Weather Analytics and Retail Sales
Wednesday, February 8, 2012 at 9:58AM After crunching the numbers, the National Climatic Data Center (NCDC) has found that January 2012 was the fourth warmest January on record across the contiguous United States. This is also the mildest January since 2006, which was the warmest in records dating back to 1895.
States with a top 10 warmest January (9 total) - AZ, KS, MO, MN, ND, NE, OK, SD, WY
The Weather Channel, LLC
Weather can have a significant impact on retail sales. Consumer’s behavior changes, distribution can be impacted, regular seasonal selling can shift, etc. Our team has recently completed analytical projects with customers using precipitation, temperature, humidity, and many more weather data points to understand retail sales patterns and then use that understanding to create forecast models. This is the beauty of store / UPC grain retail sales data. Combining retail point of sale (EDI 852) demand data with weather data, you can identify fascinating and very useful insights. Some things to keep in mind….
Useful weather analytics almost always requires day grain retail data. Week grain data is useful for some weather analytics but there are significant limitations. EDI 852 is often weekly grain, but sometimes day grain is available. Portals like Retail Link can provide daily grain (or lower if you want) retail sales reports so target your project to your retail customers that provide day grain retail sales data.
Studying the data carefully to identify statistical significance is critical. Antidotal or observational research is helpful to inform your statistics but be careful about over simplifying what you see (e.g. it rained and sales are up) until you have run the numbers.
Do apply your industry and product knowledge. If you sell a product that conventional wisdom says is impacted by precipitation or temperature, then use that as a starting point for building the model. If the output of the model challenges the conventional wisdom, then dig into the model and look for holes until you are satisfied with the accuracy of the results.
A quality weather analytics project is not an inexpensive project, so be prepared to make an investment. But on the flipside, we have seen these investments provide huge returns for highly weather dependent product categories.
Reader Comments (1)
Interesting post and something we can relate to on how weather impacts trends within digital marketing, specifically in online retail - how specific weather parameters determine fluctuations in product sales in specific sectors - please take a look at this recent post: http://www.weatherlytics.co.uk/blog/warmest-april-helps-heat-up-searches-for-garden-furniture/