Clear Demand Presents Fundamentals of Retail Science: Episode II (Grocery Science)

Grocery science is the analysis that supports the business of selling food to consumers. It’s a part of retail science the way grocery is part of retail, but grocery science emphasizes specific aspects of retail decision. Grocers have less variation in product and demand than many other retailers and they also have smaller margins. Their decisions revolve around product selection, display, promotion, and inventory.

In this paper we’ll start with product selection and display in Section 2, move on to promotion in Section 3, and cover inventory replenishment in Section 4 We’ll touch on pricing, competition, and trend analysis in Section 5.


There’s a lot of emotion surrounding the decision of what to sell, during what seasons to sell it, and how to display products. There’s a lot of science surrounding those same decisions. Selecting and displaying products scientifically means more of the shelf space is earning more money. Let’s look at some of the questions science can help answer.

  • Are the must-have products visibly displayed? Certain products must be on the shelf or shoppers will go elsewhere. Obvious must-have products are staples (milk, eggs, and bread) and impulse products (beer and cigarettes). These products should be easy to find on the shelves without restricting sales of other, higher-profit products.
  • Are the popular products displayed on the shelves? After the essential products are the products so popular that their absence will make an impression on food shoppers. Not only should the big sellers be displayed, but shoppers expect big-selling products to have large display areas.
  • Is extra display area going to produce extra money? Bigger displays sell more than smaller ones, but seldom proportionally more. Doubling the exposure adds from 10 to 30 percent to sales, so multiple-facing products should be chosen carefully.
  • Does the display mix have good product coverage? A grocer typically has shelf space for a few tens of thousands of products out of a grocery product space of hundreds of thousands. Shoppers looking for S-size of product P will accept alternatives that are not too different, perhaps a different size of their brand, perhaps an alternative, similar brand.
  • Are high-impulse, high-profit products at eye level? Stuff that shoppers are going to buy anyway can go lower or higher. An impulse-buy vegetable like asparagus goes right at reach level while staples like mixed vegetables go on the bottom shelf. Is all the middle shelf space working hard to bring in revenue and profit?

Retail science can help answer these questions using knowledge from experienced retail management and mathematical, statistical analysis.


Promotions properly planned can increase a grocer’s audience, expand the revenue footprint, and make higher profits. Less-well-planned promotions draw shoppers away from a store’s profitable products toward buying the same stuff they would have paid full price for.

Let’s look at what retail science can tell us.

  •  What are the direct effects of promotion choices? Promotions generate sales–that’s why we do them. Sale signs, endcaps, fliers, and media advertisements all increase units sold. While we almost always associate promotion and price reduction, retail scientists can separate the price effect and promotional lift. This means we can understand and forecast promotion-price combinations we have not tried before.
  •  How can a promotion event support promotion strategy? There are several reasons to promote products, but they fall into three basic areas, increasing traffic, increasing sales revenue, or increasing profit. Once a strategic direction in this traffic-revenue-profit space is established, we can use retail science to ensure all the promotions and prices in the event are aligned with the strategic goal of the event.
  • How do affinity and cannibalization affect promotion events? Affinity happens when boosting one product’s sales boosts the sales of other products. One typical affinity example is putting hot dogs on sale and finding an increase in buns and relish. Cannibalization is when boosting one product’s sales is at the expense of another. My usual example is putting Pepsi on sale and finding Coke, RC, and C&C sales going down. We put Pepsi on sale to generate new soft-drink sales and not to switch Coke, RC, and C&C buyers to Pepsi. These shoppers are buying the same amount of soft drinks on sale that they would have bought for regular price. Affinity and cannibalization require intense data-mining efforts to observe because they are secondary effects of a promotion, but very important effects.
  • What is the effect of pantry loading on grocery promotions? Some shoppers see a sale as an opportunity to stock up on a product, an effect we call pantry loading. These shoppers are buying the same amount of the product on sale now that they would have bought for regular price later. This is a bad thing for the grocer unless the regular-price-later sales would be at a competitor’s store. Like affinity and cannibalization, pantry loading is a statistically-subtle effect and requires data-mining technique to verify it is happening.

  Retail science can direct attention to real promotion effects and keep us from chasing secondary affinities, cannibalization, and pantry loading effects when they aren’t really happening.


Controlling inventory levels is a huge cost tradeoff for a grocer, more than most other retailers. Profit margins in grocery are particularly slim so small changes in inventory cost make all the difference. Since food shoppers are more predictable than other retail shoppers, we can use retail science to squeeze down inventory costs while managing risks.

  • Are inventory levels too high? Most times a shopper wants to buy a product at a store and when that product is out of stock, we lose the full value of the sale. On the other hand, inventory sitting on the shelf is costing us the same as cash in the mattress instead of money earning interest in the bank. This is a difficult trade-off compounded by how consistent shoppers are when they buy products, how often we can replenish inventory, how reliable delivery and restocking processes are. (This is not the same problem as how much shelf space to devote to selling each product in the store, but the two are closely tied together.)
  • What are appropriate order-interval strategies? The act of ordering product costs something, perhaps sending trucks and drivers to remote locations or perhaps just the cost of managing the paperwork. So we can order products frequently and keep just enough inventory to keep them in stock or order less frequently and maintain larger reserves.
  • What is the best way to manage distribution centers (DCs)? For the same reason that airlines use hub-and-spoke routing, grocers have vendors deliver merchandise to DCs who then stock stores. The efficiency of routing carries a price of complexity: inventory has to managed in more places, warehouse space is dear, and the DC-to-store link has to be optimized along with the vendor-to-DC link.
  • Can we sell perishable products before they spoil? Many food products have defined shelf lives, dates after which they may no-longer be sold. Making sure these products are on the store shelves in time to be sold is a challenging transportation and distribution problem. Spoilage is worse than being out of stock because the grocer loses the full value of the product, not just the potential profit of selling it.

The constant theme in all of these questions is dealing with uncertainty. It’s hard enough to figure out how to solve displaying, stocking, ordering, holding, and warehousing products in a perfect-information world and real-world grocers have ongoing uncertainty about what shoppers will buy. There are variations in shopper buying behavior from which we derive certain statistical distributions and then there are global variations such as weather and competition. A quantitative, scientific approach to inventory can make intelligent trade-off decisions between the risk of running out and the cost of holding too much.


These display, promotion, and inventory decisions come from deep analysis of years of sales experience. It takes serious science to figure out the effect on sales of display, price, and promotion, to find affinity, cannibalization, and pantry loading, and to understand the impact of inventory decisions. Once we have that scientific infrastructure, we can make important grocery decisions.

We’re looking for pricing in the food market but also pricing relative to local competition. We’re looking for forecasts of revenue and profit to facilitate financial planning. We’re looking for trend analysis so we can see shifts in the consumer market quickly so we can act rather than react.

  • How do we distinguish between seasonality and trends? By looking at years of sales data, we can get a clear picture of the seasonal effects. This includes the calendar seasons of spring, summer, autumn, and winter and the calendar of special occasions like Christmas, Valentines Day, U.S. income-tax refunds in May, and specific-holiday-vacation weekends. Once we have the seasonal behavior, we can look at the remaining time changes in demand, which we call residual in statistics, and decide if there is an increasing or decreasing trend.



There is a great deal of heart and soul in designing a grocery-shopping experience in a store. Having a clear mental picture of the store image is the foundation of retail, turning that image into a profitable grocery business is the job of retail science.