Chapter 3
Price Structure

Tactics for Pricing Differently Across Segments

After developing products or services that create value, a marketer must then determine how most profitably to capture that value in both volume and margin. The challenge in doing so is that customers value products differently because of different abilities to pay, different preferences, and different intended uses. Moreover, the timing of customers’ needs, the speed of their payments, and the level of service and support they require can drive significant differences in the cost to serve them. When a company tries to serve all customers with one price, or a standard markup in the case of distributors and retailers, it is forced to make large tradeoffs between volume and margin—enabling some customers to acquire the product for much less than they would be willing to pay for it, while others are excluded even though the lower price that they would pay is sufficient to cover variable costs and make a positive contribution to profit.

Except for pure commodities, such as ethanol or pork bellies, a single price per unit is rarely the best way to generate revenues. Realizing a company’s profit potential created by the differentiation in its features or services requires creating a structure of prices that aligns with the differences in economic value and cost to serve across customer segments. The goal of that structure is to mitigate the tradeoff between winning high prices for low volume and high volume for low prices. The goal is to capture more revenue from sales where value or cost to serve is higher, while accepting lower revenue where necessary to drive still profitable volume.

To illustrate the huge benefits of a well-defined segmented price structure, suppose that a supplier faced five different segments, all willing to pay a different price to get the benefits they sought from a product (see Exhibit 3-1). Segment 1 with sales potential of 50,000 units is willing to pay $20 for the firm’s product. Segment 2 with sales potential of 150,000 units is willing to pay $15, and so on. What price should the firm set? The right answer in principle

is whatever price maximizes profit contribution. If you calculate the profit contribution at each of the five prices assuming a variable cost of $5 per unit, the single price that produces the maximum contribution ($2,750) is $10.

However, a single-price strategy clearly leaves excess money on the table for many buyers who are willing to pay more: those willing to pay $20 and $15. These high-end buyers perceive significantly greater value from purchasing this product, relative to other buyers. At the price of $10, they are enjoying a lot of what economists call “consumer surplus.” The firm would be better off if it could capture some of this surplus by charging higher prices to these buyers. The second problem is that the supplier leaves nearly half of the market unsatisfied, even though it could serve those customers at prices above the $5 per unit variable cost.

For industries with high fixed costs, serving those additional customers is often very profitable and, when they constitute large amounts of volume, can be essential for a company’s survival. Railroads could not maintain, let alone expand, their costly infrastructures without a segmented price structure. Railroad tariffs are designed to reflect the differences in the value of the goods hauled. Coal and unprocessed grains are carried at a much lower cost per carload than are manufactured goods, resulting in a much lower contribution margin per carload. Still, the large volumes of coal and grain transported enables that low-priced business to make a substantial contribution to a railroad’s high fixed cost structure. If railroads were required to charge all shippers the tariff for manufactured goods, they would lose shippers whose commodities would no longer be competitive on a delivered cost basis and so would lose that profit contribution. On the other hand, if railroads had to charge all shippers the tariff currently charged for a carload of unprocessed grain, their systems would reach capacity before they generated enough contribution to cover their fixed costs and become profitable. Freight railroads survive and prosper by leveraging their capacity to serve multiple market segments at value-based prices for each segment.

Even companies that serve only the premium end of a market often find that it is risky to limit themselves to that segment when they could be leveraging some common costs to serve other segments as well. In his book, The Innovator’s Dilemma, Clayton Christensen cites numerous examples of companies that failed to meet demand from the lower-performance, lower-margin segment of a market that they dominated. Invariably, someone eventually addressed that need and used it as a base to partially support the fixed costs investments necessary to enter higher margin segments.¹ For years, Xerox owned the high end of the copier market. It lost that dominant position only after companies that had entered at the bottom of the market developed service networks of sufficient size to support the higher-priced equipment bought by customer segments, such as copy, centers that require quick service to minimize downtime.

How many segments with different price points should a supplier serve? To return to our illustration, Exhibit 3-1 shows that if the firm were to set two price points serving two general price segments—high-end buyers willing to pay $15 or more, and mid-level buyers willing to pay $8 or more—it could increase profit contribution by 40 percent. But if the supplier could charge separate prices to each of the five market segments, it could increase profit contribution by 80 percent relative to the single price strategy. In principle, more segmentation is always better. In practice, however, the extent of price segmentation is limited by the ability of the seller to enforce it at an acceptable cost.

Segments for pricing are easier to define conceptually than to maintain in practice because customers whom you intend to charge a higher price have an incentive to undermine the structure. They will not freely identify themselves as members of a relatively price-insensitive segment simply to help the seller charge them more, but will try to disguise themselves as customers who qualify for a lower price. Distributors, too, can undermine a segmented pricing strategy by buying the product for delivery to a customer entitled to a lower price but then actually sell to segments that will pay more and pocketing the difference for themselves. This is a huge problem for companies in the European Union because distributors in countries where prices are lower will ship products to one where prices are higher, which often happens simply due to changes in currency values. European law prohibits attempts by national governments to restrict such “parallel trade” even between two European Union countries that have different currencies. Thus, the manufacturer without a segmentation strategy can lose sales in the low-value country due to shortages, while losing margin to competition with “parallel traders” into the high-value countries.

So how can sellers charge different prices to different customers and for different applications? The answer is by creating a segmented price structure that varies not just the price, but also adjusts the offer or the criteria to qualify for it. A segmented price structure is one that causes revenues to vary with differences in the two key elements that drive potential profitability: the economic value that customers receive and the incremental cost to serve them. There are three mechanisms that one can use to maintain such a segmented structure: price-offer configuration, price metrics, and price fences. Each is appropriate for addressing different reasons for the existence of value-based segments.

Price-Offer Configuration

When differences in the value of an offer across segments is caused by differences in the value associated with features, services, or both, a seller can segment the market by configuring different offers for different segments. Using offer design to implement segmented pricing requires minimal enforcement of the segments because customers self-select the offers that determine their prices. The segmented pricing of airline seats described in Chapter 1 is based partially on offer design, with passengers freely choosing whether they want the price that includes the ability to cancel or change flights freely, or want to forgo that feature in return for a much more discounted price. To determine whether it would be profitable to add another offer combination to the menu of choices, you would need to create a spreadsheet analogous to the one in Exhibit 3-2. With that spreadsheet, you could analyze whether the additional offer combination costs more to administer than the incremental profits it would contribute. The right number of price points depends in each case on the sizes of the customer segments, the value and cost-to-serve differences between them, and the cost inefficiency from a proliferation of offers.

To create an effective price structure, one must first determine which features and services the firm should price à la carte, leaving customers to customize their own offers and which features and services to bundle into packages. There are multiple arguments against pricing all individual features and services separately. A single price for a bundle of features and services reduces transactions costs for both customers and sellers. The costs to make and deliver most products and services increase with the number of variations allowed, although technology is reducing the cost of mass customization. Lastly, research has shown that people are less sensitive to the cost of value-added features and services when bundled as a single expenditure.²

Optimizing an Offer Bundle

By creating more than one bundled option designed to appeal to different segments, a marketer can get most of the benefits described above along with the financial rewards of segmentation. Auto manufacturers, for example, put features together in the “sport package” or the “luxury edition” that have a single price for that bundle of options, while cable TV operators create different bundles focused on families, sports enthusiasts, and movie buffs. Since very few buyers would want just one element of the bundle without putting any value on the others, few sales are lost relative to the bundling efficiencies achieved.

Adding to the benefits of bundling, sellers can often earn more profit by pricing a bundle than they could by pricing the individual elements when a particular relationship exists among the features included in the bundle. Bundling is profit enhancing when it is possible to bundle features and services that create high value for some significant customer segments but more moderate value for another. A simple à la carte price for one feature or service that optimized profitability from one segment would necessarily over or under price other segments. Bundling, however, can facilitate more profitable, value-based pricing to each segment. The following example illustrates the principle when the same features can be priced profitably for more than one segment, but the most profitable price level for different segments is not the same.

Musical entertainment can provide an ideal opportunity for profitable bundling, where the “features” valued differently by different segments are the different types of performances. In Boston, where the authors live, one can buy tickets in a series that includes a few headline performers—such as Green Day, Jay-Z, or Kenny Chesney—as well as some lesser-known but often more “innovative” performers such as Kings of Leon or Solja Boy. The challenge is that there are two large customer segments to which these concerts appeal.

There is a large general entertainment segment that views music as just one entertainment option. People in this segment are willing to pay a lot to hear great headline performers, so revenue from them is maximized at a high ticket price (say $60 per ticket). However, they need to be induced to try a concert that is more innovative (no more than $25 per ticket). Without their support, it is unlikely that innovative concerts could attract a large enough audience to justify offering them.

Fortunately, since Boston is home to multiple music schools and music aficionados, there is a smaller segment that is willing to pay as much or more to see new, innovative performers as to see headliner performers. However, because much of this segment consists of students and musicians, they are more price sensitive to the headline performers whose music they have already experienced. The challenge is to maximize income from these two segments combined.

Based upon past research and experimentation, assume that the concert promoters believe that the ticket prices in Exhibit 3-3 represent roughly the acceptable price that would optimize price and attendance by each segment alone at each concert type. Unfortunately, if prices were set at $60 per ticket for headline performances, much of the music aficionado segment would be priced out, leaving some seats empty. Even more importantly for the survival of the concert series, if the innovative performances were priced at $40 per ticket, the large general entertainment segment would fail to show, and so those performances would probably not be viable. Charging $40 per ticket for “headliners” and $25 for “innovations” would fill the halls for both types of

EXHIBIT 3-3 Revenue Optimizing Pricing by Segment for Musical Performances

concerts but would leave a lot of potential revenue on the table. Each segment would be underpriced for some type of concert for which the revenue optimizing price was higher.

Because of this reversal of preference (“headliners” are valued more by the general segment while “innovations” are valued more by the aficionado segment), it is possible to price tickets more profitably as a bundle. After establishing single ticket prices of $60 for headliners concerts and $40 for innovative concerts, the promoter can offer a series of headliner and innovative performances at a discount from those prices that fill the halls. Since the music aficionado segment would pay up to $80 for one headliner plus one innovative performance and the general entertainment segment would pay $85 for the same combination, the series promoter could maximize revenue at $80 for the pair (or $160 for 4, or $240 for 6, so long as the subscriber must choose a specified number of concerts of each type to make up a series). The venues can then be filled and generate more revenue per patron from each bundle of concerts than would be possible with single ticket pricing only (totally only $65 for a pair). The magic behind this is that the different segments are paying the additional $15 per pair of performances for different reasons. Giving them both a reason to pay more within the same bundle facilitates the capture of that value without forgoing volume.

In practice, there are often more than two segments, segments of very different sizes, and more than two types of products to bundle. Maximizing contribution requires building a spreadsheet or employing complex optimization model to evaluate bundling alternatives.³ The principle, however, is the same for bundling features in auto packages, items to include in the four-course dinner special, items in a vacation package, or spots for advertising at different times on a television network. The key is to bundle elements that are valued differently by different segments so long as the incremental revenue earned from inducing more customers to buy an element of the bundle exceeds the incremental cost to supply it. In principle, one could maximize revenue from three segments with one bundle containing three different elements, each valued most highly by one of the segments.

Designing Segment Specific Bundles

Bundling can also facilitate segmented pricing, thus increasing profitability, when different customer segments have different price sensitivity for a “core” product or service (for example, lodging at a popular vacation spot). When it is possible to find features or services that one segment values highly and another does not (for example, access to a pro-quality golf course or a “kids’ club” where children can be left safely and entertained), it is easy to design segment-specific pricing by bundling. The golfer evaluates the sum of the room cost plus the golf cost in figuring the cost of the vacation. If the golfer values lodging at this location by $100 per night more than the family, he will pay up to $100 more per day for greens fees than he would at an equal quality course in a less desirable location. (Assuming, of course, that no cheaper but equal quality course is available near this location.) Since the family did not come to play golf, they are unaffected by high greens fees.

As rewarding but often overlooked is the potential for bundling value-added features and services to attract customer segments that require a lower price to win their patronage. Although they pay a lower price, their purchase volume may, nevertheless, be profitable, especially during off-peak periods or economic downturns when excess capacity would otherwise remain unused. Simply cutting prices to win their business would, however, make it difficult to continue charging other customer segments a higher price and could “cheapen” the image of the brand. Bundling a “free” or low-cost service or feature specifically preferred by this segment, however, can improve the value proposition for that segment without having to cut the offer price explicitly.

For example, the resort hotel could charge a higher price for the room but bundle the “kids’ club” free for one child with each paying adult, admit children free at the breakfast buffet, or provide a shuttle and discount tickets to nearby family-friendly entertainment. Since the golfers would find none of this worthwhile, the attraction to the buyer and the added cost to the seller are limited to the targeted segment. Similar bundles exist in business-to-business markets. Companies that cannot discount prices to small businesses without facing demands for lower prices from larger customers may offer their price-sensitive small business customers low-cost financing, free software for better inventory management, or anything else that they would value but that large company customers would not want.

There is an alternative to adding a feature that raises the value of the discounted offer to only the low-price segment. That is to add a feature to the lower cost offer that kills value for the higher-priced segment without affecting the value to the discount segment. Dick Harmer, a former colleague of ours, gave this practice the memorable name “selective uglification.” Chemical companies often do not have separate lines for making “food grade” and cheaper “industrial grade” chemicals. They simply add something for the “industrial grade” that makes it no longer acceptable for food manufacturers and consumers. The Saturday night stay requirement for a discount airline ticket is another example, since it has no affect on the pleasure traveler who wants the trip to include the weekend anyway, but deters most business travelers.

Unbundling Strategically

While bundling can be a profit-enhancing strategy for segmentation, it often has the opposite effect when variable cost services are bundled simply to differentiate an offering. For example, a business-to-business equipment company might try to convince customers to pay more for its machines by bundling the promise of faster warranty repair service and free delivery anywhere, and a business-to-consumer airline might hope to charge more for its tickets because they include free baggage handling and agent assistance with reservations. Such price-offer structures often undermine rather than enhance profits and can be fatal to companies that cling to them in competitive markets.

The problem arises when the cost to provide the bundled service to customers can be widely different. Customers who have a high need for the bundled services gravitate to the companies that offer them for “free.” As companies gain share among these high-cost service users, the average cost to deliver the bundle increase. If they try to add the increasing average service cost to the price, they begin losing sales to customers who are not high service users. If they avoid raising the price of the bundle to reflect the increasing cost of the service, the increasing cost erodes their margins.

Unless the cost to deliver a service is trivial relative to the overall value of the offer, bundling optional services “free” will undermine profitability. Unbundling them with per use fees or limiting the use of them, as many airlines are doing for baggage handling or for using an agent to make reservations, is in fact strategically essential when facing intense competition. Where customers have come to expect the service to be included, companies can unbundle the price structure without upsetting customers by offering rebates for forgoing use. For example, one company whose customers had become accustomed to placing orders on short notice for “free” raised its prices but, at the same time, offered a discount of more than the price increase for orders to be shipped within seven days. That enabled it to avoid disrupting relationships with customers already paying a premium for its quick service while enabling it to match competitive prices when necessary.

Price Metrics

Not all differences in value across segments reflect differences in the features or services desired. Value received is sometimes not even related to differences in the quantity of the product consumed, necessitating a price metric unrelated to quantity of product or service provide. For example, in the field of health care, both government and private payers are resisting paying for health care on a “fee for service” basis since delivery of more days in the hospital or more tests is often indicative of poor treatment choices, not better patient outcomes. Both payers and health care providers, like Kaiser Permanente and Mayo Clinic, that have a proven ability to deliver care more cost effectively than their peers, have benefited from adopting more value-based price metrics: either a “capitation” price that covers all services required by a patient during a year or a price per illness or procedure that covers all services required to treat a condition to a satisfactory outcome. By adopting such metrics, health care providers that can do that more cost-effectively can avoid the difficult problem of having to convince payers to pay more per service to reflect the value of better treatment. It is much easier to make the case that they can get patients “back on their feet” for no more than the cost per patient of less effective providers.

The example just described involved changing from a feature-based to a benefit-based price metric. Price metrics are the units to which the price is applied. They define the terms of exchange—what exactly will the buyer receive per unit of price paid. There are often a range of possible options. For example, a health club could charge per hour of use, per visit, per an “annual membership” for unlimited access, or per some measure of benefit (inches lost at the waist or gained at the chest). The club might also vary those prices by time of day (low for a midday membership, higher for peak-time membership) or by season of the year to reflect differences in the opportunity cost of capacity. Finally, it might have a multi-part metric: an annual membership with an additional hourly charge for use of the tennis courts. These reflect the common categories of price metrics: per unit, per use, per time spent consuming, per person who consumes, per amount of benefit received.

The problem with most price metrics is that they are adopted by default or tradition. For example, initially software companies charged a price per copy installed on one “server” machine. In most cases, that led to a poor alignment with value. A few creative vendors recognized that when more users accessed the software, the buyer was getting more value. Consequently, they changed the price metric from a price “per server” to a price “per seat,” resulting in customers paying more when they had more users accessing the software. When this “per seat” metric proved much more profitable for the computer-aided design and financial analysis companies that adopted it, other software companies copied it. For many of their applications, however, the number of users still aligned poorly with value, leaving many customers underpriced while pricing others out of the market. The most thoughtful among them created still better price metrics. Leaders in manufacturing software replaced “price per seat” with “price per production unit.” Storage management software suppliers replaced “price per server” with a “price per gigabit of data moved.” Each time a company discovers a better metric than its competitors, it gains margin from existing customers, incremental revenue from customers formerly priced out of its markets, or both.

Creating Good Price Metrics

There are five criteria for determining the most profitable price metrics for an offering (Exhibit 3-4). The first criterion for a good price metric is that it tracks with differences in value across segments. While offer design facilitates different pricing differently based upon what people chose to buy, a price metric not based upon units of purchase can facilitate different pricing for the same offer. For example, it often makes more sense to price drug per day of therapy rather than per milligram of the drug—as Eli Lilly did when it launched the antidepressant Prozac. Someone who requires only a 10-milligram dose gets no less value than someone who requires a 30-milligram dose to control the disease. Consequently, the company charged the same amount per pill regardless of the quantity of active ingredient it contained. Second, a good metric tracks with differences in the cost to serve across customer segments. When customers’ behavior influences the incremental cost to serve them and those costs are significant, a profit-maximizing price metric needs to reflect that as well. The cost to deliver a service is significant if it exceeds the cost of measuring, monitoring, and charging for differences in its usage. Marketers are often reluctant to charge for services, even when costs are significant, because they fear that they will become uncompetitive relative to others who do not charge for them. In fact, the opposite is the case.

Giving services for “free” attracts customers who are relatively higher users of them. Customers who want to minimize their inventories will gravitate to suppliers who offer free rush orders. Customers with a lot of employee turnover resulting in poor equipment maintenance will gravitate to equipment suppliers who offer unlimited and quick on-site service. Customers who require only minimal amounts of service will, in similar fashion, gravitate to competitors offering little or no service but lower prices. As a result, marketers often find that they have differentiated their companies into lower profitability by improving their service offerings because they lack an appropriate metric to capture the value and discourage excessive use of services.

By adding charges for services, at least for those customers who are excessively costly to serve, companies are able to keep their core product prices competitive and avoid attracting a mix of customers who are costly to serve. As their markets have become more competitive, software suppliers added charges for formerly “free” online telephone support. Banks have added charges for small account holders to use a teller, or to access a teller machine more than some authorized amount. United Parcel Service added a $1.75 charge for delivery to a residential address, a $5 charge for customers who don’t put their account number on their delivery slip, and a $10 charge for a wrong address. These charges reflect the added cost of service for such packages, and the tendency for customers to cause those costs when they don’t have to pay for them. Suppliers with separate service charges can price more competitively for the core business (the software, the checking account, the package delivery) to win the customers who are lower cost to serve, while still attracting higher cost customers if they are willing to pay for the higher service levels that they demand. In fact, companies with unbundled service can offer better service because they have a financial incentive to do so.

A third criterion for a good metric is that it is easy to implement without any ambiguity about what charge the customer has incurred. Profit-sharing or performance-based pricing are theoretically ideal ways to achieve the first two criteria for a good metric—tracking with value and cost. But in practice, these methods often end in rancorous debate about how profit or performance should be measured. At minimum, it is important to have absolute clarity in advance about what the metric is and who will measure it. That generally means that the metric must be objectively measured or verified.

We once helped to create a value-based metric at a company whose lubricant enabled manufacturers to cut through difficult materials more quickly with less wear on their tools. The company’s product was an easy sell at launch when potential customers were operating at maximum capacity. Cutting materials faster increased capacity at this stage in the production process, enabling many customers to increase revenues without additional capital cost. But when a recession hit, the value associated with increased capacity fell to zero. The value created by the company’s product was reduced to the savings in labor costs and machine wear.

In theory, the price could be adjusted to reflect the customer’s capacity utilization. However, whenever price depends upon the customer voluntarily reporting information that will lead to a higher price, the potential for conflicts and misinformation is almost always too high. Fortunately we found a published industrial sales index that seemed to track well with the customers’ capacity utilization. The company continued to charge a price per pound for its product, but in return for lower pricing during the recession, the customers accepted automatic price adjustments monthly based upon the level of an industry sales index.

The fourth criterion for evaluating a price metric is how the metric makes your pricing appear in comparison with competitors’ pricing, and the impact of that on the perceived attractiveness of your offer. A new, hosted voice-recognition software that enabled a call center to process more callers without as much need for human intervention promised to create huge differential economic value for purchasers. Unfortunately, the traditional metric for pricing and evaluating hosted call center software was a price per minute of use. Since voice-recognition software processes callers faster, minutes using traditional call center software were not comparable to minutes using the voice-recognition software. A value-based price using that per-minute metric would need to be at least three times the price per minute for traditional software—inviting resistance from purchasers.

To overcome that, the company adopted a new metric: “cost per call processed.” That metric naturally required conversion of the competitors’ cost-per-minute metric into a cost-per-call metric. While the new software was still more expensive, its percent price premium was much smaller when framed in terms of cost per call than in terms of a cost per minute (see Exhibit 3-5). Moreover, the differentiation value of the avoided operator intervention was much more dramatic when framed in terms of cost per-call rather than cost perminute basis. The total cost per call was less with the new software, despite being higher on a per-minute basis. While the favorable economics of the new software was exactly the same using either metric, the per-call basis of comparison made the sales effort a lot easier.

The fifth and final criterion for evaluating a price metric is how the metric aligns with how buyers experience the value in use of the product or service. The better the alignment—how a price metric fits the timing and magnitude of the customers’ expenditure—the more attractive the offer. Movie renters get value from watching a DVD once, not from the amount of time they have it in their possession. Netflix changed the metric for film rentals when it recognized that the decline in the cost of making disks no longer required incentives for renters to return the disk quickly. By replacing the rental metric based on time ($3.95 per day) with a metric based on the number of films “out at-a-time” ($8.99 per month for one DVD at a time, $13.99 per month for two, and so forth), Netflix eliminated the inconvenience of having to acquire the movie shortly before watching it and return it shortly thereafter. That made the Netflix metric more compelling than the video store

EXHIBIT 3-5 Hosted Call Center Software

metric for a large share of the video rental market, which Netflix won over amazingly quickly.

In some cases, it is not possible to achieve all of these criteria with one metric, but they can be achieved with a multi-part metric. Mobile telephone service providers charge a fixed monthly fee, capturing the value of simply having access to a phone when needed, plus charges for the amount of different services consumed (calls, text messages, Internet time). Amusement parks sometimes have an entry fee plus a ticket charge for each ride. Banks charge a monthly fee for an account, plus additional charges for transactions. Each of these structures is designed to strike a balance between service cost recovery, encouraging use that drives volume with pricing that is seen as aligned with value, and capturing more profit from those customers who are getting more value.

Performance-Based Metrics

An ideal price metric would tie what the customer pays for a product or service directly to the economic value received and the incremental cost to serve. In a few cases, called “performance-based” pricing, price structures can actually work that way.⁴ Attorneys often litigate civil cases for which they are paid their out-of-pocket expenses plus a share of the award if they win, rather than for hours worked. Internet ads are usually priced based on the number of “click-throughs” rather than the traditional metric for advertising: cost “per thousand” exposure. Systems that control the lights, heating, and cooling within office buildings are sometimes installed in return for contracts that share the energy cost saving, rather than charges for the equipment installed. In each case, the price metric naturally charges customers differently for the same product or service based on differences in the value they receive.

Most importantly, performance-based pricing has the effect of shifting the performance risk from the buyer to the seller. General Electric (GE) used bundling to reduce risk when it launched a new series of highly efficient aircraft engines, its GE90 series. These engines promised greater fuel efficiency and power that could make them much more profitable to operate. The catch was a high degree of uncertainty about the cost of maintenance. Some airlines feared that these high-powered engines might need to be overhauled more frequently, thus easily wiping out the financial benefits from operating them. This undermined GE’s ability to win buyers at the price premium that power and fuel efficiency would otherwise justify.

Rather than accept a lower price to account for a buyer’s perceived risk, GE absorbed the risk by changing the price metric. Instead of selling or leasing an engine alone, GE effectively rented aircraft engine for a fee per hour flown that included all costs of scheduled and unscheduled maintenance. Without the uncertainty of maintenance cost, GE90 engines quickly became popular despite a price premium. In most cases, however, “performance-based pricing” is simply impractical. It requires too much information and too much trust that the buyer will actually report the information accurately. It also leaves the buyer uncertain regarding the cost of a purchase until after it is used. In practice, therefore, marketers must design profit-driven price structures by finding measures that only roughly predict the value a customer will receive and the costs to serve. Often the difference between a good and a great pricing strategy lies in finding, or creating, such measures.

Tie-Ins as Metrics

A very common challenge for a company that sells capital goods is that the value of owning them can vary widely across segments based upon how intensely they are used. For example, a company that makes a uniquely efficient canning machine might like to sell it both to salmon packers in Alaska, who will use it intensely for only a couple months each year, as well as to fruit and vegetable packers in California, who will use it to can crops all year round. One option would be to put a meter on the machine to record every time that machine went through one cycle. That, in fact, is how Xerox priced its copiers at launch, leasing them at a price based upon machine usage and refusing to sell them outright.

For the canning machine manufacturer that did not expect to have service people at the client site on a regular basis, the idea of a usage-based lease was not practical. What was practical was a “tie-in sale” that contractually required purchasers of the canning machine to use it only with cans sold by the seller at a premium price. Thus, the true cost of the machine was not just its low explicit price but also the net present value of the price premiums paid for the tied-in cans. Since buyers who used the machine more intensely must buy more of the tied-in product to use it, they effectively paid more for the asset.

“Tie-in” sales like those that tied purchase of cans contractually to purchase of the machine were quite common until 1949, when the federal courts decided that such contracts were not enforceable under U.S. antitrust law because of their impact on the otherwise freely competitive market for the tied commodity.⁵ But although contractual tie-ins are no longer enforceable, companies still frequently use technological design to tie a unique consumable to an asset. For example, Hewlett-Packard (HP) led the industry in the development and manufacture of inkjet printers. HP strategically priced the printer—the asset—low to make the up-front cost competitive with much lower-quality printers. The replacement ink cartridge—the consumable—was designed with proprietary technology to fit uniquely with the asset and carried a remarkable wholesale margin of 60 percent. The key to HP’s pricing success is that its pricing allowed it to earn some profit from its superior printers from the many light users who bought them and its high-priced ink. But HP could earn much more from heavy users who need to replace their printer cartridges more frequently. This tie-in strategy enables HP’s inkjet division to maintain a 50 percent market share and profit per dollar sales ratio that is twice that of the company in general.

In service-based companies, tie-in contracts are frequently used to reduce the cost for new buyers to try their services. Wireless phone providers offer a digital telephone for a nominal fee, and sometimes free, if the buyer agrees to purchase a long-term service contract to use the company’s wireless network for 12 or 24 months. Satellite entertainment companies offer households a satellite dish and receiver unit for a greatly reduced price when buyers agree to subscribe to a higher-priced entertainment package of channels for a minimum of 12 or 24 months. These packages can be particularly effective for low-knowledge buyers, who perceive significant risk in investing in a new and little-known technology—and then developing them into loyal buyers who become accustomed to the firm’s technology and programming.

Price Fences

Sometimes value differs between customer segments even when all the features and measurable benefits are the same. Value can differ between customer segments and uses simply because they involve different “formulas” for converting features and benefits into economic values. The difference may be tied to differences in income, in alternatives available, or in psychological benefits that are difficult to measure objectively. Unless there is a good “proxy” metric that just happens to correlate with the resulting differences in value, the seller needs to find a price fence: a means to charge different customers different price levels for the same products and services using the same metrics.

Price fences are fixed criteria that customers must meet to qualify for a lower price. At theaters, museums, and similar venues, price fences are usually based on age (with discounts for children under 12 years of age and for seniors) but are sometimes also based on educational status (full-time students get discounts), or possession of a coupon from a local paper (benefiting “locals” who know more alternatives). All three types of customers have the same needs and cost to serve them, but perceive a different value from the purchase. Price fences are the least complicated way to charge different prices to reflect different levels of value. Unfortunately, while simple to administer, the obvious price fences sometimes create resentment and are often too easy for customers to get over whenever there is an economic incentive to do so. Thus, finding a fence that will work in your market usually requires some creativity.

Buyer Identification Fences

Occasionally pricing goods and services at different levels across segments is easy because customers have obvious characteristics that sellers can use to identify them. Barbers charge different prices for short and long hair because long hair takes more time to cut. But, during nonpeak hours, barbers also cut children’s hair at a substantial discount, despite the fact that children can be more challenging and time consuming. The rationale in this case is entirely to drive business with a discount for a more price-sensitive segment. Many parents view home haircuts as acceptable alternatives to costly barber cuts for their children, even though they would never bear the risk of letting their spouses cut their own hair. For barbers, simple observation of the customer segment, children, is the key to segmented pricing.

Issuers of credit cards resort to far more sophisticated, proprietary models to anticipate the price sensitivities and costs to serve for different types of consumers. Some are more sensitive to the annual fee, some to the interest rate, and others to the frequent flyer miles they can earn. On the cost side, some consumers are more likely to default or to use their card only infrequently, thus generating fewer fees from retailers for processing charges. Finally, the companies can see from consumers’ credit reports what competitive cards they hold and can estimate their annual fee and interest rates, thus determining the “reference value” of the next best competitive alternative (NBCA). Based upon these analyses, credit card companies very finely segment their potential customer base and send out different offers that optimize the expected profitability of each segment. The metrics are the same, but the levels vary depending on which metric the issuer can use most cost effectively to capture the most value.

Rarely is identification of customers in different segments straightforward. Yet, management can sometimes structure price discounts that induce the most price-sensitive buyers to volunteer the information necessary to identify them. Many service providers, from hotels and rental car companies to theaters and restaurants, offer “seniors” discounts to those who will show an American Associations of Retired Persons (AARP) card, Medicare card, or some other ID that confirms their eligibility. College students qualify for discounts on various types of entertainment because their low incomes and alternative sources of campus entertainment make them, as a group, price-sensitive shoppers. Seniors and students readily volunteer their identification cards to prove that they are members of the price-sensitive segment. Members of the less price-sensitive segment identify themselves by not producing such identification.

Even schools and colleges charge variable tuitions for the same education based on their estimates of their students’ price sensitivities. Although the official school catalogs list just one tuition, it is not the one most students pay at private colleges. Most receive substantial discounts called “tuition remission scholarships” obtained by revealing personal information on financial-aid applications. By evaluating family income and assets, colleges can set tuition for each student that makes attendance attractive while still maximizing the school’s income.

Deal proneness is another form of self-induced buyer identification— especially through the use of coupons and sales promotions, a frequent tool of consumer marketers. Coupons provided by the seller give deal-prone shoppers a way to identify themselves.⁶ Supermarkets and drug stores put coupons in ads circulars because people who read those ads are part of the segment that compares prices before deciding where to shop. Packaged-good and small appliance manufacturers print coupons and rebate instructions directly on the packages, expecting that only price-sensitive shoppers will make the effort to clip them out and use them for future purchases.⁷

Often a buyer’s relative price sensitivity does not depend on anything immediately observable or on factors a customer freely reveals. It depends instead on how well informed about alternatives a customer is and on the personal values the customer places on the differentiating attributes of the seller’s offer. In such cases, the classification of buyers by segment usually requires an expert salesperson trained in soliciting and evaluating the information necessary for segmented pricing.

The retail price of an automobile is typically set by the salesperson, who evaluates the buyer’s willingness to pay. Notice how the salesperson takes a personal interest in the customer, asking what the customer does for a living (ability to pay), how long he has lived in the area (knowledge of the market), what kinds of cars she has bought before (loyalty to a particular brand), where she lives (value placed on the dealer’s location), and whether she has looked at, or is planning to look at, other cars (awareness of alternatives). By the time a deal has been put together, the experienced salesperson has a fairly good idea how sensitive the buyer’s purchase decision will be to the product’s price. (Note: If you want to get the best price, show the sales rep your printout from the Internet of the wholesale cost of the car and its features, and offer $200 more. You will save yourself and the sales rep a lot of time.)

Purchase Location Fences

When customers who perceive different values buy at different locations, they can be segmented by purchase location. This is common practice for a wide range of products. Dentists, opticians, and other professionals sometimes have multiple offices in different parts of a city, each with a different price schedule reflecting differences in the target clients’ price sensitivity. Many grocery chains classify their stores by intensity of competition and apply lower markups in those localities where competition is most intense. Colorado ski resorts use purchase location to segment sales of lift tickets. Tickets purchased slope side are priced the highest and are bought by the most affluent skiers, who stay in the slope-side hotels and condos. Tickets are cheaper (approximately 10 percent less) at hotels in the nearby town of Dillon, where less affluent skiers stay in cheaper, off-slope accommodations. In Denver, tickets can be bought at grocery stores and self-serve gas stations for larger discounts (approximately 20 percent less). These discounts attract locals, who know the market well and who are generally more price-sensitive because the ticket price represents a much higher share of the total cost for them to ski.

A clever segmented pricing tactic common for pricing bulky industrial products such as steel and coal is freight absorption. Freight absorption is the agreement by the seller to bear part of the shipping costs of the product, the amount of which depends upon the buyer’s location. The purpose is to segment buyers according to the attractiveness of their alternatives. A steel mill in Pittsburgh, for example, might agree to charge buyers the cost of shipping from either Pittsburgh or from Gary Indiana, where its major competitor is located. The seller in Pittsburgh receives only the price the buyer pays, less the absorbed portion of any excess cost to ship from Pittsburgh. This enables the Pittsburgh supplier to cut price to customers nearer the competitor without having to cut price to customers for whom his Chicago competitors have no location advantage. The Chicago competitor probably uses the same tactic to become more competitive for buyers nearer Pittsburgh.

Trade barriers between countries once made segmentation by location viable even for products that were inexpensive to ship. As trade barriers have declined around the world, and especially within the European Union, the tactic has become less effective. For example, automobiles used to be sold throughout Europe at prices that varied widely across borders. German luxury cars sold in Britain were often 20 percent more expensive than when sold just across the channel in Belgium. Now, brokers in Britain will survey the continent for cars, which people can fly to pick up and drive home—or have the broker bring it back for them. To fight back, some makers of German luxury brands, which are cheaper in Germany than in some other countries where they carry a more premium image, have used their warranties to enforce location fences. A car bought in Germany and imported to Britain cannot get warranty service in the United Kingdom without paying an additional charge for warranty transfer.

Time of Purchase Fences

When customers in different market segments purchase at different times, one can segment them for pricing by time of purchase. Theaters segment their markets by offering midday matinees at substantially reduced prices, attracting price-sensitive viewers who are not employed during the day at times when the theater has ample excess capacity. Less price-sensitive evening patrons cannot so easily arrange dates or work schedules to take advantage of the cheaper midday ticket prices. Restaurants usually charge more to their evening patrons, even if they cater to peak crowds at lunch, because demand (in the United States, but not in Europe) is more price sensitive for the midday meal. Why? There are more numerous inexpensive substitutes for lunches than there are for dinners. A Big Mac or a brown bag, acceptable for lunch, is generally viewed as a poor substitute for a formal dinner as part of an evening’s entertainment.

Priority pricing is one example of segmenting by time of purchase. New products in a retail store are offered at full price, or sometimes premium surcharges over full price in the case of extreme excess demand. Over time, as product appeal fades in comparison to newer competitive alternatives, buyers discount the product’s value until they are willing to pay only a fraction of its original price for leftover models. This is a common tactic in the retail fashion and automobile industries, where customers with high incomes and low price sensitivity pay premium prices for the latest styles and models and can choose from a full inventory of sizes and colors. Over time, as inventories age and the availability of sizes and colors declines, prices are reduced in successive rounds of promotions to appeal to more price-sensitive buyers who are willing to wait for the opportunity to buy high-quality, but less trendy, inventory and with less certainty of obtaining their preferred size or color.

Priority pricing also applies in business-to-business purchases. A favorite strategy of Intel is to introduce a leading-edge semiconductor at a premium price, and then discount its existing semiconductor product lines. Leading-edge original equipment manufacturer (OEM) computer manufacturers that produce and sell the fastest and latest computers to innovative professional buyers with low price sensitivity pay the price premium for the latest chip technology. More price-sensitive buyers who are willing to accept slightly outdated technology are then offered older-model computers equipped with Intel’s now older semiconductors at lower prices.

Predictable, periodic sales offering the same merchandise at discounted prices can also segment markets. This tactic is most successful in markets with a combination of occasional buyers who are relatively unfamiliar with the market, and with more regular buyers who know when the sales are and plan their purchases accordingly. Furniture manufacturers employ this tactic with sales every February and August, months when most people usually would not think about buying furniture. However, people who regularly buy home furnishings, and who are more price sensitive because of the reference price and total expenditure effects, know to plan their purchases to coincide with these sales.

Time is also a useful fence when demand varies significantly with the time of purchase but the product or service is not storable. The problem plagues airlines, hotels and restaurants, electric utilities, theaters, computer time-sharing companies, beauty salons, toll roads, and parking garages. Unable to move supplies of their products from one time to another, their only option is to manage demand. One way of doing so is with peak-load pricing, the implementation of which we will discuss in Chapter 8 when we explain how and when to adapt pricing to changes in the cost of capacity.⁸

Pricing for travel through the Eurotunnel between England and France is an interesting application of segmented pricing for a product with fixed capacity. The channel tunnel allows transport of an automobile and its occupants for a flat price between Folkestone, England, and Calais, France. Prices that allow travel at whatever time of day you choose are twice as high as during the off-peak evening and night periods. This reflects the opportunity cost of limited capacity. More interesting is the fact that rates increase with the time elapsed between the outbound and the return trips. Roundtrip use of the tunnel for a two-day, one-night visit from the United Kingdom to France costs from £44 per auto while roundtrip use for a three- to seven-day visit costs from £78 per auto for use of the tunnel at the same times of day. Clearly, this has nothing to do with cost or available capacity, so what drives it? The answer is that the value of having your own car with you on the trip, versus having to rent one after traveling by plane or train, increases with the length of the stay.⁹

Purchase Quantity Fences

When customers in different segments buy different quantities, one can sometimes segment them for pricing with quantity discounts. There are four types of quantity discount tactics: volume discounts, order discounts, step discounts, and two-part prices. All are common when dealing with differences in price sensitivity, costs, and competition.¹⁰ Customers who buy in large volume are usually more price sensitive. They have a larger financial incentive to learn about all alternatives and to negotiate the best possible deal. Moreover, the attractiveness of selling to them generally increases competition for their business. Large buyers are often less costly to serve. Costs of selling and servicing an account generally do not increase proportionately with the volume of purchases. In such cases, volume discounting is a useful tactic for segmented pricing.

Volume discounts are most common when selling products to business customers. Steel manufacturers grant auto companies substantially lower prices than they offer other industrial buyers. They do so because auto manufacturers use such large volumes they could easily operate their own mills or send negotiators around the world to secure better prices. Volume discounts are based on the customer’s total purchases over a month or year rather than on the amount purchased at any one time. At some companies, the discount is calculated on the volume of all purchases; at others, it is calculated by product or product class. Many companies give discounts for multiple purchases of a single model but, in addition, give discounts based on a buyer’s total expenditure on all products from the company.

Although less common, some consumer products are volume discounted as well. Larger packages of most food, health, and cleaning products usually cost less per ounce, and canned beverages cost less in twelve-packs than in six-packs. These differences reflect both cost economies for suppliers and the greater price sensitivity for these products by large families. Warehouse food stores, such as Wal-Mart, Costco, Sam’s, and BJ’s often require consumers to buy in large-quantity packages to qualify for discounted prices.

Often sellers vary prices by the size of an order rather than by the size of a customer’s total purchase volume. Order discounts are the most common of all quantity discounts. Almost all office supplies are sold with order discounts. Copier paper, for example, can be purchased for about $20 per case of about 10 reams, but purchased individually it costs several dollars per ream. The logic for this is that many of the costs of processing an order are unrelated to the size of it. Consequently, the per-unit cost of processing and shipping declines with the quantity ordered. For this reason, sellers generally prefer that buyers place large, infrequent orders, rather than small frequent ones. To encourage them to do so, sellers give discounts based on the order quantity. Order discounts may be offered in addition to volume discounts for total purchases in a year, because volume discounts and order discounts serve separate purposes. The volume discount is given to retain the business of large customers. The order discount is given to encourage customers to place large orders.

Step discounts differ from volume or order discounts in that they do not apply to the total quantity purchased, but only to the purchase beyond a specified amount. The rationale is to encourage individual buyers to purchase more of a product without having to cut the price on smaller quantities for which they would pay a higher price. Thus, in contrast to other segmentation tactics, step discounting may segment not only different customers, but also different purchases by the same customers. Such pricing is common for public utilities, from which customers buy water and electricity for multiple uses and place a different value on it for each use.

Consider, for example, the dilemma that local electric companies face when pricing their product. Most people place a very high value on having some electricity for general use, such as lighting and running appliances. The substitutes (gaslights, oil lamps, and hand-cranked appliances) are not very acceptable. For heating, however, most people use alternative fuels (gas, oil, coal, and kerosene) because of their lower cost. Utilities would like to sell more power for heating and could do so at a price above the cost of generating it. They do not want to cut the price of electricity across the board, however, since that would involve unnecessary discounts on power for higher-valued uses.

One solution to this dilemma is a step-price schedule. Assume that the electric company could charge a typical consumer $0.06 per kilowatt-hour (KWH) for general electricity usage but that it must cut its price to $0.04 per KWH to make electricity competitive for heating. If the company charged the lower price to encourage electricity usage for heating, it would forgo a third of the revenue it could earn from supplying power for other uses. By replacing a single price with a block-price schedule, $0.06 per KWH for the first block of 100 KWH and $0.04 for usage thereafter, the company could encourage people to install electric heating without forgoing the higher income it can earn on power for other purposes. To encourage people to use electricity for still more uses, such as charging their car batteries during off-peak hours, utilities often add another step discount for quantities in excess of those for general use and heating. Exhibit 3-6 illustrates a step-price schedule for an electric utility.

Given the clear increase in profit from offering step discounts, effectively moving along an individual customer’s demand curve, why do most companies still offer each individual customer volume at only one price? The answer is that segmenting different purchases by each customer is possible only under limited conditions. It is profitable only when the volume demanded by individual buyers is significantly price sensitive.

Summary

Designing an optimal price structure that effectively segments your market and maximizes your profitable sales opportunities is clearly among the most difficult, but potentially rewarding, aspects of pricing strategy. For companies that are launching an offering with differentiated benefits or employing a business model with a different cost structure, creating a new price structure that aligns with those differences is usually necessary to capture the profit potential associated with them. Even without such a change, a company that can incrementally improve the price structure can gain profitable incremental volume. The principles of price structure discussed in this chapter, and the examples cited to illustrate them, can serve as a guide to a better basis for collecting revenues across segments. There is no simple formula. Each case requires creativity to find the best means to implement those principles within your market. It is, however, one of the most important activities that a marketer can do to improve profitability, since the investment required is small relative to other marketing investments, and the payoff is often very large.