The adverse financial impact of average costing is greatest for those companies, such as service providers, whose products are not storable. Such companies face the problem of having to build capacity to serve temporary but predictable “peaks” in demand. This creates the interesting situation where the cost of capacity goes from being incremental to sunk, and back to incremental again, over the period of a year, a month, a week, or even a day. Airlines face peaks at the beginning and ends of weeks but have excess capacity midweek and on weekends. Telecom companies face peaks in the middle of each weekday but have excess capacity in the evenings and on weekends. Restaurants, car rental companies, marketers of advertising space, commercial printers, health clubs, resorts, electric utilities, and landscape maintenance companies all face substantial peaks and valleys in demand for nonstorable products or services. One way to manage capacity in those cases, and to thus maximize profitability, is with price.
The key to using price to manage capacity profitably is to understand how to allocate the capacity costs over time. Most companies make the mistake of averaging the capacity cost over all of the units produced. If an electric utility sells 40 percent of its kilowatts during a few peak hours, and 60 percent during the other 21 hours in the day, then 40 percent of the capacity cost (the depreciation and maintenance cost for the power plants) would be allocated to the peak hours. This results in each kilowatt being assigned the same capacity charge. Although this is the usual approach (utilities have even been required to cost and price this way by regulation), it makes no sense in principle and undermines profitability in practice. Why? Because the need for the capacity is created entirely by the peak period demand. Off-peak demand can be satisfied without the additional capacity, so capacity costs are not incremental to decisions that affect the volume of off-peak sales. Consequently, the cost of capacity above what is necessary to meet off-peak demand should be allocated entirely to the peak period sales.
One effect of allocating those costs only to sales in the peak period is to raise the hurdle required to justify peak-period investment. The way to ensure that capacity costs are covered is to make no capacity investments that cannot be entirely justified by the revenue from peak demand. If additional capacity really is required only for a few hours a day, a few days a week, or a few months a year, then prices in those periods should be covering all the cost of the capacity, or the capacity should not be built. The other effect is to reveal the surprisingly high profitability of the lower-price, lower-margin sales in off-peak periods. Because the contribution from off-peak sales is not required to cover the cost of capacity, which will be there whether or not the capacity is used, that contribution falls directly to the bottom line. Companies that fail to realize this overinvest for peak demand and then are forced either to cut their prices to fill their off-peak capacity or to suffer even greater losses during off-peak periods. On net, they invest themselves into unprofitability.
As a company moves toward pricing differently for the peaks and valleys, its average price decreases, but its profit and return on capital invested will increase. For companies with a peak capacity problem, it is usually far more important that they earn a high return per unit of capacity than it is for them to earn a high contribution margin per sale. For many years, hotels mis-measured their success by their ability to command and increase their “average daily rate.” Of course, one way to increase average daily rate is simply to rent no rooms except at times of peak demand when the hotel can ask for and get its highest rates. That is unlikely, however, to yield a good return on assets. When hotels began being managed more rationally, the industry adopted a new measure, “revenue per available room,” that changed the incentive to manage capacity. The bottom line became “Get all you can get at peak, but make sure you fill the room and earn something at off peak.”
The same argument would apply when costing the use of a manufacturing facility, a railroad right-of-way, or the seat capacity of an airline. The historical cost of those assets is entirely irrelevant and potentially a very misleading guide to pricing. There is often, however, a current cost of using those assets that is very relevant. That cost occurs whenever capacity used either could be used to make and sell some other product, or could be rented or sold to some other company. Even though the historical cost is sunk, the relevant cost of using those assets is positive whenever there are competing profitable uses. That opportunity cost is the contribution that must be forgone if the assets are not sold or used to produce the alternative product or service. It can easily exceed not only the historical cost but also even the replacement cost of the capacity.
Even if a company has current excess capacity but there is some probability that future business might have to be turned away, the capacity should be assigned an opportunity cost for pricing. Airlines, for example, stop selling discounted seats for a particular flight long before the flight is full. The opportunity cost of selling a discounted seat is near zero only if that seat would otherwise certainly be empty at flight time. As a plane's capacity fills, however, the probability increases that selling a discounted seat will require turning away a passenger who would have paid full fare on the day of the flight. The probability of such a passenger wanting the seat, times the contribution that would be earned at full price, is the opportunity cost of selling a discounted seat in advance.
Obviously, moving beyond these costing principles to estimating the true cost of a sale is not easy. Too often, however, managers shrink from the task because of the cost and complexity of measuring true costs on an ongoing basis. Usually, in our experience, it is possible to get much closer to true costs by doing even a simple point-in-time study of cost drivers. We have, for example, worked with a company that charged every item produced the same amount for paint, even though some items were produced in large lots and others in small lots. By doing a simple statistical regression, using prior year data, paint purchases by color as a function of production of that color product, and average lot size, we rationally reallocated paint costs to reflect the higher costs of small batches. In other cases, we have relied on cost drivers as subjective as a plant foreman's judgment about the relative difficulty of making different types of products. Are such judgments highly accurate? Probably not. That is not a reason to avoid making them if that is the best you can do with the time and money available. It is better to make pricing decisions based on rough approximations of the true costs of products or services than on precise accounting of costs that are sure to be, at best, irrelevant and, at worst, highly misleading.