Rite Aid Exec Talks About the Weather
With three consecutive quarters of same store sales and adjusted EBITDA growth, could weather be a contributing factor in Rite Aid’s success? On Tuesday at the annual Category Management Conference in Atlanta, Rob George, director of market research for Rite Aid, talked about how better use of business weather intelligence is helping in the revitalization of the company.
Using raw weather data alone is too unreliable, according to Mr. George, because it relies on subjective, individual interpretation of the situation, which is inefficient as too many people try to solve the same problem. So Rite Aid works with Planalytics to try to get to a "single version of the truth." Since weather is highly variable and is only similar to the previous year 20 percent of the time, Mr. George said Rite Aid aims to do two things: 1) ‘Deweatherization’ — removing weather-driven volatility from historical results so they can plan and allocate from weather-neutral positions; and 2) ‘Weatherize’ its sales plan — adjusting in-season weather impact on business by accounting for weather events.
Weather driven demand (WDD) is the measurement of how much lift or drag weather provides and does not include any factors other than weather. Rite Aid looks at this information on a weekly basis, down to the store level. Looking at WDD allows Rite Aid to change assortment and stock levels accordingly, heavying up on rain gear for one part of the country and ordering less for another, for example.
The company is able to see long-term forecasts up to 11 months out, and the forecast gets a major fine-tuning two weeks out, enabling last-minute adjustments. Rite Aid plans stock, promotions, displays, ads, and even markdowns according to weather forecasts. And they can manage their facilities better, knowing when to start seasonal landscaping work, when to plan for snow removal, and even when to verify snow removal bills.
Additional benefits of weather forecasting at retail might include social media — offering coupons or special deals to shoppers on Facebook or Twitter based on local conditions and/or long and short term forecasts.
Discussion Questions
Discussion Questions: How useful do you think long term weather forecasting is, or should be, to retailers? How should it be properly factored into other decisions related to advertising, promotion and merchandising?
If consumer wants and needs drive their purchasing behavior, I have to believe that long term weather forecasting can be vital. It’s easy enough to trend. Success will warrant continued use and more success. Am I missing something?
So with all our “sophisticated” technological advances like yesterday’s discussion about tracking shoppers through their phones — it all gets down to the weather. Funny how nature has a way of bringing a needed dose of humility to human kind. Maybe that’s why virtually every conversation on the planet begins with a comment on the weather. Reminds me of that old quote about the attendance at your funeral being largely dependent on the weather!
Sorry, but I just don’t buy this. If the local weather forecasters can’t get it right; how does Rite Aid expect to accurately forecast to get the right inventory levels? This is a stretch.
While weather is certainly a key variable for retail forecasting, it’s hard to control and hardly predictable other than within a narrow time frame. Customers are inherently more predictable not only in terms of behavior, but also in their ability to drive sales and profit growth. I wonder how much Rite Aid is spending on weather insights versus customer insights?
Weather has an amazing impact on the supply and demand of consumer goods. Recently, most of the stores on the East Coast were out of stock on flashlights and other storm mode goods almost immediately after the forecast of the hurricane. The same happened in the Midwest with approaching ice storms last winter. It’s impossible to know for certain the weather many months in advance, however, seasonal trends and the likelihood of certain trends are well known in this day and age of a much more intelligent weather science.
This is big. Improving the accuracy of weather analytics and forecasts will both drive sales and reduce seasonal markdown losses. The bonus: shoppers will get it and gravitate to the store that is not out-of-stock.
Weather has always been a huge factor in retail. The bottom line is, people don’t come out to play when the weather is bad. Can long-term forecasting play a role? Yes, in areas such as marketing and facility planning, but at the store level, schedules should remain a one week or 2 week plan and be flexible. If there is a huge snowstorm and I have no customers, there is no point in having 10 associates sitting around doing nothing. That doesn’t mean I have to send them home (although I’m sure you would readily have volunteers) but there should always be a rainy-day plan. It’s a great time to spot check the entire store, execute side projects or just empty out the back room. One particular company I worked for even had a small 5 day forecast window directly on the workforce management screen.
While there may be some value on the margin from this, it is just not critical to how retailers win. Retailers need to find the actions they control that really drive demand or switch demand to their stores. Certainly, if analysis is not done well enough, weather can easily cloud (pun intended) a retailer’s analysis of whether a given program works, so it is important to net that effect out. But the focus should be on understanding sales driving programs, not predicting the weather.
Sell snow to Eskimos in the summer time? Nothing new here without seeing the actual plans — before Rite Aid implemented this “new methodology” vs. prior. Certainly if advanced forecasting can support decisions — both good and bad — with a rationale as to why goods should have been stocked heavier in one Rite Aid district or region vs. another, it’s a sound investment. I doubt that advanced weather forecasting is making customer insights through advanced analytics a lower priority. One is leveraged for the merchant organization and the other for their Marketing/CRM pillar.
I can’t help but remember trying to explain sales blips to senior management because of snow storms and being admonished to “not give a weather report” but just the facts. I’m sorry but this sounds like a desperate attempt to take attention away from much more important sales issues to me.
I think integration of weather data into forecasting applications is just the beginning. With more and more data being collected, the impact of exogenous factors on store sales becomes easier to quantify and utilize. In the case of weather, you need to capture both the actual weather and the forecast. We all know what the forecast of snow, whether it ever actually falls or not, does to supermarket sales. I am surprised Rite Aid has not been doing this for a long time.
I remember an undergraduate book on forecasting describing it as “Like driving a car blindfolded while following directions from someone looking out the back window.” You really only have two options for generating forecasts: historical data and contemporaneous correlations. The challenge with historical data is that there are many factors, including weather, that impact the results (what week was Easter last year?). The good news is that many of the other factors such as holidays, school schedules, seasonality, transportation schedules (buses, trains, etc) are usually known and can be factored into the final forecasting algorithm. It is a lucky situation where a correlated factor can give you good causal data for anticipating sales. In Rite Aid’s case, I imagine they could expect cold remedy sales to pick up a few days after a snow or cold air event. Cold remedies could be forecast using a combination of seasonal and correlated data factors.
Having said all this and firmly believing that weather data is good to have, the truth is that weather forecasting is still very much a guess. Despite all the sophisticated algorithms, going beyond 48 hours is still a tricky proposition. This means that the other critical factor in all this is replenishment lead time. A delivery schedule that includes a delivery every two days is operating right at the cusp of the forecasting capability. This is where the whole concept of in-store safety stock and layered inventory become critical to ensuring the service level can be achieved. Merchandisers hate back rooms, they’d rather have the stock on the selling floor, but for highly weather-affected merchandise, it may be better to have local safety stock in the backroom during the appropriate season.
Rob is a really smart retailer! We always used to say, “weather plays”; meaning that no matter how well you nailed your product, people, place, etc., weather was a similar factor. So, to be able to have some knowledge of whether or not you’re going to get snowed out this winter or have fantastic weather, would be a major advantage in terms of planning. It’s a totally over-looked factor and usually the biggest reason you see huge markdowns post-crummy weather. Good thinking — as long as it’s at least close to right, of course.
Leveraging information about the weather is another form of personalization, which can be used to let a customer know that the company is thinking about them as individuals and their local if not specific needs. “Weatherization,” to be successful, should be combined with other information that a retailer knows about the customer to create a deeper, more substantial conversation and relationship, which will be longer and have a greater lifetime value.