When Do Pricing Strategies Work for Retailers?
More sales can mean less profit. Marco Vriens examines when discounts, loss leaders and price increases pay off for retailers, and how data and testing can guide better pricing decisions.
Pricing strategies
Retailers can use promotions and loss-leader pricing, both of which involve a temporary or permanent price cut. They can also use hybrid strategies where some items may be discounted while other prices are increased, or they can just increase prices on selected items. Do such pricing tactics work? Promotions and loss leaders both aim to raise traffic, sales and, ideally, profit. They often don't pay off for retailers, but if well-chosen, they can. The difference lies in which items (categories, brands) you pick and who buys them. Price increases can work when you can identify through modelling which ones lead to profit increases.
Loss leaders and promotions
One of the first studies [1] cited showed that loss leaders can raise traffic but rarely raise total store profit: the extra visits don't bring enough full-margin purchases with them. A second study measured the profit impact of every promotion a large national retail pharmacy chain in the US ran (CVS) [2]. Across different categories and brands, on average, the net profit impact was negative. More than half lost money. Across about $9 billion in front-store sales, the net profit effect of all its promotions was roughly minus $25 million. The promotions that paid off were mostly manufacturer-funded buy-one-get-one offers, promotions on CVS’s own store brand, and promotions on high-priced or frequently promoted brands. A large-scale study [3] found that price promotions typically reduce retailers’ category margins and, even after counting cross-category and store-traffic effects, are usually not beneficial to the retailer, while manufacturers’ revenues mostly rise. Funding is part of the story: many promotions are paid for partly or wholly by manufacturers. Ordinary promotions often benefit brands more than retailers. A brand gains when shoppers switch to it from competing brands, and about a third of the typical promotion bump comes from such switching [4]. For the retailer, that switch often just moves a sale from one brand to another on the same shelf, at a lower margin.
A meta-analysis of 1,851 published price elasticities [5] puts the typical short-term response to a promotional price cut at about −3.63. Using the results from this study shows why promotions often don’t work. Say an item sells for $2.50 with a 30% gross margin, so it earns $0.75 per unit. Cut the price by 10% to $2.25, and the margin falls to $0.50 (22%). At 10,000 units, the original price yields $7,500 in gross profit. To earn the same at $0.50 per unit, the retailer must sell 15,000 units, an increase of 50%. With a promotional price elasticity of about −3.6, a 10% price cut lifts sales by only about 36%, to roughly 13,600 units, which earns about $6,800. Gross profit falls by about 9%. A promotion pays only if something else makes up the gap: manufacturer funding, purchases of other items, or new customers.
Retailers often justify promotions by their spillover to other categories. A promotion ripples, but only into a few specific neighbours [6]. Spillovers were larger when supported by an ad in the weekly flyer (or online) or a display (for example, an end-of-aisle placement), and when the related categories sat close together in the store.
Price increase strategies
A discount pays only if its cost is recovered somewhere else, and many retailers already act on this: they manage prices as a portfolio rather than category by category [7]. When a retailer cuts prices in one category, prices in other categories tend to rise. In other words, visible discounts can be funded with margin on items customers watch less closely.
Studies of promotion profitability usually take regular prices as given. The CVS researchers point out that for a retailer that relies on frequent promotions, regular prices are an inflated benchmark: without promotions, everyday prices would be lower. With a more realistic benchmark, CVS’s promotions looked less unprofitable, although the same promotions remained the weakest [2].
The DIY retailer bauMax embarked on an ambitious project to measure how price changes across its 60,000 SKUs affected demand and profits. After the results of their modelling, they tested the predictions with a small number of items in a handful of stores. Once the increased profits of this small roll-out were confirmed and in line with the predictions, they used optimised prices for about 20,000 of its roughly 60,000 items, across all its stores. It deliberately left about 100 items that competitors tracked untouched, to avoid a price reaction. They also tracked overall store perception to guard against the store gradually being seen as more expensive. They increased gross profit by about 8%[8].
Practical tips
1. Diagnose and test. Promotions and loss leaders often are not profitable. The effects vary across categories and brands. CVS used this analysis to pinpoint the categories where promotions consistently lost money. They conducted a field test and showed that eliminating those promotions chainwide would reduce sales by about $7.8 million but raise profit by about $52.6 million a year [9]. The lesson carries over to loss leaders: measure each promotion's net profit, keep the ones that pay, cut the ones that don't, and test before rolling out. At bauMax, extensive econometric modelling showed price increases led to profit increases. They implemented a gradual roll-out, starting with a small set of items in a few stores. Once the profit predictions were validated, a wider roll-out was pursued.
2. If you use promotions and loss leaders, build traffic by featuring items many customers buy often [10]. Expect smaller baskets from those trips, though, and avoid discounting many items in the same category, which lowered store margins in that study. Retailers with larger assortments, or more to cross-sell, have more to gain from loss leaders. To increase spillover, make any temporary discount visible, for example with a feature or a display, and keep related items close[6].
3. Pick items your most valuable customers buy. A theoretical model by DeGraba predicts that the best loss leaders are items bought mainly by high-spending customers[11]. That would explain why grocers discount turkeys on Thanksgiving and eggs at Easter, when baskets are large, but not flowers on Mother's Day.
4. When deploying price increases, use econometric modelling as bauMax did, so you can select the price increases that yield net profit gains.
5. To make promotions and loss leaders profitable, you can fund them from the long tail. Still, you should measure whether it raises profit, and you should watch the overall price perceptions of the store.
Few retailers can answer from intuition which promotions, loss leaders or price increases work best. It takes transaction data and econometric models to see who buys what, a basket analysis to find the items that pull others along, and test-and-control stores to check whether profit shows up under optimised prices. This subject is discussed in part 2 of this series.