5 minute read

Unraveling the Layered Retail Forecast: Part 1

Conner Snodgrass, Implementation Advisor (US)

Mark Twain once said, “Climate is what we expect, weather is what we get.” While the terms weather and climate are closely related, both dealing with changes in atmospheric variables such as humidity, temperature and wind, they do have different meanings. The climate of an area is determined over decades while weather refers to the same variables over a shorter period, typically days or hours. Thinking back to Mr. Twain’s quote, it is easy to anticipate the climate for an area. Predicting the day-to-day weather is considerably harder. What if we could accurately predict the weather? What if we could also consider what impact the weather will have in that region? What if we could layer in these impacts to existing projections to improve their accuracy? In fact, the top forecasting solutions are already considering all these factors and more. In today’s post, let’s investigate what role the weather plays in forecasting, and how we can make weather considerations an input into our forecasting process.

What makes a good forecast, and why is it important?

Before we dive into the specific impact of the weather on the forecasts on which retailers rely, what makes a good forecast and why is having one so important? A forecast is only as good as its inputs. The most important input to the forecasting process is the data itself. To generate an accurate forecast, you need data that has “Quantity and Quality.” Leading forecasting techniques utilize artificial intelligence and machine learning to build a mathematical model based on “training data.” The greater quantity of “training data” the model has, the better result you will get. Quality data is devoid of errors and erroneous outliers. Once the inputs are settled, a great forecast can be produced.

And why do we want accurate forecasts? Empirical evidence suggests that forecast accuracy leads to lower inventory levels and higher profits, which adds value to any enterprise. In the retail world, recent trends to retain employees and to comply with predictive scheduling requirements have retailers forecasting and writing schedules weeks in advance. The value of an accurate forecast in these circumstances cannot be understated. An accurate forecast of metrics like sales, items and customers sets the foundation for an accurate labor forecast to ultimately schedule the right people, at the right time, doing the right things, with minimal need for costly last-minute changes.

What role does the weather play in forecasting?

Now that we understand the value of an accurate forecast, what role does the weather play? The answer to that question will vary by region. 75 degrees and sunny will have a different effect on shoppers in Michigan vs. Southern California. The weather impacts the decisions of people every single day, from what they have for lunch, to what outfit they wear, and even the tasks they will perform that day. The conditions outside can be the deciding factor on whether a consumer shops online or visits the store and can even shift how their hard-earned money is spent across certain products and services. Weather patterns also affect a myriad of business operations from the supply chain all the way down to frontline staffing considerations. The impact of the weather is not only felt during the day-to-day operations, but also weekly, monthly and quarterly through seasonality. The most advanced forecasting solutions can consider not only the weather forecast and seasonality, but also scale the impact to the local store or region being forecasted.

How can weather considerations become an important input into our forecast?

How can we go about making sure our forecasts incorporate this weather data to become more accurate? The most important consideration here will be selecting a forecast vendor who can automate this process while still considering the most detailed weather data available. Traditionally, accounting for weather in retail forecasts has been limited to omitting and/or tagging historical data. While effective, these have many limitations, such as manual oversight, lack of flexibility, and lack of granularity. The best forecasting solutions will understand the local weather, how that compares to the average for that time of year, and layer in other adjustments such as seasonality and automatically incorporate them into the forecast.

Advanced forecasting solutions are now able to partner with companies that provide detailed weather data for every region across the globe. This weather data includes variables such as temperature, precipitation and wind speed. By comparing the weather forecast to historical weather during the same period, in the same region, forecasters can answer a very critical question, “is this a normal condition for this location?” If the conditions are out of the ordinary, then an adjustment to the forecast either up or down might be in order.

Let’s look at a real-world example of how including regional weather considerations in your forecast can improve operations. A local retailer notices that the weather forecast for next week is cold and rainy, which is abnormal in their state Texas for this time of year and will certainly affect shopping patterns and store traffic. How can they use this knowledge to improve their forecasts and make sure that the store is prepared for this decrease in shopping activity? Once the impact of say, a big rainstorm, has been quantified it is then much easier to project how that will affect future forecasts and schedules. If a rainy week will decrease shopping activity, fewer employees are needed to deal with the reduced number of customers, and there is an opportunity to optimize schedules around lower labor costs.

An additional benefit to having a weather event quantified is to understand the impact if the weather forecast changes. What about the times where the 5-day weather forecast calls for rain, but each day ends up 75 degrees and sunny? The most accurate forecast is often the one with the most up-to-date data. In the forecasting world, this act of incorporating the most recent data is known as “Reforecasting.” Forecasting solutions that are constantly reforecasting to include the latest data points, whether it be the local weather or a new sales trend, will empower management to make the most informed decisions with the latest data. Let’s apply that scenario where the weather forecast was wrong to our real-world example. On Monday of the previously forecasted “rainy” week, the reforecast identifies that the weather patterns have changed and it will now be a week with average temperate and sunny weather. Now that we understand the impact that the downpour had on our forecast, we can simply adjust the schedules back to a more normal level to prepare for the newly increased level of shopping activity. This may involve putting up bid shifts, gig scheduling or even crowdsourcing due to the short lead time when additional labor is needed. The most advanced forecasting solutions are constantly reforecasting, including weather impacts, to allow the retailer to see the impact and make any alterations with the most up-to-date information.


The weather impacts the decisions of people every single day and can be the deciding factor as to whether a consumer shops, shops online or visits the store and what they buy. The best forecast solutions incorporate detailed, region specific, weather data to forecast the sizable impact of the local weather more accurately on retail patterns. By leveraging climatology, retailers can have more accurate forecasts and schedules, allowing them to operate more efficiently while increasing customer satisfaction. Consider these advantages when selecting a forecast provider in the future.

Try the retail industry’s most accurate AI and M-driven metric forecasting for free! Complementary 8-week forecast at no cost to you. Learn more.

Continue reading

Let’s Connect