Markdown Management: In Pursuit of the Optimal Discount

Will your customers buy an item immediately? Or hold out for a sale?

If they’re waiting, what can you do to spur a purchase?

In new research, Darden Professor Manel Baucells argues that retailers can grow profits by cutting prices sooner and more deeply than conventional modeling suggests.

Baucells’ model, called dPTT (discount, probability and time trade-off,) accounts for how consumers’ buying habits are influenced by “psychological distance,” a measure of how sensitive a consumer is to delays and risks in product availability, whether perceived or real. In a series of experiments, Baucells discovered that the optimal markdown using dPTT could be about twice the discount advocated by the standard markdown model (e.g., 20 percent instead of 10 percent in their base cases situation).

Further, because it better approximates the ways humans actually behave, Baucells says that dPTT has the potential to increase markdown revenue 20 to 25 percent, and overall revenues by 1.5 percent. A typical retailer operates with a net margin of 3 percent, which means that “each percent of extra markdown revenues translates into major profit increases,” write Baucells and colleagues Nikolay Osadchiy of Emory University and Anton Ovchinnikov of Queen’s University, Ontario.

“Effectively, every time a customer enters the store, he or she mentally ‘solves’ a buy-or-wait problem.” The idea, Baucells says, is to “set a markdown that is large enough to attract new customers, yet small enough to control strategic waiting.”

Markdown Mania

Given the prevalence of markdowns and their complexity to administer, retailers are always looking for more effective strategies. According to Baucells, nearly one-third of items are sold at discounted price, generating about 20 percent of retail revenues.

In the past, according to Baucells, retailers would slash prices — on average 50 percent — in hopes of attracting new customers.

“Price discounts were seen as a form to expand the market by selling to consumers that are not willing to pay the full price. But the cannibalization effect — the loss of sales at full price — was difficult to calculate and often ignored,” he says. “The implicit assumption here is that consumers were myopic, and would ignore the possibility of waiting and buy at a discounted price.”

Of course, this backfired: Customers became accustomed to waiting for sales.

Researchers came up with a new model: discount expected utility, or DEU, which advocated for either an “everyday low price” strategy or very modest (e.g., 10 percent) discounts. DEU has multiple strengths and is “directionally correct,” according to Baucells. It holds under a broad set of circumstances, yields exact formulas for the price discounts, and predicts how many consumers will buy now and how many will wait. Its main weakness, however, is it assumes consumers behave rationally at all times.

Rational, but Not Completely

 In reality, consumers fall into a gray area. “Actual consumers are not irrational, but they are not 100 percent rational either,” Baucells notes.  He says that dPTT is, in essence, a “modification of DEU that better approximates how individuals feel about time and risk trade-offs.”

For instance, the DEU model assumes that consumers treat changes in time and availability in a consistent, linear way. For instance, under DEU, a one-week delay in getting a product is a consistent level of “penalty,” whether it is a new delay that takes away immediate availability, or an extension of an already existing wait.

In reality, Baucells says, “the typical consumers will be reluctant to wait for one week instead of getting the product today, but may not mind waiting for four weeks instead of three.”

“Once a certain psychological distance has been created, there is not much to be gained from having additional distance,” Baucells observes.

Additionally, a customer is going to feel a risk in a product selling out more keenly when that product was previously perceived as fully available than when the chance of getting a product already looked remote.

In summary, the buy-or-wait decision is “a multidimensional trade-off,” Baucells says, and by factoring in how consumers weigh risks and delays, “dPTT creates a pricing model that can be more credible and useful.”

Manel Baucells co-authored “Behavioral Anomalies in Consumer Wait-or-Buy Decisions and Their Implications for Markdown Management,” forthcoming in Operations Research, with Nikolay Osadchiy of Goizueta Business School at Emory University and Anton Ovchinnikov of Smith School of Business at Queen’s University.