3 Ways Big Data Is Improving Telecom
Big data has hit nearly every industry; healthcare, retail, defense, and others have all reaped in major benefits like increased revenues and simpler solutions. It seems that telecommunication services, too, are jumping on the bandwagon. In seeing that telecom operators have lost a large amount of revenue due to the recent elimination of roaming charges in the European Union -- not to mention, the rise of Skype and WhatsApp and the decline of landlines, big data analytics might just be how telecommunication providers counter these revenue losses. How? Let's find out.
Boost Customer Experience
It's well-known that telecommunications and wireless service providers handle and store a lot of data; they have to deal with hundreds of thousands of users' call detail records, internet traffic and transaction data, usage patterns, technical fault data, and location data, to mention some. Unfortunately, because operating costs are so expensive, most of it just sits there -- but what's the point of retaining all this data if not to analyze it and put it to good use?
Fortunately today, big data analytics can help telecommunications providers analyze network data quickly and inexpensively. This way, they can focus on the data that matters, gain insights on important trends, and optimize their services accordingly.
In Israel, for example, a leading mobile operator found themselves unable to react to important events (including call drops, signal strength, and handovers) in a real-time fashion while extracting network statistics offline -- meaning, they were basing network performance and user experience improvements on historical events rather than online troubleshooting. It was too expensive, unfortunately, to implement a real-time solution because it required installing probes on each base station. In adopting SQream DB, an economical alternative which uses GPU-based technology to boost analytics performance through massive parallel computing, the mobile operator could suddenly store more data over longer periods and also monitor the entire system over a broader timeline in real-time -- permitting the mobile operator to identify network failures and accordingly improve their customer service representatives' responsiveness.
If you didn't already know, communications fraud, or using telecom products and services without paying, is a major problem for telecom providers today -- in 2015, experts estimated fraud losses at $38.1B! In the same survey, the Communications Fraud Control Association (CFCA) discovered that the most popular types of fraud are: international revenue share fraud, interconnect bypass, premium rate service, arbitrage, and theft or stolen goods. And, the most popular methods for committing fraud are: PBX hacking, IP PBX hacking, subscription fraud, and dealer fraud.
It seems, that with the prevalence of smart devices and the growing use of IP networks, communications fraud is only getting worse. But big data can do a lot today to prevent the onslaught of fraudulent activity -- for example, by correlating internal location, usage, and account data with credit reports.
This is pretty much what Argyle Data did for Vodafone. Vodafone was suffering; in addition to substantial revenue losses, fraudulent activity was costing them millions (!) of customers by negatively impacting Vodafone's brand perception and net promoter scores. It was difficult for Vodafone to keep up with fraudsters' increasingly sophisticated techniques, especially when fraudsters learned just how to bypass standard fraud prevention algorithms. But, in partnering with Argyle Data, a platform just for detecting revenue threats, Vodafone hoped for change: Argyle Data streamed Vodafone's traffic data and searched for abnormalities, problems, and inconsistencies using adversarial machine learning, or machine learning designed to catch criminals.
It's no longer possible for telecom companies, much less any large corporation, to depend on old-school marketing techniques like cold calling, mail flyers, and billboards. Rather, because of the intense competition among businesses, corporations must work harder to distinguish themselves in order to grab new customers and retain old ones.
Usually, this means spending more money. But, of course, big data analytics is able to serve these needs while still remaining cost-effective. In adopting a data-driven and targeted marketing approach, telecom companies can increase customer onboarding, interaction, and satisfaction.
Ufone found themselves in a highly competitive market as one of the five major telecom operators in Pakistan; their churn rate was quite large. The customer-care department, marketing department, and third parties at Ufone relied on outdated versions of their customers' call detail records to conduct campaigns -- plus each department was drafting campaigns completely unilaterally, unconcerned with what the other departments were creating. Ufone recognized this inefficiency and deployed several IBM products, including IBM InfoSphere Streams, IBM Unica Campaign Management, and IBM Websphere. Using these solutions, Ufone was able to target and engage various customer segments, like those who seemed likely to leave the network, those who texted often, those who called often, etc. Then, after segmenting their users, Ufone could text certain segments special offers, hoping they respond positively and purchase the upgrade.
Follow Jasmine Cohen on Twitter