Telco Big Data and real-time analytics – the penguins start jumping into the sea

The Telco Big Data & Real Time Analytics Summit held in London last week spanned a surprisingly wide range of data management, BI and analytics-related topics and delivered useful insights from operators, vendors and other industry players. However, it was the operator presentations that stood out, as much because we’re now getting to a stage where operators are beginning to dip their toes in the water with projects or launches that include elements of Big Data or real-time analytics.

As a speaker from Sprint Advanced Analytics Lab pointed out, for operators the coming twelve months will largely revolve around seeing if the penguins that jump into the water get eaten or not. That will then determine what happens next in the form of solid telco investment in this space.

It should be stressed that many of these early operator initiatives will still involve a strong element of traditional business intelligence and analytics, rather than any large-scale investment in Big Data and associated distributed data processing systems as such.
Conference presentations from Telefonica Digital, Sprint and Deutsche Telekom were particularly interesting from this perspective.

Sprint, for one, claims to have had some early success selling telecoms data to marketing agencies. The operator has been experimenting with positioning itself as an information broker and has already worked on some early engagements where it has created location-based customer insights for retailers. The operator sees retail and gambling industries as obvious targets, although early opportunities have been wide-ranging, including amusement parks wanting information on the location and behaviour of visitors.

Sprint has mainly focused on call records believing there is more than enough untapped information in these without having to look further afield at this initial stage, but the potential sources of data are many.

Structured data sources include: call, billing, electronic data, network and location data records. Unstructured data sources include: phone calls and text messages, social media, web browsing records, media downloads, eBooks and newsreader content as well as apps usage and interaction.

All of the above can be used in combination or matched with sample data from other sources in order to micro-segment or make inferences about behaviour patterns.

There is also potential for operators to make use of analytics to position themselves as OTT enablers in new markets. Deutsche Telekom, for example, is about to launch a series of real-time insurance services that are enabled by a rich mix of location-based, demographic and other data sources. The first of these independently-branded services is a real-time insurance offering aimed at young skiers, a frequently uninsured demographic . It targets them via a smartphone app that identifies when someone fitting a particular profile arrives at an appropriate airport or ski resort.

A lot of the conference presentation examples or use cases focused on making better use of real-time analytics or Big Data to support micro-segmentation. It is inevitable that some of the more sexy areas – such as data brokerage and targeted real-time marketing analytics and customer care – will attract attention, but it is important to remember that are just as many opportunities on the operational side. The role Big Data and real-time analytics can play in improving operational efficiencies should not be underestimated as they can provide detailed context for decisions about network and service quality as well as support faster decision making and so allow operators to be more proactive in addressing network or service problems.

It is clearly still early days with regards to the adoption of Big Data or real-time analytics in the telecoms space. There do not, for example, appear to be many Apache Hadoop platform implementation plans as yet, or if there are then operators are keeping very quiet about it. Proof enough of this was the fact that we may have been attending a Big Data-branded conference this week but many of the use cases discussed at the event involved more traditional BI or analytics approaches and where Big Data was specifically discussed it was frequently in the context of adding value to existing projects.

This is partly a reflection of the relative immaturity of this market, but it may also be an early indication that Big Data in the telecoms vertical – in the medium term at least – is as likely to be positioned as an incremental add-on to other information management and BI initiatives rather than a stand-alone implementation that has its own clear business case. Apart from anything else, Big Data can be expensive and there is a much that can be achieved with existing relational database management systems and more traditional approaches. For the foreseeable future operators are likely to continue to need tools that can query both Hadoop and other categories of data and information systems. It’s highly unlikely that anyone is going to throw everything away and replace it with a ‘Hadoop cluster’.

Volume, velocity and variety – the three Vs that are the drivers of Big Data – are clearly stumbling blocks to making effective real-time use of the vast quantities of data swilling around a typical operator’s business, but there is as much urgency to introducing some coherence to the maddening variety of data sources currently available to operators.

Tools that can analyse and add value to vast quantities of data are clearly going to create benefits for operators but it could be argued that data consolidation and the creation of something approximating a single data system – and hence a ‘single version of the truth’ – is at this stage probably just as important for operators to get to grips with.

The elephant in the room – privacy and trust

It’s inevitable that at a Europe-based conference, regulatory issues around the use of customer data figured highly, especially as two of the speakers were Dutch and the Netherlands has particularly stringent rules in this area. But if anything, regulatory issues played second fiddle to the whole issue of customer trust.

During one of the conference panel sessions focusing on this topic there was a majority opinion that operators needed at all times to be upfront about what they were doing with customer data and specific customer opt-ins were required. Burying what you are doing in the Terms & Conditions might be sufficient from a legal point of view but seriously risks losing customer trust.

Interestingly the US operator on the panel took a contrary view largely because the legal situation in the US is different to Europe, with operators currently interpreting the law as requiring only an opt-out, although of course that could potentially change if challenged.

The Carrier IQ incident last year – where AT&T, Sprint, T-Mobile USA and Apple became involved in a legal case with mobile phone customers – shows that it is necessary to actively engage with customers and win over public opinion on privacy issues rather than just rely on legal argument. The risks to customer trust are just too great otherwise.


Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.


Do you agree public funding should be used to support mobile operators to more broadly deploy Open RAN?

Loading ... Loading ...