Business forecasting with online buzz

Schaer, Oliver and Kourentzes, Nikos and Fildes, Robert (2019) Business forecasting with online buzz. PhD thesis, Lancaster University.

Text (2019SchaerPhD)
2019SchaerPhD.pdf - Published Version
Restricted to Repository staff only until 31 December 2024.
Available under License Creative Commons Attribution-NonCommercial-NoDerivs.
Download (2MB)

Abstract

In our fast-paced business world with changing consumer preferences, the demand for product life-cycles have shortened and become more volatile. This is the case for both own and competitors' products. To cope with those challenges, practitioners and researchers focus their attention on augmenting forecasting models by incorporating additional information such as online buzz. Online buzz reflects discussions around corporate and user-generated online content, such as social media interactions, forum discussions and online search. A majority of studies report forecast accuracy gains from its inclusion. However, on closer examination, many of these studies exhibit design weaknesses including lack of adequate benchmarking or rigorous evaluation, questioning their reported estimates. This thesis focuses on three research questions: (i)~What is the predictive value of post-release buzz and usability for operational decision making? (ii)~Can pre-release buzz help to estimate the market potential for sequential released new products? (iii)~How suitable is pre-release buzz to predict competitors' new product success? Our findings are mixed, by demonstrating both its potential, but also limitations of its usefulness. We conclude that online buzz is beneficial during the pre-release case, but not for the post-release phase. In the latter case, online buzz has little if any predictive signal, when considering realistic business lead times, explained by the frequently instant buying decisions and mixed pre- and post-purchase signals. On the other hand, pre-release buzz, which has a clear anticipatory characteristic, enhances new product forecasts. Our proposed framework demonstrates that pre-release buzz is beneficial also in predicting market potential. Moreover, it can be used to construct reliable forecasts for competitors' new products, providing market critical competitive intelligence. The thesis concludes with managerial implications and identifies future research directions stemming from this work and the critical reflection of the often hyped online buzz.