Big Data in India: A CxO Perspective

A growing level of digitization means that data is now being churned out at an even more frantic pace than before. With India’s scale and diversity of data, it has been a highly challenging task for CIOs to create the foundation for intelligent analysis of data. To get a clearer perspective on the state of big data in India, Dataquest spoke to leading CxOs from different verticals to gauge the usage of big data in their respective organizations and how it has made a difference to the competitiveness. A first-person account of spokespersons from different organizations follows

With a strong network of 400+ branches and being a consumer lending enterprise focused on financing the underserved, we rely on data-driven decision making, which in turn, is the bedrock to our customer segmentation and acquisition strategy. We have information of over 1 mn customers, which provides us an immense opportunity to run statistical models for a responsible financial growth of the franchise.

For any meaningful analysis, it is important that we get accurate customer information upfront so that we can apply decision science rules for quicker and seamless customer acquisition. The core of business intelligence is the transformation of data into insight and insight into action that can add value to the enterprise.

Over the last three years, we have invested extensively in analytics and technology. We believe analytics is critical for businesses to make better high-volume decisions to achieve breakthrough business performance; deeply understand and engage customers to earn and inspire their loyalty; and change the game and drive competitive advantage through big data and technology.

The industry is witnessing an accelerated growth in the amount and variety of useful information, thanks to growth in digital, increasing mobile penetration, hyperconnectivity and related trends. When we combine these data sources with information from our high-quality decision science models, we get the best of both worlds— high quality, highly granular, highly descriptive information that when used well, creates big opportunities to improve the effectiveness and efficiency of our marketing, sales, and innovation activities. The dynamic nature of our business requires decision sciences, an interdisciplinary approach of business, applied math, technology, design thinking, and behavioural sciences to solve constantly shifting and evolving business needs. This largely summarizes our need for quality analytics.

How analytics gives us a competitive advantage

There are seven inter-connected steps, which explain how the usage of analytics has given us an advantage. Firstly, data capture is the key. Customer information is made accessible through a data warehouse platform. Secondly, historic customer information is used for segmentation, and in turn, create appropriate customer acquisition strategy. But to move beyond segmentation and customization requires a more complex data-rich and event-driven analysis. Thirdly, many customer insights are intuitively obvious. One can predict the behavior pattern of top 5% but the remaining 95% need extensive modelling and robust analytics capability.

Fourthly, once the data streaming happens in real-time, we anticipate opportunities to board the right set of customers. Fifthly, the power of real-time data analysis is its ability to enable real-time decisions. We need to differentiate between our existing borrowers and new borrowers and need to inculcate the culture to promote customer loyalty. Sixthly, we have automated information management so that our frontline staff can take not just faster but accurate decisions for a unified and delightful customer experience. Lastly, we measure and evaluate the efficacy of our processes on customer acquisition. This continuous, closed-loop process incorporates strategic and operational intelligence and in doing so enables granular analysis, refines event-driven models, and generates new insights and opportunities.

Challenges

While big data has increased the opportunities available to businesses, it also creates more challenges to capturing, storing, and accessing information. Utility of some of the data points declines rapidly which means that we need to churn data into information very quickly. In India, there is a challenge of getting relevant information (like age, income levels) especially in the rural markets, but the heartening news is that this is soon changing given the penetration of banking and unique identification number in most of the markets that we operate in.

Our organizational data size of the transactional systems is about 14 TB. Mostly, this covers the structured data. Complexity includes cross-functional, multi-geographical, distributed and diversified people, business processes, infrastructure, and enterprise systems. We do structured data analytics with the data at the employee and project level. It gets aggregated at the customer, vertical, horizontal, geo, business unit, and the project type level. The primary focus of our efforts in analytics are in the areas of customer centricity and delight, employee passion, and delivering innovative services. Due to analytics, we can view the organizational dashboard of various metrics. This helps us in time management, tracking insights, and reporting accurately to stakeholders for quick actions and decision making. Better reporting and decision making enables excellence in sales, delivery, and operational areas. Today, we can present monthly dashboards for various levels of management. Thanks to analytics, we also have more structured and data driven management and reviews.

At Vodafone India, we have a mature BI and analytics ecosystem which comprises of an Enterprise Data Warehouse (EDW) platform, an integrated analytics platform and we have recently set up an extension of a big data platform to analyze real time, high volume unstructured data. The volume of data available in these platforms put together is in access of 500+ terabytes.

The EDW platform has a 3NF (Third Normal Form) data model which holds the atomic business data, whereas the analytics platform stores aggregated data in dimensional models in multiple data marts each catering to a different line of business like marketing, sales, customer service, finance, etc. Around 1,300+ business users from different business functions across 23 locations pan-India access the EDW and data marts on a daily basis.

Top priorities

Being a telecom firm, we like all decisions to be based on facts rather than guess work, hence, analytics is one of the top priorities for us. Based on business needs, the following key areas have been identified as a part of analytics roadmap for our organization:

Self-Service BI: With this, business users will be empowered to create BI reports as per their requirements. This has already been rolled out to the key user segments.

CxO Dashboards on Handhelds: Business KPI dashboards will be made available to Cx Os over the handheld devices.

Big Data: The big data pilot project showed us that there is enormous value in tapping the dark data available in our organization and also outside. We have plans to take this initiative forward and include more use cases.

Network Analytics: Data available in our network assets holds tremendous amount of value and we are converting the information available in this humongous data into insights for business use.

Prescriptive Analytics: We are also working towards developing a prescriptive analytics framework. With this, we will be able to better anticipate our customer needs and cater to them.

The business case for big data

We have introduced the big data platform as an extension to our analytics environment and enabled a few use cases for our business users including: (1) Real-time triggers for gauging the propensity to international travel, first time OTT (Over the Top) applications like WhatsApp, Facebook, etc, have helped in targeting customers on a near real time with right offers. This has helped in increasing campaign conversions substantially and hence revenue. (2) Social network analytics has helped in predicting customers churn in advance based on their behavior patterns. This has enabled us to control churn in good measures. (3) Integration of unstructured data, ie, social media data has helped in generating more insights about customers, ie, preferences, hobbies, and profession which have helped in creating the right products for customers. Integration has enabled to understand customer’s sentiments which was not available earlier.

After we complemented out traditional analytics platform with the big data platform, we could do much more than ever before, including: (a) Quick identification of propensity of a customer to take certain decisions, and on a near real time provide tailored solutions to meet the customer needs. (b) Identify users who are having issues with a service and proactively address that. (c) Understand social media chatter related to our service to address any large-scale concerns and enable social media as a customer service channel.

Being in the logistics industry, it is very important for us to get close to the data as this helps us in getting the actual picture of the situation prevailing in the market. The quantum of data that is being handled is relatively not so large but the complexity comes into the picture when the data comes from multiple platforms and is of real time in nature. It gets very complex to verify and validate data of such nature.

Analytics supplements the organizational efforts in many ways and for a logistic firm, keeping track of all important domains of business is possible only through analytics. It supplies us with data that reduces the ideal time for vehicles. We can also plan our route details to increase efficiency of our fleet in terms of loading, delivery, and transit time.

Analytics plays a big role in realization of achieving optimum customer satisfaction by supplying us with all the necessary details that are used in the formulation of customer-centric strategies and to control any deviation. This makes sure that we deliver what we want to and with the desired level of excellence.

We have developed customized tools to manage our data made on Java and other platforms. Analytics provides us with a 360 degree view from the perspective of all the departments and helps us improve internal as well as external functions.

After the deployment of the solution, there have been many constructive changes as now the decision making has become far easier. Proactive decisions are being made with the help of data supplied and forecasting can be done based on reliable supplies. Planning is getting facilitated at every stage and is making our commitments even easier to achieve.

Insurance being an extremely data-intensive business, depends on huge amount of data and it also generates a huge amount of data in course of its operations. Both these types of data are stored in structured fashion in the way that helps stakeholders get unique insights and helps in the overall improvement in client service and compliance with regulation.

In their course of routine business, insurance companies collect massive amount of data—both structured as well as unstructured. This is the data that insurance companies generate through transactions (digital and textual) and in varied degrees from external world, be it government agencies, data bureaus, or social media. In various countries, there are several agencies, government and non-government, which make data available to the companies in a lawful manner, though in India sources are limited. With increasing complexity of business and fierce competition, success of any organization depends heavily on its ability to process this data effectively, analyze the same, and use in real-time decision making.

This data is used in building complex statistical models for pricing, understanding customer demographics and behavior, thereby sharpening campaigns, cross sell/upsell, fraud analytics, customer service, etc. With the increasing influence of social media, customers are freely talking about product, brand, companies day in and day out, both negative and positive. This generates wealth of data in an unstructured format. It is complex to analyze unstructured data and draw meaningful influences, but insurance companies are heavily investing in technologies that help collect this data, analyze it in real-time to understand consumers sentiments, and take appropriate actions. Some companies are crowdsourcing ideas from these channels and designing new products.

Insurance companies are also using a technology called telematics, where a device is installed in insured vehicles and data is collected about the way a customer drives. This vast amount of data is used in arriving at customized pricing for the customer. This is a unique concept, used by various companies in several countries and gaining popularity, where a good driver is rewarded with lower premium, whereas a not-so-good driver pays a relatively higher premium. In this process, the device installed in the vehicle continuously sends the data to insurance companies’ servers and complex algorithms analyze the driving patterns. Similar monitoring of data, ie, sensor data offers huge opportunity for business, like recommendation to prevent larger damages, better planning of resources (claim handling), early detection, and underwriting excellence, etc.