Real-time Big Data Analytics for Telecommunications

InfoSphere Streams Product Manager

The telecommunications industry has dealt with big data for years, for example tracking individual wires into peoples’ homes and places of business. Detailed billing to track time and duration of each telephone call, even for local billing in the 1980s, introduced a huge influx of data to manage and process. Cell phones have exacerbated big data requirements. With each cell phone call, one or more Call Detail Records (CDRs) is created – if you are driving and you have a long call, you might get a CDR from each cell tower you pass through. With each browse of a web site, an Event Detail Record (EDR) is created with information like the URL browsed and the amount of data downloaded. This data is required to generate the monthly bill with what’s owed for postpaid plans. Prepaid plans do it in real time to prevent misuse.

With the huge volume of data, and the desire to process and analyze it in real time, IBM InfoSphere Streams (“Streams”) has been a natural in the telecommunications industry. Starting with some telcos in India where churn rate was 11% per month[1], Streams was used to perform mediation of these call and event detail records (xDRs) in under a minute. This huge improvement over the prior 12- to 48-hour process not only saved time and money for the IT department, but it allowed the telcos to immediately analyze the data. Churn prediction models and revenue assurance applications could more quickly identify issues that could be rectified.

Over the years, dozens of telcos have adopted Streams for a variety of applications beyond mediation. Applications like real-time campaign management, real-time advertising, cybersecurity detection, and law enforcement deep-packet inspection are a few other examples.

Real-time campaign management uses Streams to understand immediately when campaign tactics have been achieved, like sending a certain number of text messages or using a certain number of calling minutes. This allows providers to immediately recognize their customers’ usage patterns and suggest further offerings to upsell. With Streams and a rules-based campaign management system, one telco improved marketing effectiveness 600% and reduced development time and cost by 95%. Several companies are using a combination of Streams and IBM Campaign (formerly Unica Campaign) to deliver and track the actual campaign tactics. Read more about how PTML Ufone uses these for Smarter Campaigns.

Fraud detection is growing in importance. One telco believes they can save up to $300M per year by more quickly observing voice and data fraud and preventing further misuse. Network and set-top box monitoring has been implemented by one European telco to ensure quality distribution for video through their network. A predictive analytic algorithm monitors the set-top box log messages to predict outages. After all – if the network quality is poor, or the set-top box is out of service, customers may call to complain, resulting in lost revenue and added expense.

With many success stories across multiple applications already completed, IBM has announced our intention to purchase THENOWFACTORY. This privately held provider of analytics software helps telcos deliver better customer experiences and drive new revenue opportunities. Integrating The Now Factory into IBM means we can deliver Streams-based applications for telcos to help them perform real-time analytic processing (RTAP.) The result for telcos is increased revenues and lower costs.

Related Content 

Animation: Big Data, Big Opportunities: Telecommunications 

Blog: XO Communications uncovers millions of savings with PureData System for Analytics

Video: IBM PureData System Testimonial featuring XO Communications

Blog: Connecting Network & Subscriber Domains with Big Data & Analytics: IBM’s Intent to Acquire The Now Factory

[1] SOURCE: The Wall Street Journal, December 22, 2011