A unique piece of research, which comprehensively analyses the impact of weather on the e-commerce buying habits of high-end fashion consumers, suggests huge savings could be made by more effectively harnessing weather data.
Research indicates short-term buying habits could be better modelled in sales forecasts, and this then could support improved utilisation of resources across the supply chain.
Previous academic analysis on the effect of weather on retail store settings was limited to single retail locations and narrow product categories. Weather impact on online sales remained unclear, and it was felt to be an area ripe for research, especially as the global electronic commerce boom seemingly goes from strength to strength. The aim of The Value of Weather Information for E-Commerce Operations, carried out by the University of Cologne and Kuhne Logistics University, was to quantify the “weather effect” and the aspects of weather that stimulate e-commerce sales.
Given that online sales have multiplied over recent years, discerning shifting demand patterns are becoming increasingly valuable for the supply chain sector. Take China, figures quoted in the South China Morning Post suggest retail sales for mobile e-commerce in China are soon set to reach $1 trillion. Chinese consumers are increasingly sourcing online luxury items and cosmetics products and other items. A similar pattern is emerging in the US, where e-commerce revenues accounted for almost 6% of total retail sales in 2013, according to the US Bureau of Census. In Europe, figures released by The Centre for Retail Research suggested purchasing via e-commerce platforms jumped by 18.4 percent in 2015.
This huge shift in buying habits to online platforms has had significant implications for shops operating from physical locations in towns and cities and forced most major retailers to utilise online channels in conjunction with brick-and-mortar shopping. However, though some goods, such as ebooks, can be delivered online, most merchandise require tangible physical fulfilment through a traditional supply chain.
The development of online sales has brought with it complex challenges for supply chain operators: diversifying product portfolios, smaller order quantities, and high levels of customer service and delivery requirements. According to the Cologne-Kuhne research, a key competitive advantage for any e-commerce platform is the quality of the logistics processes that lay behind the commercial operation.
The planning behind logistics operations can be particularly troublesome: on the one hand, online sales are “volatile and difficult to forecast”, while on the other, logistics capability can be quite rigid and difficult to adjust in the short term. The pick-and-pack workforce in the warehouse must be able to cater for fluctuating sales to prevent holdups and promote labour efficiency. The university researchers cited instances where demand surges during the holiday season and unexpected weather changes caught out logistics companies involved in online sales.
It is against this background that Europe’s major online fashion platform Zalando became interested in how weather affected their demand patterns, and how weather data could ultimately reshape their sales forecasts and lead to a more efficient planning process that is more closely aligned to sales.
This university research used robust statistical techniques to analyse the effect of weather on e-commerce sales, and therefore how weather forecasts can be used to improve sales forecast accuracy and benefits to workforce planning. The location for the analysis was a state in Germany which was deemed suitable for scientific and rigorous statistical research techniques for investigating sales-weather mapping.
This study adopted a seven-day weather forecast model, which considered different weather aspects including temperature, sunshine and rain forecasts.
The authors of the report concluded, “The results of our analyses indicate that good weather decreases online demand and that the effect is stronger on the weekends than during the week.”
However, the researchers had some caveats for supply chain management depending on the time horizon, “Long-term planning processes that require demand forecasts over time horizons that span weeks or months, as is the case for many inventory planning problems, are unlikely to benefit much from weather forecasts.
“However, short-term planning processes that rely on sales forecasts over a few days can benefit considerably from the sales forecast error reductions that weather information delivers.
“More importantly, we show that daily sales forecasts can be significantly improved by augmenting our baseline forecasting model with weather data.
“Another key finding is that weather can help smooth out extreme forecast errors from sales forecasts, which indicates that weather may be useful in avoiding major misalignments in operational plans and marketing campaigns.
“We find that workforce plans based on the weather-augmented sales forecasting model generally lead to lower costs and better performance compared with a model without weather data.”
The results were deemed “robust”. Then they quantified the likely savings in euros, “Our results indicate that including weather information in demand forecasts results in substantial savings in the range of hundred thousand Euros at a single location. For large online retailers with several warehouses, we expect the annual cost reduction to be in the range of a million Euro.”
The researchers suggested the results of the study have other applications, such as in marketing, where more accurate sales forecasts could help companies better align promotional activities. In the fast-paced fashion retail industry, daily weather conditions could improve the reach of marketing campaigns. Other applications such as better short-term forecasting could help in setting pricing strategies.
The research has some limitations as data is taken from a single retailer. However, the authors point out that this retailer, Zalando, is Europe’s largest fashion retailer with a vast product assortment. Another possible limitation of the research is the weather effect is universal and could not be limited to any one product. Despite these drawbacks, the researchers maintain that they have found the direct effect weather has on people’s behaviours and allocation of leisure time. Therefore, the conclusions should be of interest to other online – especially, but not limited to, fashion – retailers.