Professor of Operations, Division Chair of Operations
Why the allocation mechanism smoothes demand fluctuations
An odd thing happened in Kellogg, Idaho, during the 1990s. Despite the town’s small size — population 2,422 — and relative isolation — Spokane, Washington is over an hour away — local car dealer Dave Smith Motors started selling more Dodge trucks than anyone else in the nation. By 1996, it was pushing nearly 4,000 extended-cab Ram pickup trucks out the door.
We’ll send you one email a week with content you actually want to read, curated by the Insight team.
Part of the surge was due to an Internet ad touting no-haggle pricing that attracted customers from around the country. But there was more to Dave Smith Motors’ impressive sales than just savvy marketing. “They also had a GMC dealership that ran the same way on the Internet but was not dominating the market the same way,” says Martin Lariviere, a professor of managerial economics and decision sciences at the Kellogg School of Management. The difference between the Dodge and GMC dealerships was not the way the two were run, but the way Dodge and GMC allocated trucks to dealers.
“Dodge was running strictly on a turn-and-earn scheme, and if you moved the metal, they gave you more,” Lariviere says, referring to the method commonly used by the automotive industry to assign capacity when manufacturers cannot keep up with demand. Dealer allocations are based on previous sales, with high-selling dealerships receiving more capacity in the next round. Because of this, the sales reps at Dave Smith Motors “were really cleaning up on Dodges,” he says.
Needless to say, this was upsetting other dealers, some as far away as Boise, Idaho, and Great Falls, Montana. Not only did they feel they were losing potential customers, some thought they were not getting the number of trucks they deserved. Dealers are valuable assets to auto companies, and Dave Smith Motors’ exploitation of the turn-and-earn system was making some of them upset. Given the pitfalls of the scheme, you would be forgiven if you thought Dodge quickly dropped turn-and-earn when it learned of Dave Smith Motors’ hijinks. It didn’t.
Dodge clearly thought turn-and-earn was an advantageous scheme, or else it would have allocated stock differently. But researchers have had a difficult time identifying precisely why turn-and-earn is an efficient way to dole out products. One of Lariviere’s earlier papers showed that the scheme can put geographically separated dealers into competition, like what happened in the case of Dave Smith Motors. But that reason alone did not satisfy him. Lariviere had another hunch, though. In his new paper, written along with Lauren Xiaoyuan Lu, an assistant professor at the University of North Carolina at Chapel Hill, he shows turn-and-earn can reduce uncertainties that propagate up the supply chain, taming the so-called bullwhip effect.
The bullwhip effect can be a nightmare for supply-chain managers. “The term supposedly originated with Proctor & Gamble and diapers,” Lariviere says. “The end consumer, so to speak, uses Pampers at a pretty constant rate, but when P&G looked at the way supermarkets were ordering, they saw big swings, and when they looked upstream at how they were placing orders on their suppliers, like 3M, they saw even bigger swings.” Forecasting material orders or planning new factories requires confident estimates of future demand, and anything that dampens the bullwhip effect allows companies to better estimate demand and operate more profitably.
Auto manufacturers in particular have large incentives to keep demand for their products smooth. Keeping cars and trucks moving to dealers keeps costly inventory levels low. Furthermore, adding production capacity by building a new factory or retooling an old one can be prohibitively expensive and risky. For example, demand for a new model may evaporate before a retooling is completed.
While auto makers are most commonly associated with turn-and-earn, allocating capacity based on past sales also happens elsewhere, Lariviere says. A short while back, supplies of Eggo waffles were constrained because of the two factories that produce them, one was shut down for maintenance and the other was flooded. Bounty paper towels were on allocation for a period of time, too. Demand spiked when a reformulation of the product proved popular, and Proctor & Gamble’s production capacity could not keep up. One solution was to build a multimillion-dollar paper mill, a decision that would be profitable only if demand remained high, Lariviere points out. Turn-and-earn bought the company time while it evaluated demand.
Taming the Bullwhip
Lariviere and Lu suspected turn-and-earn may have some role in reducing the bullwhip effect, so they constructed a model in which dealers are assigned capacity based on the allocation scheme. Unlike previous models ‚which had only two time periods, where capacity in the second was allocated based on sales in the first, this one ran ad infinitum. Previous publications had to settle for two time periods as the mathematics proved difficult. Extending the model’s horizon required Lariviere and Lu to apply the tools of computational economics.
High-sales dealers not only have a clear advantage under turn-and-earn when demand is high — they receive the highest allocation and so can sell more — they will fight hard to keep their lead when demand dips to the medium level.
Next, they mimicked real-world fluctuations in demand by introducing three levels — low, medium, and high — rather than one. Not only did this more faithfully represent the real world, it also allowed them to see under what conditions a dealer like Dave Smith Motors would fight to retain sales dominance. Finally, Lariviere and Lu’s new model introduced fluctuations in local demand, such as when the road in front of a business is being repaired.
Lariviere and Lu found that high-sales dealers not only have a clear advantage under turn-and-earn when demand is high — they receive the highest allocation and so can sell more — they will fight hard to keep their lead when demand dips to the medium level. Lariviere explains from a dealer’s perspective: “If I’ve got a lead on a competitor and we get down to that medium state, I will sacrifice something so I can clean up when the market is great.”
In other words, by substituting higher sales for higher profits during slight sags in demand, dealers can guarantee high future allocations. When things pick back up, they can recover their losses through renewed strong, high-profit sales. “But,” Lariviere continues, “if demand gets really soft and the market is really slow, then I will give up.” When demand drops off a cliff to the low state, Lariviere and Lu’s model predicts that previous sales leaders cannot continue to sell at a loss to guarantee future allocation.
Up Side to Down Sales
Low demand is not bad for everyone, though. It can give low-sales dealerships a chance to catch up, Lariviere and Lu found. The recent recession confirms their result. “Toyota has traditionally been strong on the coasts, particularly the West Coast” but not in the Midwest, Lariviere says. As American car companies started to flounder in the early 2000s, Toyota sales in the Midwest started to pick up. But Toyota ran into a problem, he adds. Because it allocates based on turn-and-earn, dealers on the coasts continued to receive more of the hot sellers like the Sienna minivan. It was not until sales for the entire industry dropped later in the decade that Midwest dealers had a chance to catch up.
“Back in 2007 and 2008 as the overall economy started to slow down, there were various dealers in the Midwest who were being quoted in the trade press saying, ‘I can finally get Siennas’,” Lariviere says. East and West Coast dealers simply could not maintain their high sales and so lost allocation. Should the recession end quickly, then Midwest dealerships will likely emerge with better sales allocations than before. However, Lariviere cautions, if sales are low for an extended period of time, everyone suffers.
The dynamics are similar on the local level. Turn-and-earn can mollify short-term disturbances in demand. Let’s say the road in front of a car dealer is being repaired by the city and customers are finding it a hassle to turn into the dealership. The dealership knows the construction will be over in a few months, and rather than reduce orders, they will keep orders high to ensure future allocations. They may lose some money during construction, but would likely lose more money in the future if the manufacturer reduced their allocation.
“Turn-and-earn is a way of mitigating variability in the supply chain, which is something previous models didn’t actually address,” Lariviere says. By encouraging dealers and retailers to absorb local variability in demand — reducing the bullwhip effect — manufacturers allocating based on turn-and-earn have an easier time planning for the future.
Related reading on Kellogg Insight
Professor of Operations, Division Chair of Operations
Tim De Chant was science writer and editor of Kellogg Insight between 2009 and 2012.
Lu, Lauren Xiaoyuan and Martin Lariviere. 2011. “Capacity Allocation over a Long Horizon: The Return on Turn-and-Earn.” Manufacturing & Service Operations Management. DOI: 10.1287/msom.1110.0346.
Getting children to make healthy choices is tricky — and the wrong message can backfire.
A conversation between researchers at Kellogg and Microsoft explores how behavioral science can best be applied.
Acquiring another firm’s trade secrets — even unintentionally — could prove costly.
Common biases can cause companies to overlook a wealth of top talent.
A new study suggests that firms are at their most innovative after a financial windfall.
Don’t let a lack of prep work sabotage your great ideas.
Training physicians to be better communicators builds trust with patients and their loved ones.
The fallout can hinge on how much a country’s people trust each other.
Tim Calkins’s blog draws lessons from brand missteps and triumphs.
Three experts discuss the challenges and rewards of sourcing coffee from the Democratic Republic of Congo.