BFSI

Case 1

Client: One of the largest commercial credit bureaus in the world

Opportunity: Improve the matching on commercial credit to 100% and collect revenue for all answers

Solution:

  • Segment existing response rate and revenue coverage on the queries
  • Analyze the factors due to which revenue is not earned including sub-optimal matching algorithm, data not available, data can’t be productized and external data sources not used
  • Simulated profitability if external data sources are used based on richness of data
  • Analyzed the correlation between data quality, buy rate and pricing
  • Defined control and test group of customers to understand walkaways and ways to prevent it

Result:

  • Improved match rate to 100% driving $48 MM annualized revenue

 

Case 2

Client: One of the largest commercial credit bureaus in the world

Opportunity: Analyze market potential for a new product line; create a plan to double revenue in 3 years

Solution: 

  • Analyze renewal, upsell, cross-sell and attrition data to develop predictors for these behaviors
  • Define the product segments aligned to small businesses (core targets) needs
  • Segment the customer data using decision tree classifier models
  • Test the models with un-seen data and further improve their precision

Result:

  • Developed the plan to double the revenue using the models in 2.5 years

 

Case 3

Client: One of the largest credit-card networks globally

Opportunity: Build cutting edge product recommendation algorithm for the premium card portfolio

Solution: 

  • Focus on restaurants, the sizeable discretionary & elastic spend
  • Analyze hundreds of millions of transactions to understand customer behavior individually and as part of social networks spanning merchant establishments
  • Build powerful meta-data for restaurants in the network by reinterpreting existing data and integrating Zagat data
  • Build an algorithm taking into account cuisine preferences, geographic footprint and spending preferences of a customer

Result:

  • Algorithm patented by the client
  • Algorithm predicted with as much as 50% accuracy where a person will dine or lunch in the near future leading to a range of marketing opportunities

 

Case 4

Client: One of the largest credit-card networks globally

Opportunity: Improve response rate to email marketing campaigns

Solution: 

  • Analyze customer profile and email response behavior to form preliminary hypotheses based on historical data
  • Craft integrated tests encompassing key hypothesis for each element of the campaign – population selection, timing & creative – to generate experimental insights

Result: 

  • 20% increase in response rate to emails