Marketing events specialist, Interactions Marketing, has tied up with Google to use the search engine’s cloud solution for big data analysis plus visual analytics software to predict the impact of weather on retail sales.
Using Google BigQuery and Tableau’s visual analytics software, Interactions layered transactional sales data with regional weather data to understand shopper behaviour.
It focused on identifying repetitive weather events, classifying them on a severity scale, and measuring the effects major weather events had on sales before, during and after those events. With the analysis, Interactions was able to track new patterns in sales and shopper behavior, which it claims can help retailers and CPGs plan successful in-store activities in advance of major weather events.
According to the company, this insight will be used as a predictive tool during and prior to specific weather events to help retailers minimise or eliminate out-of-stock issues, optimise item assortments in high demand categories, and increase sales.
Interactions Marketing said early results suggest these insights could deliver significant benefits, including boosting the return on investment for marketing and advertising spend and improving targeted and individualised shopper communications resulting in increased basket size and cannibalisation of competitors’ shoppers.
“Interactive performance of Google BigQuery, combined with Tableau’s intuitive visualisation tools enabled our analysts to interactively explore huge quantities of data – hundreds of millions of rows – with incredible efficiency. In some cases taking analysis that would ordinarily require a week down to just hours and minutes. This time-to-insight was previously impossible,” said Giovanni DeMeo, vice president of global marketing and analytics for Interactions.
“It enabled us to visually share that information with our retailer and CPG partners, and use it to enhance in-store activity and increase sales. This is only one of an infinite number of ways that we will now be using big data to improve the revenue and profitability of our partners.”
The total store analysis revealed, down to the individual product level, which items had the most significant change in sales as well as what varied in shopper behaviour for similar weather events when considering time of day, day of the week, geographic location, and even proximity to competitor locations.
The data identified 28 categories with significant changes in sales versus the control. The data showed that one day prior to statistically similar weather events the sales in these categories spiked from 20% to 261% over same day previous year, but also noted a drop in sales during the peak of the event and for four days after.
These behaviours were evident in regions that experienced the event, along with regions where the event was predicted but did not actually occur. If weather reports predicted a storm a week ahead, for example, people still waited until the day before the event to do their event-specific shopping. In one scenario, and contrary to every other shopping behavior, this resulted in a huge spike in Monday sales over the preceding weekend for a predicted Tuesday weather event.
“Retailers have access to massive amounts of complex data to help them make good decisions. The trick is to find a way to easily visualize and analyse it effectively,” said Francois Ajenstat, director of product management at Tableau Software.
“By combining the flexibility and horsepower of Google BigQuery with Tableau’s visual analytics, Interactions has delivered insights that were not previously revealed. Retailers and CPGs will now be able to make real-time data driven decisions to inform their business.”
Interactions, Google and Tableau will be sharing many more details about this use case during the 2013 Tableau European Customer Conference on 11 June 2013 in London.