Globys uses big data analytics as the underpinning for its offerings for mobile operators, particularly in terms of marketing. RCR Wireless News posed a
series of questions on big data trends to Globys’ CEO Derek Edwards, with an eye toward what value big data brings to the telecom industry, as well as his perspective on overarching trends and the impact that big data will have on the bottom line.
RCR: What major trends will drive telecom analytics in 2015?
Edwards: Technologies that can handle big data in real time and allow for a granular view of the customer will take center stage in 2015. Operators need a more comprehensive, current and actionable view of the customer, and it’s these next-generation technologies that will enable the shift from point-in-time analytics and reporting to real-time decisions and actions – in a sophisticated manner and at scale.
This move toward “taking the right actions” based on better understanding individual customers and their behavior will be driven by operators facing increased pressure to derive long-term business value from their existing customers. As we’ve seen in other industries, this need to quickly and cost-effectively increase the rate and scale at which data is acted upon to solve specific problems will increase the uptake of the “as-a-service” model. Operators continue to focus on the measurable return on investment of their big data investments. In 2015, we’ll see these investments help move the needle for operators in the areas of revenue, retention and customer experience.
RCR: What use cases will dominate BDA applications for telecom?
Edwards: We’re seeing a lot of operators invest in big data analytics to help them better market their service. Specifically, they want the expertise to better understand individual customer behavior and then tailor interactions in ways that drive specific outcomes.
The use cases of most significant interest will be those that drive improvements around revenue and retention. We’re already seeing the application of big data analytics to the areas of network and customer support. Moving forward, we’ll see big data analytics driving more marketing and customer experience initiatives: Marketers will turn to big data analytics to help drive more personalized, real-time customer engagement. This capability presents considerable opportunity for operators to proactively intervene and influence customers’ behaviors and decisions at the time it matters. Ultimately, operators will prioritize use cases according to the direct impact they have on business key performance indicators such as revenue, retention, satisfaction, etc.
RCR: Are there still technical challenges to be met with BDA? If so, what are they and how is the industry and/or Globys working on them? If not, how has the industry addressed them?
Edwards: With big data analytics, operators still face the technical challenges around data latency. Marketers in particular have to be able to get data in near real time if they want to pursue timely analysis, insights and action. The technical challenge is in getting as close as possible to the actual data sources, especially when data warehouses are in place and workflows for getting data into them already exist. Typically data reaches the data warehouse with some delay, sometimes a couple of days, which limits the ability to execute contextual communications and offers.
Getting access to raw, unfiltered data that hasn’t already had a business process run on it is another technical challenge operators face. The goal is to avoid getting data that has been aggregated or has been altered from its raw form in any way. Marketers want the rich detail, not data that’s been averaged or summarized into a single line item. They want to know, for example, the date, time, place and frequency of the dropped calls that occurred in a given month for an individual customer, not just the number of events.
What’s exciting is that big data analytics technology is making this cost-effective. Massive amounts of data can be gathered closer to the source and stored at a reasonable cost, making granular customer views and real-time analysis and decisions possible. APIs are also opening up that allow for data to be accessed sooner and used in ways other than through the traditional warehouse, which presents new opportunities.
RCR: What value do you see telecom analytics bringing to the mobile operator – in business intelligence, marketing, network optimization, etc. – and what do you see as the translation for the bottom line?
Edwards: Big data analytics offers mobile operators a new way to solve the chronic problems of churn and declining ARPUs. What’s different about the sophistication of the analytics available today is the level of granularity at which behaviors can be analyzed, the rate at which learnings can be applied and scaled, and the accuracy at which sustainable success can be measured.
Every operator knows that when customers feel “known,” they stay longer and spend more. What they don’t know is how in a systematic fashion to make them feel known. What this capability means for operators is significant in terms of revenue and retention benefit.
RCR: Mobile operators generate huge amounts of data, and not all of it is useful or actionable. Sometimes the most valuable information only occurs a very small percent of the time (such as with, say, churn). What is Globys’ approach to sifting out the most important data and drawing from it actionable intelligence?
Edwards: One of the biggest misperceptions around big data analytics is that more data, more insights and more results generate better outcomes. In reality, it’s not about having the most data but about having the right data to generate actionable insights that drive better results. To put it in perspective, think about 20% of a base with a $1 lift in ARPU vs. 2% of a base with a $10 lift. Both result in the same net gain, but the data (volumes and varieties), marketing strategies and offers required to achieve the outcome would be completely different.
At Globys, our approach is to determine the most relevant data for a specific use case but not by using human intuition or assumptions. Our approach starts with unsupervised data exploration using big data technology, such as Hadoop, which enables the analysis of all available data to determine behavioral patterns for each customer. This understanding allows us to determine which behaviors are attributed to specific outcomes, such as churn, lower than normal usage, social sharing, handset upgrades, or adoption of a new product. Focusing on those identified behaviors, our system leverages robust experimentation to test messages and offers across customers, contexts and channels to determine what is most effective. Then machine learning automatically determines what works and what doesn’t, driving dynamic optimization that maximizes performance results over time.
Follow me on Twitter: @khillrcr
Image Copyright: deniskot /123RF Stock Photo
