Most people would probably agree that weather influences our lives in manifold ways. But how about our shopping habits? It is not unusual to read in the newspaper about the weather as a major influence factor of retail sales. In fact, some retailers such as British TESCO have already started to analyze weather in order to improve their operations planning and inventory management.

While the assumption that weather impacts retail sales seems plausible, empirical evidence from academic papers is scarce. There are some studies that investigate the effects of weather on classical bricks-and-mortar retailing, but until now there were no studies that analyze the weather-effect in an e-commerce setting

Collaboration with a large German fashion online retailer

In a recently published study the authors Steinker, Hoberg and Thonemann from the Department of Supply Chain Management & Management from the University of Cologne and the Department of Supply Chain and Operations Strategy from Kühne Logistics University in Hamburg address this gap in academic research. In collaboration with a large German fashion online retailer the authors analyze how the weather affects online demand over a period of two and a half years.
After a thorough discussion on how weather might relate to online demand via direct or indirect effects, the authors take a deep dive into an empirical analysis.

Their findings support the idea that weather has a strong influence on consumer shopping behavior and online-demand. Interestingly, the weather effect seems to be “asymmetric” according the study’s findings: While good weather has a strong negative effect on demand (i.e., fewer sales with good weather) the opposite effect (i.e., more sales with bad weather) seems to be much weaker.  The findings also suggest that the weather effect is more pronounced on weekends compared to work-days and also stronger during the summer months compared to the cold-season.

Large cost savings possible

To quantify the potential cost-savings from using weather data in operational planning, the authors analyzed by how much sales forecasts could be improved when weather data was used in statistical forecasting models. Compared to the baseline specification of the time-series models where no weather data was used, the forecast models that used additional weather variables performed significantly better and reductions in forecast errors of up to 50% were observed. A subsequent backtesting-analysis for one of the company’s warehouses revealed that cost savings of several hundred thousand euros could be achievable by including weather data in the workforce planning.

Implications for practice

The study provides guidelines on how to augment operational planning with weather data. In particular the benefits for increased forecasting accuracy and the implications for workforce planning are highlighted. High forecasting accuracy is a competitive advantage in e-commerce operations since it enables high responsiveness and short-delivery times at lower costs. The findings of the study also indicate that weather data could be successfully applied to other use cases such as search engine advertising (SEA), dynamic pricing or to increase the effectiveness of marketing campaigns.

Publication:

The Value of Weather Information for E-Commerce Operations (2017), S. Steinker, K. Hoberg, U. Thonemann, Production and Operations Management (Early View Articles - 29 June 2017), DOI: 10.1111/poms.12721