BFSI VISION | Analytics

   
 

V.V.Balaji
Executive Vice President – Technology & Service Delivery, ICICI Prudential Life Insurance Solutions

Expectations, Fears, Challenges in Big Data

Business Drivers:
  • The business driver is to get more wallet share from existing customers.
  • Visibility into the business and predicting the customer behavior
  • Being all time right for customer service- understanding the customer behavior and need
  • Big data can primarily be used for quick, on-the-fly, insights to the field force

Acceptance of analytics:
  • What's in it for me? – question from large section of the company
  • Ad-Hoc data massaging and gut feel behavior
  • Copy paste collaboration of various islands of reporting
  • Change management

Challenges faced in Big data Analytics:
  • First and foremost – big data needs use cases, unless you have a compelling use case, there's no use getting into big data – linkage of big data to business context
  • Big data is not about creating a big data store or repository, you need clear definition of insights you expect from it
  • One must look at using all the structured data exhaustively and extensively before venturing into unstructured data in an analytics way.
  • But this does not stop one from subscribing to analytics providers who crawl through social networks and provide customer insights and ready analytics for company to use.

Fears in Big Data Analytics
  • No fears per se in big data analytics, but one must practice caution to ensure data confidentiality and access controls – data should not reach wrong hands
  • One should also be vary about customers' perception and response on usage of larger public data.

Expectations from Analytics vendor
  • While having a strong base and fundamental stability, one must also be futuristic and capable of moving with the world in terms of exploiting new capabilities

Key success factors in Big data analysis
  • Having valid use case
  • Data security and access control must be well defined
  • Sources of information need to be mapped well and should be relevant to business context