Volume Discount? In the Chip Industry, Don’t Count on One

Semiconductor makers’ pricing is based not just on quantities ordered but also on “capacity rationing”

A standard assumption in business is that you get a discount for buying in quantity. Someone buying 1,000 widgets would reasonably expect to pay less per widget than someone buying 100. At the very least, the bigger-volume buyer wouldn’t expect to pay more per widget.

But a study of business-to-business sales of computer chips seems to turn classic thinking about pricing on its head. In a paper published in the journal Management Science, University of Hong Kong’s Wei Zhang, University of Southern California’s Sriram Dasu and UCLA Anderson’s Reza Ahmadi detail their work tracking how one semiconductor manufacturer (not identified) priced its chips relative to the number ordered by different corporate customers.

In a surprising number of cases — about 26% of transactions studied over a three-year period — the manufacturer charged more per chip for large quantities than for smaller quantities of the same chip.

In some cases the premium charged to larger-quantity buyers was modest, under 5%. But some premiums were far bigger. One desktop computer chip sold for an average price of $29.13 to small-batch buyers, while large-batch customers paid a hefty 54% more, or $45.01 per chip.

The authors suggest chip manufacturers’ pricing decisions are based not just on the quantities ordered but also on “capacity rationing” — that is, how a manufacturer values its total production capacity when negotiating with buyers.

The study notes that the semiconductor market typically has short product life cycles (because of constant innovation) and high capital investment requirements, and is subject to intense initial price negotiations between manufacturers and buyers. That all makes it critical for producers to accurately gauge demand for an individual chip from buyers across the order-size spectrum.

A chip producer needs a certain number of large-quantity buyers to “accelerate the selling process,” the authors write. Smaller-quantity buyers, meanwhile, are needed to consume “residual” manufacturing capacity. Then the producer needs to allow for mid-sized buyers to take up capacity that is “too much for small buyers but not enough for large buyers,” the paper says.

Given the impact of any one transaction on subsequent transactions, “it is important for the seller to control the capacity allocated to each buyer, if possible, prior to price negotiations,” the study says.

The chip industry’s selling regime differs from those of most other manufacturing sectors. Many companies naturally seek to achieve maximum sales, and price their goods accordingly, rewarding bigger-volume buyers with discounts if feasible. They can do so as long as production can be ramped up to satisfy demand.

But in the case of semiconductors, “Because manufacturing facilities are costly and construction lead times are long, capacities are inflexible during a selling season,” the authors write.

There are some industries that resemble the semiconductor industry in pricing structure, the study says. “For example, in the travel industry, airline companies and hotels have limited capacities and these capacities have to be sold within a limited period. Customers in these industries include bulk buyers such as travel agencies and resellers who purchase different quantities and negotiated prices.

“Findings from our study may carry over to such businesses,” the researchers write. “Other examples may include movie theaters, concerts and sports events.”