Big data can deliver substantial business value, but only if it breaks out of organizational silos. That was one of the key conclusions from a panel of experts discussing big data and customer engagement at the TM Forum Live in Nice in June.
Featuring experts from leading communications service providers (CSPs), the panel identified numerous use cases for big data. For example, Dr. Susan Wegner, VP Internet & Services, who heads the Deutsche Telekom group wide Big Data Competence Team, described how her company is running a proof of concept in Hungary where it is making personalized offers to (opted-in) consumers when they are in the vicinity of relevant shops. She also described how big data can be used to reduce credit card fraud. For example, if your phone is in Nice and there is a payment from your credit card in Paris, then there is probably something wrong. A mobile operator could provide real-time location data to the credit card company.
However, using big data in this way requires careful planning, given the potential privacy implications. Dr. Wegner stressed the importance of getting representatives from different functions to work together to oversee the use of big data and evaluate potential use cases. She noted that valuable data can be dispersed throughout an organization, spanning multiple departments and Hadoop clusters.
At the same time, a CSP needs to avoid being held back by organizational inertia. Geoffrey Zbinden, VP Big Data & Business Intelligence, Orange Group, stressed the importance of high-level sponsorship to counter the “resistance to change” that can arise in large operators. “Big data can dramatically improve efficiency…it can reduce incoming calls and increase customer satisfaction,” he said. “We have the opportunity to break the silos and to merge all the information, not only about customer usage, but also about interaction with customer care in shops, on the web and through portals.”
Orange is in “test and learn mode”, exploring how big data can be use to improve customer care, marketing and other internal processes. For example, it is using big data to detect when a set top box has been damaged by lightning during a storm, so Orange can proactively fix the problem rather than reacting to an incoming call from the customer.
“Orange has identified more than 100 operational use cases, but not all will generate value and maybe 20% will generate 80% of the EBITDA impact,” Mr. Zbinden said. “Clearly you have to put in place a pragmatic approach.” The use cases that yield the most business value will depend on the maturity of the market, the competition and other local factors.
Making structural changes
One way to break big data out of silos is to give IT staff clear business objectives. Ahmed Saady Yaamin, VP, Business IT, Robi Axiata, described how his company has embedded IT staff into a marketing analytics team, to ensure they focus on business outcomes. With pricing and ARPU so low in Bangladesh, Robi Axiata is focused on using big data to provide a differentiated customer experience that will cut churn and increase customer lifetime value.
“In order to make the maximum leverage from big data, you need to solve the challenge of organizational structure,” Mr. Yaamin told the Forum. “We have been talking about the gap between business and IT for so long… it is very important to align these two teams to make sure big data is in action.” In particular, CSPs need to ensure that the team implementing big data is also responsible for the business outcome. “We need to have an end-to-end responsibility and that needs to be reflected in the KPI of the team,” Mr. Yaamin added.
Creating a common language
Historically, valuable, but often overlapping, information has been trapped in standalone data warehouses making it difficult to leverage and expensive to manage. If the industry repeats these mistakes, “we will have not only big data, but a big problem,” warned Gadi Solotorevsky, CTO of cVidya.
Mr. Solotorevsky described how the TM Forum is looking to create a common language, a common blueprint, that will give everyone a common understanding of what is big data, what the building blocks are, what data is needed for what scenario and how this data can be shared between different scenarios. “There is a big data team that is making these definitions,” he explained. “There is a catalyst project in which we are taking these definitions and deploying them on 15 different use cases built by five different companies, to reuse information, to reuse building blocks.”
For me, the key takeaway from the panel discussion is that CSPs are now tentatively taking the steps that will be needed to really harness the potential of big data. But there is a long way to go: Our industry is embarking on an exciting, and potentially rewarding, journey.
Click here to watch the full panel video.