(When) Can Social Media Buzz Data Replace Traditional Surveys for Sales Forecasting?

Abstract

Producing reliable sales forecasts for new products is notoriously difficult. Traditional surveys have been popular for decades but they are relatively cumbersome and expensive to implement. Use of readily available data from social networks is becoming increasingly popular. Can the customer buzz measures obtained from social networks be a convenient and inexpensive substitute of traditional tools? Using movie industry as a backdrop, the authors compare forecasts derived from a large-scale purchase intention survey with those obtained via social networks. They find that while the survey approach is more reliable in predicting performance overall, customer buzz-based forecasts outperform surveys under conditions of high uncertainly, e.g., for niche and low-budget movies.

Keywords: big data, social media, marketing, marketing research