A large financial institution wanted to create highly targeted, effective marketing campaigns to acquire and retain profitable customers. This required the consolidation and analysis of large amounts of information from both traditional and emerging sources of data.
A Teradata Aster platform was employed to extract new insights from multi-structured data by combining data warehouse information with other, non-relational sources such as weblogs, social media sites and clickstream data. It seamlessly analyzed this information using built-in analytic techniques implemented in SQL-MapReduce-a patented approach to marrying set-based SQL logic for analysis with MapReduce-based procedural programming.
Applied to data that the bank had already collected on its customers, such as credit and debit card purchasing details and granular Automated Clearing House transaction data, it could now quickly and easily answer questions it had struggled with previously:
How are customers connected and how do they influence one another?
How can we identify consumers who influence the behavior of their friends and family members and use this information to enhance existing churn models and increase customer retention?
Who is most likely to be a brand champion and key influencer for this organization?
Armed with details on metrics on the influence a consumer's peers had over decisions as varied as product selection and customer churn, precisely targeted offers could be aimed at the most influential customers at exactly the right time, resulting in:
A greater influx of new customers
Better brand loyalty from existing customers
More innovative personalization strategies
Increased overall profits
The mining and analytics of big data sets is not restricted to only the marketing segments. Even high-precision manufacturing companies like the ones involved in semiconductor, and medical manufacturers are witnessing a deluge of data and can take better business decisions by being able to store and analyse this data. The big picture is all about being able to understand the operational and manufacturing issues and relate it to achieve a better customer experience. And this can be clearly achieved by being able to collect, store and analyse all the multi-structured data floating around.
Noel Pettitt is area vice president - South East Asia and India, Teradata
Sign up for Computerworld eNewsletters.