However, there is surprisingly little research into some aspects of the relationship between weather and sales. That’s particularly true when we consider the impact of weather on sales at the whole UK level, rather than just on individual products or specific locations. Despite regular commentary blaming or praising the weather for how retailers are doing, how much the weather affects sales is not well understood. Through application of machine learning techniques to their combined datasets, the BRC and the Met Office sought to shed some light on this question, focussing on to what extent temperature can predict growth in Non-Food sales. The findings of the study have implications for both retailers and market observers:
• Differences in temperature, are a good predictor of growth in Non-Food sales growth, but only at certain times of the year.
• While some periods, particularly the Christmas season, show little relationship with the weather others are highly sensitive to temperature. During the Summer to Autumn transition, nearly half of the variation in weekly sales growth from year to year over this period can be accounted for by differences in temperature between years. The analysis implies that for each degree warmer the first week of September is relative to the previous year, the rate of growth of Non-Food sales is reduced by 1.1%, equivalent to around £40m of sales.
• When there is a relationship, it doesn’t always work in the same direction. At sometimes of year warmer weather can help sales, at other times it can hinder them.
•The study found little evidence that, at an aggregate level, temperature has a permanent impact on sales. If September is warmer and consumers do not buy as many cold weather products, they spend more on these products once cooler weather finally arrives. However, unseasonal weather can strongly affect monthly growth figures and be easily mistaken for changes in underlying consumer confidence. It can also impact profitability, with retailers selling more stock on discount following a period of unexpectedly weak sales.
• The strength and pattern of relationship to the weather differs across categories of Non-Food (See Table 1). It is particularly strong in Clothing, during the transition to Autumn. Temperature differences can explain more than 70% of year to year variations in growth of Men’s Clothing in mid-September to early October and can reduce the rate of sales growth by 2.4 percentage points for each degree it is warmer than the previous year, equivalent to £6m a week.