complete Customer View

Another use case of data analysis is to understand, analyze and predict the customer behavior so that necessary actions can be taken to continue to win the customer’s approval. Some factors that can be analyzed are:

1. Demographics (gender, wage, age, etc.)

2. Customer feedback, opinion, sentiment analysis of online posting including social media

3. Usage pattern based on location

4. Usage pattern based on websites visited or advertisements clicked

5. Correlation of customer to predict churn

Use Case – Detection of Threat detection

Using data analytics, network breaches or vulnerabilities can be quickly identified and remediated. Machine learning and pattern recognition can quickly find out the root cause of erratic behavior exhibited by a network node or equipment. Similarly fraudulent use of resources to buy services, devices etc. can be detected sooner than later. Location information of customer, their usage pattern and device/service portfolio can raise flags if some unusual purchase is observed.

Further Areas of Improvement

For a network carrier, successful deployment of their service in a new geographical location can be analyzed using the sentiments over social media as well as customer reviews on an online portal. Network metrics will also show the degraded service in the form of drop calls, coverage complains and slow speed for the given area where new network deployment is planned thus success and improvement can be predicted for accurate return-of-investment calculation after the deployment of new network assets.

Using the information of customer’s billing cycle and contractual plan, network carrier can analyze if customer start to browse competitors network’s offerings which can point to possible defect in future at the end of current contract. Analysis of latest customer interaction with customer support and posting of reviews over social media can also be a good indicator to detect future churn.

Besides using network’s own information, publicly available information of competitor’s network can be analyzed to identify advantage and disadvantage areas for possible marketing campaigns and remediation plans respectively.

Periodic customer engagement surveys can serve as a good source to detect the trends of industry and what customer is looking for. Service quality, quantity and price can be compared against other carrier’s using surveys to detect the factors behind customer’s perception of a given network.

Reference

MAPR (n.d.). Big data opportunities for telecommunications. Retrieved from https://mapr.com/blog/big-data-opportunities-telecommunications/