m
Our Mission Statement
This is Photoshop's version of Loremer Ipsn gravida nibh vel velit auctoregorie sam alquet.Aenean sollicitudin, lorem quis bibendum auci elit consequat ipsutis sem nibh id elit.
Follow Us
Top
Why Telcos Need a Real-Time Analytics Strategy - VoltDB
9376
post-template-default,single,single-post,postid-9376,single-format-standard,mkd-core-1.0,highrise-ver-1.0,,mkd-smooth-page-transitions,mkd-ajax,mkd-grid-1300,mkd-blog-installed,mkd-header-standard,mkd-sticky-header-on-scroll-up,mkd-default-mobile-header,mkd-sticky-up-mobile-header,mkd-dropdown-slide-from-bottom,mkd-dark-header,mkd-header-style-on-scroll,mkd-full-width-wide-menu,mkd-header-standard-in-grid-shadow-disable,mkd-search-dropdown,mkd-side-menu-slide-from-right,wpb-js-composer js-comp-ver-5.1.1,vc_responsive
VoltDB / No Compromises  / Why Telcos Need a Real-Time Analytics Strategy

Blog

Why Telcos Need a Real-Time Analytics Strategy

Historically, telco analytics have been limited and difficult. Telco networks and the systems that support those networks are some of the most advanced technology solutions in existence. The performance requirements for telco networks are staggering – the ability to process millions of transactions per second, the ability to manage billions of events per day, and the ability to manage tens of millions of devices simultaneously – and to do so without any perceptible downtime.

Handling the data generated by these networks is a herculean task. Telco networks are unique, and their success can be defined in two parts:

  • Being able to successfully process volumes of data off network elements and systems without losing any information;
  • Being able to render an accurate bill from this information and create revenue from the services provided.

Analytics and insights have always taken a back seat to the first two priorities – accurate data processing and billing. In the telco world, batch processing has been the de facto choice for data processing and billing, so any analytics strategy had to be based on a batch processing model.

The types of analytics a batch processing system can provide are generally limited to generic insights about the numbers and volumes of subscribers, devices, and applications, across large time windows. Examples include:

  • Average number of subscribers using the network per day;
  • Trends and changes to the numbers of subscribers seen in a network per day or per week;
  • Average number of a certain type of device in the network; and,
  • What’s the “busy hour” in a network for usage?

Telco companies used this information and insights to plan capital investment projects for network expansion and marketing.  While these insights are valuable, an entire class of insights even more valuable to operators has been left on the table –real-time insights.

Watch the Video of our Telco Expert

There is a class of real-time data that has not been used to benefit telco insights

As telco networks mature, the trend lines point squarely to two things:

  • There’s exponentially more information, and
  • Information needs to be handled in real-time, not via batch processing.

There are several reasons for this – the evolution of 4G to 5G, IoT adoption, the proliferation of devices, and more – but the need for new revenue models and network controls is equally pressing. New revenue opportunities include pre-paid services, integrated subscriber policy management, and other scenarios.  The bottom line is this – all the data is generated in real-time and telco operators are having to process it in real-time.  Does this affect our analytics strategy?

The answer: Absolutely!

Using the example of a mobile network, let’s look at the kinds of data that exist for a subscriber if I am an operator:

  • I know where exactly you are using your device, and exactly where you came from beforehand;
  • I know the time of day, and how often you use your device in this location;
  • I know what application you’re using – email, Skype, Facebook, etc.;
  • I know your usage profile – how much usage you have left in your plan, how much money is left in your pre-paid account; and,
  • I can probably infer where you’re likely to be during working hours versus non-working hours, based on your history.

None of this should be a surprise. If you have a mobile phone, you’re broadcasting your life story 24/7. But think of the richness of data – Steve Jobs might say data and insights like this are “insanely great!”  So, what can we do with such incredibly rich data?

Change the telco game from the ground up.

Real-time data is a key way to improve QoS and CEM

By interacting with rich, real-time data as it streams through the network, we invite a sea-change in how telco operators interact with, retain, grow, and improve their customer relationships. Extending the example of the mobile network above, consider the strategic advantages and insights a telco operator might gain with a real-time analytics strategy:

  • I can detect declining Quality of Service (QoS) in the network, and address the situation before the caller puts the phone back in their pocket;
  • I can adjust network traffic in real-time to fix any performance perception issues (e.g., downgrading the streaming rate on video). For passive services like email, this wouldn’t be as necessary. Will a subscriber really know that an email that arrives five seconds late is “late”? Probably not.
  • I can make real-time offers to upgrade service, or extend contract terms to subscribers, while they are a captive audience;
  • I can gain insights into “soft errors” in the network. For example, if I see a subscriber calling the same phone number three times in 60 seconds, a network engineer might say that’s three successful calls. But a Customer Retention Director would probably speculate that’s because the first two calls were garbled, and the customer might be pretty incensed with my brand.
  • I can offer real-time messaging to customers who are having a bad network experience to alert them, “We know! We’re working on it!”, showing pro-activity and increasing customer retention.

Real-time data is incredibly useful – but only in the moment – after that subscribers have left

There is no substitute for real-time analytics and action. In today’s competitive landscape, telco operators are under staggering pressure to retain and grow customers. Phones are being sold unlocked. OTT providers offer many ways around the telco billing infrastructure. Insights at the end of the month are no longer good enough, especially if a subscriber has had a few weeks of bad experiences.  They’re likely looking – or gone – by the time that end of the month report gets generated.

Data lakes are useful for some trends – but not real-time insights

Nothing about this trend towards real-time analytics takes anything away from the traditional role of analytics in telco networks. All the insights described above still apply:

  • Average number of subscribers in the network
  • Predicted usage for each type of device
  • Capacity trends based on historical usage
  • etc.

But let’s not ignore the immeasurable benefit from interacting with those real-time data streams. They drive a new and exciting opportunity to interact with, grow, and retain subscribers – in ways a traditional analytics strategy cannot.

There’s only one way to get that data, and it isn’t with an OLAP solution. You need a real-time OLTP solution

Telco Operators often have huge investments in data warehouses to do analytics. These systems drive deep insights into network engineering and planning, capacity management, customer retention, and other key metrics. None of the trends towards real-time data will change that.

Traditionally, Telco operators had a choice to do their analytics, either

  • To process the network today, and analyze it tomorrow, because data volumes are too overwhelming, or
  • To analyze the network data for very specific information – but you couldn’t apply that analysis over the huge volume of data.

Either way, there’s a compromise.

Most operators have chosen the first approach, to retain the data and deal with it tomorrow.

The Achilles heel of these solutions is often that they are based on OLAP (On Line Analytical Processing) technologies, for example NoSQL solutions. OLAP solutions are excellent choices for doing deep-field, long range analyses of data, perhaps empowering a team of data scientists who are looking for insights in terabytes of stored data. But often those valuable insights are stale by the time they’re discovered.

Real-time analytics can only come from one type of solution, one that lives in the data path of the telco network – and it must be OLTP (On Line Transactional Processing) technology.  Each piece of data can be handled and analyzed in real time – Telco real-time.

Batching and micro-batching will get you faux-real-time. It isn’t good enough when your decisions depend not just on the events that you process, but also on the order of arrival of those events, or, perhaps more importantly, on the events that do not even arrive. These are the events that are critical to identifying issues in QoS and other insights, and the only system that can help you there is a per-event processing OLTP system. Being able to process events in a per-event fashion in the network, in real-time is the only way for Telcos to truly leverage the potential value in the data.

No need to compromise

Telco Operators finally have a “No Compromises” solution to their real-time analytics challenges.

This game-changing approach to handling real-time data and enabling new analytics and insights is the reason dozens of the world’s leading telco providers are powering their networks with VoltDB technology, across hundreds of millions of subscribers every day. Try it yourself – download VoltDB today. Or, learn more when you visit our telco strategy page.

Watch the Video of our Telco Expert