Big Data and Predictive Modeling
With the advent of the Internet came the ability to gather large amounts of information on consumers’ opinions, shopping and purchase behaviors, satisfaction and loyalty. Coupled with the development of sophisticated data warehousing capabilities where companies collect and store detailed transaction data, many prognosticators have announced that we have entered the world of “Big Data” and Predictive Modeling.
Because “big data” is largely unstructured, the information contained within it is neither readily apparent or readily accessible. Turning such data into information has been a challenge for market researchers.
Morpace has been finding information within “big data” for its clients for quite some time. Sometimes it has entailed merging survey data with corporate data on transactions, sales or revenue. Sometimes it has entailed coupling multiple data bases from multiple sources.
Some recent examples:
- Revenue and profitability for multiple big box retail outlets modeled as a function of customer satisfaction.
- Customer retention and profitability for retail bank customers modeled as a function of customer satisfaction.
- Automotive dealership customer loyalty and defection modeled as a function of dealership management, facility and staff training standards performance.
- New vehicle market success metrics (e.g., brand image, online inquiries, share of segment sales) modeled as a function of central location product clinic evaluations.