The legacy 'cash cow' of voice services is disappearing quickly, but new business models are emerging and the data analytics gold rush will reward the innovators that act quickly and deliver compelling business cases.
The term 'big data' is associated with capturing and analysing consumer behaviour using the Web. The velocity and volume of digital transactions captured and processed to achieve business value require changes in the software systems to support petabytes of structured and unstructured data. These systems must be distributed, delivered in a cloud-computing environment and need to support in-memory database management. The data can be structured or unstructured. Structured data includes HLR records, IP application flows, geo-positioning and billing detail records. Unstructured data includes tweets, Facebook posts and any other form of social content.
Revenue generation and data monetisation are driving the market for big data. The vast amount of information that operators hold about their subscribers has been largely untapped, but could be harvested to answer key questions and predict future scenarios (see Figure 1). Smartphones, the availability of broadband and maturity of ecommerce enable more cost-effective advertising and big data will play a pivotal role in understanding buying behaviour and spurring customer demand. However, big data projects will require real-time data capture and analysis. CSPs can profit from big data internally and sell it to advertisers. Telefónica's Smart Steps service takes this approach: it captures, processes and then sells the anonymous data of its customers to retailers and public-sector organisations.
Figure 1: Harvesting real-time network data to act now and predict future scenarios [Source: Analysys Mason, 2012]
Our research has found immediate benefits of big data projects that can best be articulated in use case studies that we have published in our Telecoms Software Strategies programme. Each use case study is relevant to specific business objectives with clear paybacks. Rotational churn in a prepaid market uses data analytics to understand usage patterns, social network connections and business intelligence to predict churn. This predictive analysis can be used to proactively connect or promote loyalty offers before the customer is lost. Other revenue-enhancement strategies focus on yield optimisation. CSPs can measure load by cell site and target high-value users with promotional offers that encourage network usage. This generates incremental revenue without affecting service quality and uses investment in sunk network cost. In the M2M market, big data can be used to identify crowd movements and traffic jams using location-based data. This data with analysis can be sold as a value-add application to redirect customers to storefronts or seek alternate routes. M2M instrumentation in the transport and automotive industry tracks assets and environmental conditions, and supports remote surveillance. Data streams can be processed quickly to alert a parent to the erratic driving of their teenage child, for example, or to interact with electronic billboards to display ads based on the demographics of drivers passing at specific points in time.
The following are the three greatest inhibitors to realising the benefits of big data.
- Multinational CSPs operate as independent entities, which restricts their ability to unify marketing campaigns, standardise product sets, and rapidly launch new services.
- Security and privacy concerns could spur regulators to enact tougher regulation and increase churn. (Telefónica recently withdrew support for its Smart Steps service in Germany because of consumer privacy concerns.)
- Technology-driven projects that do not assess the business requirements and do not deliver actionable reports.
We recommend that CSPs embarking on a big data analytics project identify a specific business objective that is well defined in scope. CSPs should be able to classify the project either as revenue-generating or productivity-enhancing with a clear cost-benefit business case to back it up. M2M is revenue-generating, whereas network optimisation is focused on capital efficiency. Be aware of privacy issues that could spur regulatory investigations. Understand how unstructured data will be used, which accounts for a large percentage of the value of big data analytics. And finally, use the supplier community where 80% of the solution can be achieved quickly, and customise the remainder.