2
This chapter introduces the basic concepts behind pricing and revenue optimization. We first look at some of the common pricing challenges faced by organizations—most notably a lack of consistent management, discipline, and analysis across pricing decisions. The chapter then describes three purist approaches to pricing—cost-plus, market based, and value based—and discusses some of their shortcomings. It then introduces pricing and revenue optimization. At the highest level, pricing and revenue optimization is a process for managing and updating pricing decisions in a consistent and effective fashion. At the core of this process is an approach to finding the set of prices that will maximize total expected contribution, subject to a set of constraints. The constraints reflect either business goals set by the organization or physical limitations, such as limited capacity and inventory. While the use of constrained optimization is common to all pricing and revenue optimization applications, the type of problem to be solved depends on the specific characteristics of the market. Markets vary in terms of timing or cadence of pricing decisions, the nature of the goods and services being sold, and the type of customer commitment being priced.
2.1 THE CHALLENGES OF PRICING
For many organizations, pricing includes a remarkably complex set of decisions. While most companies have a good idea of the list prices they have established for their products, they are often unclear on the prices that customers are actually paying. A multitude of different discounts, adjustments, and rebates are often applied to each sale. For this reason it is critical to distinguish between the list price of a good and its pocket price—that is, what a particular customer ends up actually paying. The list price is generic, while the pocket price may be different for each customer. The price waterfall was introduced by McKinsey and Company as a graphical way of illustrating the discounts that occur between the list price of a good and its pocket price. A consumer package goods (CPG) example is shown in Figure 2.1. In this case, 12 price reductions or discounts are applied between the list price and the pocket price. These include an 8% competitive discount, a 3% sales special, a 1% exception deal, and so on, down to a 1% freight allowance. The net result is that the pocket price for this customer is 29% less than the list price.
Figure 2.1 Price waterfall for a consumer package goods (CPG) company. Source: Courtesy of Mike Reopel of A. T. Kearney.
The price waterfall illustrates quite neatly that the pocket price paid by an individual customer is often not the result of a single decision, but the cumulative result of many different decisions. In fact, for the majority of companies, many discounts are independently set by different parts of the organization, without consistent measurement or tracking. In Figure 2.1, the competitive discount might have been authorized by the regional sales manager, while the product bonus was determined as part of a general marketing program and the freight allowance was given by the local salesperson in response to a last-minute call by the purchaser. As a result, no one is truly in charge in the sense that no one in the organization is responsible for the fact that the discount offered to this customer was 29%, while that offered to another might be 18%. In fact, not only is no one in charge, but it is often remarkably difficult to determine what the pocket price paid by a particular customer even is. As two McKinsey consultants put it:
The complexity and volume of transactions tend to create a smoke screen that makes it nearly impossible for even the rare senior managers who show an interest to understand what is actually happening at the transaction level. Management information systems most often do not report on transaction price performance, or report only average prices and thus shed no real light on pricing opportunities lost transaction by transaction. (Marn and Rosiello 1992, 86)
Without a consistent process of analysis and evaluation, the probability that a particular pocket price maximizes customer profitability is like the probability that a blindfolded dart player will hit a bull’s-eye—not zero but not very high. In fact, the situation can be even worse. Sophisticated buyers often understand a seller’s pricing process better than does the seller. A sophisticated procurement department, faced with a price waterfall such as that shown in Figure 2.1, would quickly learn how to divide and conquer to obtain the lowest pocket price. The buyer’s procurement agent will seek to extract additional concessions from sales, will argue for relief from strict interpretation of volume purchasing agreements, and will seek extended payment terms from accounts payable. Smart buyers will quickly detect a disorganized or dispersed pricing organization and exploit it to their advantage.
Management often concentrates primarily on invoice prices or list prices. However, the price waterfall illustrates that the majority of important pricing adjustments often take place after the list price and even the invoice price. A typical trucking company sells less than 5% of its business at list price—all the rest is sold at a discount. Only a tiny minority of customers purchase a car at the manufacturer’s suggested retail price (MSRP). Yet, in many cases, managers spend long hours arguing over and analyzing list prices, despite the fact that list prices have little or no relationship to what most of their customers will be quoted. As Figure 2.1 shows, even companies that focus on the invoice price are still missing much of the important action. In this figure, six different discounts were granted after the invoice price.
The distribution of pocket-price discounts given by a CPG company to its various customers over a year is shown in Figure 2.2. In this case, 9% of the customers received a discount of greater than 40%, while 16% received discounts between 35% and 40%, and only 3% paid list price. This distribution represents a fairly typical spread of discounts for the CPG industry. Note that the distribution in itself does not tell us anything about the quality of the pricing decisions being made by the company. However, it immediately demonstrates two facts.
• The item being sold is not a commodity. The variation in pocket prices means that customers are willing to pay a range of prices for the item and the company is willing to sell the product at different prices.
• Only 3% of customers bought at list price. For this item, setting the list price is not the critical pricing and revenue optimization (PRO) decision. Rather, list price is being set high and discounts are being used to target prices to individual customers. The key decisions are what discounts to offer each customer.
Figure 2.2 Pocket-price distribution for a consumer package goods company.
Source: Courtesy of Mike Reopel of A. T. Kearney.
It should be stressed that the existence of a pocket-price distribution such as the one shown in Figure 2.2 does not by itself say anything about the quality of the pricing decisions. A company practicing sophisticated pricing optimization may also show a wide distribution of prices. The question is: Is the pocket-price distribution the result of a conscious corporate process based on sound analysis, or is it the result of an arbitrary process?
A key measure of the quality of PRO decision making is the extent to which the pocket price correlates with customer characteristics that are indicators of price sensitivity. In many cases, large customers are more price sensitive than smaller customers. For this reason, many companies offer higher discounts to their larger customers. If customer size is determining discounts, a seller would expect discount levels as a function of customer size to lie within a band something like that shown in panel A of Figure 2.3. The actual mapping of discount level against customer size for the CPG company is shown in panel B of Figure 2.3. The correlation of discount to customer size was only about 0.09 for this company—statistically indistinguishable from random. This means that something other than customer size is determining the level of discounts.
Figure 2.3 Correlation of discount with customer size—consumer package goods example. Panel A shows management expectations; panel B shows the actual distribution. Source: Courtesy of Mike Reopel of A. T. Kearney.
Graphics such as the price waterfall in Figure 2.1 and the pocket-price distribution histogram in Figure 2.2 are useful to help companies assess the current state of their pricing. The price waterfall can identify how pricing responsibilities are dispersed across the organization and where sophisticated buyers may be utilizing divide-and-conquer techniques to win higher discounts. The pocket-price histogram can give an idea of the breadth of discounts being given to customers. Graphs such as those shown in Figure 2.3 can show the extent to which discounts correlate to customer characteristics such as size of customer, size of account, and mix of customer business. These graphs often form part of a preliminary diagnostic analysis and can be used to illustrate the need for better pricing decisions.
2.2 TRADITIONAL APPROACHES TO PRICING
Pricing and revenue optimization incorporates costs, customer demand (or willingness to pay), and the competitive environment to determine the prices that maximize expected net contribution. Other approaches to pricing tend to weigh one of these three aspects more than the others, as shown in Table 2.1. Cost-plus pricing calculates prices based on cost plus a standard margin. Market-based pricing bases prices on competitors’ prices. Value-based pricing bases prices on estimates of how customers value the good or service being sold. The finance department tends to like cost-based pricing because it guarantees that each sale produces an adequate margin, which seems fiscally prudent. The sales department tends to like market-based pricing because it helps them sell against competition. The marketing department is often the natural supporter of value-based pricing.
TABLE 2.1
Alternative approaches to pricing
Each of these purist approaches has its strengths and weaknesses, which are discussed in the following sections.
2.2.1 Cost-Based Pricing
Cost-based pricing, also known as cost-plus pricing, is perhaps the oldest approach to setting prices and still one of the most popular. It has a compelling simplicity—determine the cost of each product and add a percentage surcharge to determine price. The surcharge is often calculated to reflect an allocation of fixed costs plus a required return on capital. Alternatively, it may simply be based on tradition or a rule of thumb. For example, an old rule of thumb in the restaurant industry is “food is marked up three times direct costs, beer four times, and liquor six times” (Godin and Conley 1987, 58).
The cost-plus pricing approach appears to be objective and defensible. If all competitors in a market have similar cost structures, it would appear to be a reasonable way to ensure consistency with the competition. In addition, cost-based pricing gives the appearance of financial prudence. After all, if all of our products are priced with the right surcharge, the company is guaranteed to make back the cost of production plus fixed costs plus the required return on capital. Everybody, including the shareholders, should be happy. It is not surprising that cost-based pricing, with its dual appeals to objectivity and financial prudence, often appeals to finance departments.
The major drawback of cost-plus pricing is widely recognized: It is an entirely inward-focused exercise that has nothing to do with the market. Calculating prices without any reference to what customers might (or might not) being willing to pay for your product is an obvious folly: Customers do not care about your costs—they care about what they have to pay for the product. Furthermore, cost-based pricing does not support price differentiation—the ability to charge different prices to different customer segments—which is at the heart of pricing and revenue optimization.
Another problem with cost-plus pricing is that the costs used as its basis are often nowhere near as objective as they seem (and as the finance department may believe them to be). The calculation of the variable and fixed costs involved in the production of a complex slate of products involves innumerable subjective judgments. Furthermore, all of the hard cost numbers available to the organization are based on historical performance—production costs in the future may be widely different as the mix of business changes and as production efficiency changes. Basing pricing decisions strictly on costs plus a surcharge can lead to highly distorted prices driving lower-than-expected results. This can yield results that seem puzzling, since the prices were based on seemingly objective costs.1
Given these drawbacks, it should not be surprising that experts are uniformly harsh toward cost-plus pricing.
• “The problem with cost-driven pricing is fundamental” (Nagle and Müller 2018, 4).
• “Cost-plus pricing is not an acceptable method” (Dolan and Simon 1996, 38).
• “Does the use of a rigid, customary markup over cost make logical sense in the pricing of products? Generally, the answer is no” (Lilien, Kotler, and Moorthy 1992, 207).
Despite nearly uniform condemnation, cost-based pricing is surprisingly resilient. As one author put it, “Cost-plus pricing is a lot like the romance novel genre, in that it’s widely ridiculed yet tremendously popular” (Dholakia 2018). In fact, surveys consistently find that cost-plus pricing is the most widely used approach to pricing across different countries and industries. For example, a 2009 survey across multiple industries in the United States, India, and Singapore found that “the most frequently used strategy was cost-plus pricing (47.2 percent of firms)” (Rao and Kartono 2009, 17). A 2010 survey of British industry found that 44% of respondents priced all or some of their products using cost-plus pricing (Greenslade and Parker 2012).2 The only good news about the persistence of cost-based pricing is that it means there is substantial opportunity for improvement.
2.2.2 Market-Based Pricing
Market-based pricing means different things in different contexts. In this book it refers to the practice of pricing based solely on the prices being offered by the competition. It is commonly applied by smaller players in situations in which there is a clear market leader—for example, a small cola brand might set its price based on the price of Coca-Cola. A survey found that 31.7% of firms in the United States, Singapore, and India “match the price set by the overall market or price leader” (Rao and Kartono 2009, 20). Market-based pricing is, of course, also the practice in pure commodity markets, such as bulk chemicals, in which offerings are completely identical and transaction prices are rapidly, perfectly communicated. In these cases, there is no pricing decision per se—all companies take the price as given and adjust their production accordingly. For a commodity, there is no alternative to market-based pricing.
Market-based pricing can also be an effective strategy for a low-cost supplier seeking to enter a new market. For example, Alamo Car Rental started as a low-cost rental car company targeting the price-sensitive leisure market. Alamo’s initial strategy was to ensure it was always priced at least $1.00 lower than both Hertz and Avis on the reservation system displays used by travel agents. This strategy was effective at meeting the strategic goal of rapid growth and penetration of the leisure market. Similarly, the European airline Ryanair has a strategy of having the lowest fare on the market compared to all competitors serving the same route within two hours of departure (Eisenaecher 2019).
While market-based pricing is appropriate in certain cases—in a commodity market, for small players in a market dominated by a large competitor, and as a way to drive market share—it is often used in cases where it is less appropriate. At its most extreme, it means letting the competition set our prices. Slavishly following competitive prices does not allow us to capitalize on the changing value perceptions of customers in the marketplace. Furthermore, it does not allow us to capitalize on the differential perception that customers hold of us versus the competition. We should charge a higher price to customers who value our product or brand more highly. Monitoring competitive prices and making sure we maintain a realistic pricing relationship with key competitors is always important—but we also need to adjust our position relative to our competitors to reflect current market conditions if we want to maximize profitability.
If all of the competitors in a market use market pricing, it is not clear where the price actually comes from and pathological effects can occur. One of these pathological effects was uncovered on April 18, 2010, when a postdoctoral student working in Michael Eisen’s lab at the University of California, Berkeley, searched on Amazon for a book titled The Making of a Fly: The Genetics of Animal Design, by Peter Lawrence. He was surprised to find that the book was being offered new by two different independent book sellers on the Amazon marketplace at prices of $18,651,718.08 by Profnath and $23,698,655.93 by Bordeebook. Adding insult to injury, both sellers asked an additional $3.00 for shipping.
Digging in a bit, Eisen found that both sellers were using pure market pricing—that is, each seller was basing its price entirely on the price being offered by the other. Profnath was setting its price at 99.83% of Bordee’s price—presumably to be priced just below its rival. However, Bordee was setting its price 27% higher than Profnath. In the absence of orders, the price began to spiral upward. Assuming that Profnath originally priced the book at its list price of $70, Bordee would respond with a price of $88.90. Profnath would then slightly undercut the price at $88.75, and Bordee would respond with a price of $112.71, which Profnath would undercut at $112.52. The power of geometric growth meant it would take only about 53 of these cycles to arrive at prices over $18 million.3
It seems clear that the process that produced the exorbitant book pricing was the result of algorithmic pricing on the part of both sellers using extremely simplistic market-based pricing rules. One lesson is that algorithmic pricing requires sensible guardrails to avoid having prices wander into unacceptable values. However, the more important lesson is that setting prices based entirely on competitor prices can easily lose touch with the marketplace, particularly if all the competitors are also basing their prices on competition. Like the other pure pricing tactics, market-based pricing does not generally lead to an optimal price.
2.2.3 Value-Based Pricing
Like market-based pricing, value-based pricing (or value pricing) means different things in different contexts. In its broadest sense it refers to the unexceptional proposition that price should relate to customer value. In its narrowest sense, it is sometimes used as a synonym for personalized or one-on-one pricing, in which each customer is quoted a different price based on her value for the product being sold. In this book it refers to the belief that customer value should be the key driver of price.4 Historically, value-based pricing usually referred to the use of methodologies such as customer surveys, focus groups, and conjoint analysis to estimate how customers value a product relative to the alternatives, which is then used to determine price. This type of value-based pricing is employed most frequently for consumer goods—especially when a new product is being introduced. In a 2003 survey, 52% of companies responding from the United States claimed that they “price this product based on our customers’ perceptions of the product’s value” (Rao and Kartono 2009, 15).
There is nothing wrong with the basic idea behind value-based pricing. If a seller faces no competition and the seller can determine the value each customer places on his product and charge that value (assuming it is higher than incremental cost) and the seller does not need to worry about arbitrage or cannibalization (or about being regulated), then that is what he should do to maximize profit.5 This pure approach to pricing has one serious drawback: it is impossible. For one thing, there is no way to discern individual customer value for a product at the point of sale. In addition, the possibility of arbitrage and cannibalization almost always limits the ability to charge different prices for different products. Finally, competitive pressure means that companies almost always have to set prices lower than they would prefer to any group of customers.
The competitive restriction on value-based pricing is worth emphasizing. There is a great difference between the value that a potential buyer might place on our product in isolation and what we can actually get that customer to pay in a competitive market. A customer may value our product or services highly, but she also almost always has alternatives. For example, management consulting firms routinely discuss changing from cost-based pricing (hours worked times rate per hour) to value-based pricing, under the belief that “a good consultant could boost earnings using a value-based model” (McLaughlin 2002, 1). Yet less than 5% of consultants use value-based pricing. Why should this be? Consider a brilliant management consulting organization that can provide services to a client that will lead to improved profitability of $2 million yet only cost $500,000 to provide. If the consulting company has a true monopoly, it should be able to close the deal at $1.95 million, leaving the client $50,000 ahead. But what if there is another, slightly less brilliant consulting company with the same cost structure that can provide similar services that would lead to improved profitability of only $1.5 million? That company could counterpropose a project at $1.4 million, which would be a better deal for the client, who would be $100,000 ahead. The upshot is that competition can severely restrict the ability of a company to set a price based on value, even when the competition is offering an inferior product. Even the existence of an inferior substitute will mean that a company cannot charge full value.
2.2.4 Summary of the Traditional Approaches
Cost-based pricing, market-based pricing, and value-based pricing are pure pricing approaches—each with serious drawbacks. In reality, most companies are not purists. While a company is likely to have a dominant philosophy, very few companies use a single approach 100% of the time, and companies will modify their approaches to achieve different goals. When Xerox wanted to increase market share, it would place the pricing function in the sales organization—prices would drop, and sales would soar, but unit margins would drop. When Xerox wanted to increase profits, it would move the pricing function into the finance department—prices would increase, as would profits, although sales would drop (Eric Mitchell, pers. comm., 2003). Other companies are less disciplined—their approach to pricing may change with the flavor of the month: market based when the emphasis is on market share, value based when focusing on the customer comes into vogue, and cost based when profits are required.
Vacillating among pricing approaches may actually be better than strict devotion to one approach. Any company that sticks tenaciously to any one of the three pure approaches would likely soon find itself in deep trouble. What is often seen in reality is a hybrid—companies espouse a particular philosophy but use pieces of all three, supplemented by a considerable amount of improvisation. The upshot is pricing confusion; rarely is a consistent justification or approach applied across all pricing decisions. One of the fundamental ideas behind pricing and revenue optimization is that a company should clearly articulate its goal—that is, what it wants to achieve from pricing. Finally, pricing and revenue optimization as described in this book involves combining costs, competitive pricing, and customer value information in determining the price to offer. In that sense, it represents a blend or balance of all three approaches.
2.3 THE SCOPE OF PRICING AND REVENUE OPTIMIZATION
One of the goals of pricing and revenue optimization is to provide a consistent approach to pricing decisions across the organization. This means that a company needs to have a clear view of all the prices it is setting in the marketplace and the ways in which those prices are set.
2.3.1 The PRO Cube
The goal of pricing and revenue optimization is to provide the right price:
• For every product
• To every customer segment
• Through every channel
In addition, the goal is to update those prices over time in response to changing market conditions. The scope of pricing and revenue management is defined by these three dimensions, which can be represented as a cube, as shown in Figure 2.4. Each element within the cube represents a combination of product, channel, and customer segment. Each element (or cell) has an associated price. For example, one element might be
Medium-size turbines sold to large customers in the Northeast via the direct sales channel
Another element might be
Replacement gears sold to small companies via online sales
In theory, a different price might be associated with each cell within the PRO cube. In practice, some cells may not be meaningful. Some products may not be offered through some channels, for example. It is also the case that the prices within the PRO cube may not always be independent of each other. Our ability (or desire) to charge different prices through different channels may be constrained either by practical considerations or by strategic goals. If we want to encourage small customers to purchase online rather than through our direct sales channel, we might institute a constraint that the online price for small customers for all products must be less than or equal to the direct sales channel price. Considerations such as these need to be incorporated in the business rules applied to the pricing and revenue optimization process as described in Section 2.4.
Figure 2.4 Dimensions of the pricing and revenue optimization cube.
Companies may offer even more prices than the PRO cube would imply. Certain products might be subject to tiered pricing or volume discounts. Or a company might offer bundles of products at different promotional rates or discounts that are not available for individual products. Each of these bundles or quantity combinations can be treated as an additional, virtual product in the PRO cube.
The PRO cube is a useful starting point for a company seeking to understand the magnitude of the pricing challenge that it faces. Enumerating the combinations of products, market segments, and channels gives a rough estimate of the total number of prices a company needs to manage. Many companies are astonished at the sheer volume of prices they already offer. For example, a bank offered home equity loans with rates that could vary by any combination of credit band, loan size, term, geography, and loan-to-value ratio. When all the possible combinations were enumerated, the bank was surprised to discover that it actually had more than 2 million prices in play for home equity loans at any one time. Such astonishment is often the first step in realizing the need to establish a consistent pricing and revenue optimization process.
2.3.2 Customer Commitments
A core concept in pricing and revenue optimization is the idea of a customer commitment. Specifically, a customer commitment occurs whenever a seller agrees to provide a customer with products or services, now or in the future, at a price. The elements of a customer commitment include:
• The products and services being offered
• The price
• The time period over which the commitment will be delivered
• The time for which the offered commitment is valid—that is, how long the customer has to make up her mind
• Other elements of the contract or transaction (e.g., payment terms, return policy)
• Firmness of the commitment and risk sharing, including contingencies that might change the price paid by the customer
Some examples of customer commitments include the following:
• A list price is possibly the most familiar form of customer commitment. A list price is a commitment that a buyer can obtain the item by paying the posted price. List price is often (but not always) a nonnegotiable or take-it-or-leave-it price. Common examples are the shelf prices at the grocery store and drugstore, price per gallon displayed on the gasoline pump, and most online retail prices.
• Coupons and other types of promotion are also customer commitments. They allow certain customers to obtain a good at a price lower than list for some period.
• In a business-to-business setting, prices for large purchases are often individually negotiated. The seller may have published list prices, but in many industries items are rarely, if ever, purchased at the list price. Rather, each buyer negotiates an individual discount. The discount that a particular buyer receives can depend on the buyer’s purchasing history with the seller, the skills of the purchasing agent and the salesperson, and the desire of the seller to make the sale. Chapter 13 describes an approach to optimizing such customized prices.
• In other business-to-business settings, the buyer and seller negotiate a contract that establishes prices that will be in place for six months, a year, or longer. For example, in less-than-truckload (LTL) freight, shippers agree on a schedule of tariffs that will govern all the shipments they will make over the next year. These schedules are usually expressed in terms of a discount from a standard list tariff. Contracts are individually negotiated between each shipper and each carrier. Contracts may be exclusive (as they often are in package express), or a shipper may establish contracts with two or more carriers, as is common practice in trucking and container shipping.
• Contracts between electronics distributors, such as Ingram Micro and Tech Data, and wholesalers, such as Hewlett-Packard and Sun Microsystems, are often on a tiered-pricing basis, where each tier includes a volume target and an incremental discount for hitting that target. An example is shown in Table 2.2. In this case, the purchaser will receive at least a 5% discount on each unit purchased during the current quarter. If she purchases at least 50,001 units, the discount increases to 6%. And if she purchases 80,001 units, the discount increases to 7%. This type of tiered pricing provides an incentive for the distributor to push the seller’s products rather than those of its competitors. Tiered discounts can be based on either units sold or revenue. The increased discounts themselves can apply either to incremental sales above the tier or to all sales during the quarter. We examine approaches to setting tiered prices in Section 6.6 on nonlinear pricing.
• Online auction houses such as eBay enable a seller to establish a reserve price for an item she wants to sell. On eBay, the highest bidder above the reserve price will be sold the item. The commitment by the seller is to honor this policy. In addition, the seller may establish a buy-now price that allows immediate purchase. Auctions, either online or offline, are also common for many types of used goods. In an auction setting, the pricing decision on the part of the seller is the reserve and the buy-now price, if applicable.
TABLE 2.2
Tiered pricing
The forms and types of commitments that sellers make (and buyers expect) vary from industry to industry. As a simple example, airline tickets were historically refundable, meaning that the airline took all the risk of a customer not showing up—a no-show—a risk the airlines sought to mitigate via overbooking. On the other hand, tickets to Broadway shows have historically been nonrefundable, meaning the customer takes on the risk of her no-show decision. As a result, Broadway theaters do not overbook. The details of the types of commitments made in different industries are often the result of complex and contingent historical factors and may change over time—just as airlines are increasingly beginning to adopt nonrefundable ticketing.
In fact, it is not uncommon for individual sellers to utilize different approaches for different products through different channels. For example, an automobile manufacturer is likely to sell the majority of vehicles through dealers, using a combination of list price (in this case, the wholesale price) and promotions. However, General Motors Corporation receives about 11% of its revenue in North America from group sales of cars to corporate, government, and commercial fleets. These sales are generally priced through individual negotiations or bids. Finally, a manufacturer may also be selling some cars directly to the public through a deal with an online channel. In this case, the manufacturer needs to apply pricing and revenue optimization across the entire range of its channels and pricing mechanisms in order to maximize total profitability.
2.4 THE PRICING AND REVENUE OPTIMIZATION PROCESS
Successful pricing and revenue optimization involves two components:
• A consistent business process and organization focused on pricing as a critical set of decisions
• The software and analytical capabilities required to support the process
Figure 2.5 The pricing and revenue optimization process.
Much of the interest in pricing and revenue optimization has focused on its use of mathematical analysis. Quantitative analysis is indeed critical in most pricing and revenue optimization settings, but it cannot provide sustainable improvement unless it is embedded in the right process. A diagram of a closed-loop pricing process is shown in Figure 2.5. In this figure, the overall pricing and revenue optimization process has been divided into eight activities. Four of these activities are part of operational PRO—that is, they are executed every time the company needs to change the prices it is offering in the marketplace. The remaining four activities—supporting PRO activities—occur at longer intervals. As the figure shows, pricing and revenue optimization is ideally a closed-loop process; that is, feedback from the market must be incorporated into both the operational activities and the more periodic activity of updating market-response functions. Each of the individual activities is discussed below.
2.4.1 Operational PRO Activities
Operational PRO activities work continuously to set and update prices in the marketplace. The timing of the decisions will depend on the application—airline revenue management systems and online retailers can change prices from moment to moment. Other companies may update prices weekly or monthly. Independent of the speed at which prices are changed, the core operational activities will be similar.
Analyze alternatives. This activity is the one most widely identified with pricing and revenue optimization or revenue management. For many companies, it has historically involved the use of spreadsheets to compare pricing alternatives under different scenarios.
Choose the best alternative. The choice of the best alternative requires a clear understanding of the goal for each price in the PRO cube—is it to maximize revenue or maximize price or some combination of the two? When many prices are changing quickly, an algorithm is typically required to solve the underlying optimization problem.
Execute pricing. Finally, the prices that have been calculated need to be communicated to the market. This step is also referred to as pricing execution. The mechanics of price transmission and distribution vary greatly from industry to industry, company to company, and even channel to channel within the same company. For an airline or hotel, prices are updated by opening or closing fare classes—a process discussed in more detail in Chapter 8. In other cases, a pricing matrix in a database specifies which prices are available to which customers through which channels. In other cases, new discount guidelines are communicated to the sales force by email. For a company with many products being sold through many channels, communicating the correct prices correctly can be a challenge in itself. A number of pricing execution or pricing management vendors, such as Vendavo, sell software whose primary function is to ensure that complex pricing changes are correctly and efficiently transmitted through all channels. In the face of a complex PRO cube with thousands or even millions of prices, ensuring that each price is correctly calculated and implemented can be a daunting task. One of the important functions of pricing execution systems is to prevent price leakage, in which incorrect prices are published or salespeople price outside established guidelines.
Monitor and evaluate performance. As sellers, when we receive results from the marketplace we need to compare the results with expectations and evaluate overall performance against our goals. Are we achieving the lift we anticipated from a promotion? If not, why not? Was total market demand less than expected? Did the competition react more aggressively than we had anticipated? Were customers less responsive to price than we anticipated? Or was it a combination of all three?
Figure 2.5 illustrates that market feedback occurs at two levels. Analysis of alternatives receives the most immediate feedback. Here, the effects of the most recent actions are monitored so that immediate action can be taken if necessary. The required actions may not always be directly related to pricing and revenue optimization. For example, the markdown management system used by a major retailer indicated that a particular style of pants was selling well below expectations. An investigation found that the pants had been hemmed improperly. Once the problem had been corrected, the pants began to sell at the rate initially expected. In the more common case, a big difference between realized sales and expected sales is likely to be due to an unforeseen change in the marketplace.
The second level of feedback updates the parameters of the market-response functions. If sales of some product are slower than expected, it may indicate that the market is more price sensitive than expected and the future market-response function for that product should be adjusted accordingly.
2.4.2 Supporting PRO Activities
The primary role of the supporting activities is to provide key inputs to the operational PRO activities. The supporting activities occur in a much longer time frame. Typically, goals and business rules will change quarterly or less frequently. Market segmentations may be updated annually. Price-response functions will be updated much more frequently, typically weekly or monthly in the case of markdown management or customized pricing.
Set goals and business rules. A key initial step in pricing and revenue optimization is to specify the overall goal of the process. Without a clear goal, it is impossible to make consistent decisions and to evaluate decisions in order to improve the process over time. In general, the goal of pricing and revenue optimization is to maximize expected total contribution. However, at certain times, in certain markets, a company may wish to increase market share or attain a volume sales goal. Whatever the goal, it is most important that it be stated clearly and explicitly. As discussed in Section 2.3, the corporate goal for a product, segment, and channel combination determines the form of the objective function that will be used when optimizing that cell of the PRO cube.
Business rules apply whenever there is a restriction that may prevent us from pricing optimally in some market. Is there a minimum per-unit margin requirement we need to meet for all products within a category? Is there a minimum discount we need to offer to certain strategic customers? Do we want our retail list prices to be no lower than our internet prices? Such rules are sometimes referred to as guardrails. Each business rule needs to be implemented as a constraint in the optimization problem at the heart of PRO.
Example 2.1
A heavy-equipment manufacturer wants to make sure that prices in the Northeast are never more than 15% higher than the comparable model being offered by the leading competitor. This decision results in a constraint to be included in the PRO optimization problem: Our price in the Northeast ≤ 115% competitor A’s price in the Northeast.
Note that business rules are inputs into PRO—that is, they are decisions made not on the basis of short-term optimization but on the basis of other considerations a company wants to incorporate in its pricing and its willingness to accept the resulting reduction in short-run performance.
Segment the market. An important theme within pricing and revenue optimization is segmenting the market in a way that maximizes the opportunity to extract profit. We discuss techniques for doing this in Chapter 6. The market segmentation needs to be updated periodically to reflect changes in the underlying market. However, this updating should be performed much less frequently than the operational PRO processes. For this reason, the “Segment market” step in Figure 2.5 is shown with a dashed line outside the main process loop.
Estimate price response. For each market segment identified we need to estimate the corresponding price-response functions. How to do this is the topic of Chapter 4.
Update price response. It is not sufficient merely to monitor the performance of PRO actions against expectations. A company needs to update its models to incorporate what it has learned. As long as performance closely matches expectations, the parameters of the models need not be changed. However, if performance is significantly different from expectations, models need to be updated to reflect the new information.
The most critical message to take away from the pricing and revenue optimization process illustrated in Figure 2.5 is that PRO should be treated like any other critical business process. For a seller to make effective PRO decisions, he needs to clearly identify what he is trying to achieve, the constraints he is facing, and the alternatives available. Based on his understanding of the market and the constraints, he should choose the alternative most likely to achieve his goals. Once this alternative is implemented, he should monitor and measure the results against expectations and update his understanding of the market so that he can make better decisions in the future. The PRO organization, responsibilities, systems, and data-capture methodologies should all be designed to support this process.
We must recognize that however prices are analyzed and calculated, a human being ultimately needs to be responsible for the prices in the market. This does not mean that every price needs to be inspected individually by a human being—a practical impossibility when thousands or millions of prices are changing each day. However, it does mean that prices need to be monitored with appropriate high-level metrics such as the distribution of contribution margin and the distribution of prices relative to competition. Extreme deviations—either high or low—should be audited to ensure they are consistent with the pricing methodology and not the result of calculation errors or bad data.
The PRO process presented in this section is highly general. We return to it in later chapters as we look at how pricing and revenue optimization is applied in specific situations, such as revenue management, customized pricing, and markdown management.
2.4.3 The Time Dimension
Much of the growing interest in pricing and revenue optimization is being driven by the increasing velocity of pricing decisions. Prices in the past changed once a quarter for many industries; now they change once a week, once a day, or even once a minute. Increased frequency and complexity of pricing creates its own momentum. A company that is able to change its prices more rapidly in response to changing market conditions can gain an advantage over its more slow-footed competitors. This creates a strong incentive for other companies to match (or exceed) the frequency of change. E-commerce has been a proven driver of increased price velocity. As discussed in Chapter 1, the adoption of computerized distribution systems—an early form of e-commerce—by the passenger airlines was a direct contributor to the ability of the airlines to develop, manage, and update highly complex fare structures.
Pricing follows different patterns in different industries. The rack (wholesale) price for gasoline fluctuates randomly, largely following the crude oil price, with little obvious trend. Fashion goods are priced high at the beginning of the season and are then subsequently marked down as the season progresses. By contrast, airline prices tend to rise as departure approaches. Different pricing approaches need to be used in each of these markets. Gasoline rack prices are set using dynamic pricing approaches that primarily track changes in the underlying supply and demand imbalance. When supply is low (e.g., Iraqi oil production is interrupted) or demand is high (e.g., the Northeast has a cold winter), the price tends to rise. The price of a fashion item falls across the season, since such items become less valuable as the season progresses. On the other hand, the passenger airlines have successfully segmented their customer base into early-booking leisure customers and later-booking business customers. To support this segmentation, ticket prices rise as departure approaches. These different patterns lead to very different optimization approaches in different markets.
Frequent price changes create a challenge for an organization. As price changes become more frequent and the pricing structure becomes more complex, it becomes less and less feasible for a company to apply an in-depth analysis to each pricing decision. The spreadsheet analysis method of making pricing decisions begins to break down. As the complexity and rapidity of pricing decisions continues to accelerate, many organizations become overwhelmed. A common symptom is that despite expending more and more effort on pricing decisions, the company continually seems to be one step behind the market and the competition. At this point, most companies need to adopt a computerized pricing support system.
2.4.4 The Role of Optimization
As the name implies, optimization plays a central role in pricing and revenue optimization. As we see in Chapter 5, formulating and solving pricing decisions as constrained optimization problems is at the heart of pricing and revenue optimization. The formulation and solution of these constrained optimization problems draws on techniques from statistics, operations research, management science, and machine learning.
While constrained optimization is at the heart of pricing and revenue optimization, it is also the case that no company in the world actually optimizes its prices. The reason is that determining optimal prices is, in general, impossible—or at least well beyond our current ability. Many pricing and revenue optimization approaches (including those we study) are based on solving stylized representations of the underlying problem. These representations include many of the important features of the real-world problem, but they exclude others. For example, many hotel companies have become reasonably proficient at pricing and revenue optimization. Yet the following is a list of some of the factors often not incorporated in hotel pricing and revenue optimization decisions:
• How the price we offer a potential customer now affects her propensity to consider this property (or this chain) in the future
• How group prices should be optimized to trade off the amount received from each group with the number of rooms they will take, considering that we may need to refuse bookings to future independently booking customers
• How the probability that this particular customer will cancel, understay (i.e., check out early), not show, or overstay (i.e., check out late) should influence the price we quote her
• How the price we quote a customer will affect her propensity to overstay or understay
• How the price we quote a customer should be influenced by the prices currently being offered by major competitors in this market
• How the price we display in the system might influence competitors to change their prices
• How a group of properties serving a single location (e.g., all the Marriotts in Manhattan) should be priced to jointly optimize use of all the capacity
• How the expected booking order of future customers by length of stay affects the optimal price to offer this customer
Each of these and many other influences is either often not incorporated or is incorporated in a very basic fashion in most hotel pricing and revenue management systems. The point is not that hotels are poor at pricing and revenue optimization. On the contrary, the industry has become increasingly sophisticated over the past two decades. The point is, rather, that hotels have been able to improve their pricing significantly without becoming perfect.
Every industry that uses pricing and revenue optimization relies on a simplified representation of the underlying problem. The only justification for doing this is that it works. By capturing 75% or so of the real-world complexity, mathematical analysis almost always does better than either unaided human judgment or any of the purist approaches to pricing described in Section 2.2. Of course, companies and vendors continue to invest to improve the sophistication of their pricing and revenue optimization systems. Over time, systems capture more and more aspects of the real world, and the quality of pricing decisions improves. However, it is unlikely that any system or approach will ever generate perfect recommendations. What pricing and revenue optimization really does is use quantitative analysis to make better pricing decisions more quickly. Part of the philosophy of pricing and revenue optimization is that good prices on time are far better than perfect prices late. The improvement realized on any individual pricing decision may be small, but the aggregate effect over hundreds, thousands, or millions of pricing decisions can be very large indeed.
An alternative approach to modeling demand and explicitly optimizing price (model-based optimization) is to use a test-and-learn, or learning-and-earning, approach. Under this approach, different prices are tested under different market conditions, and the prices that lead to better outcomes are chosen more often in the future when those (or similar) conditions recur. This approach has the advantage that it does not require explicit modeling of the relationship between price and demand. It has the drawback that it requires a considerable amount of testing to reach acceptable performance. For many companies, price testing may be difficult or impossible. Furthermore, when it is possible, it may require a considerable amount of testing suboptimal prices to find the optimal price, which can be costly. Finally, test-and-learn approaches are no better than explicit modeling in determining the longer-term implications of pricing decisions.
In this book, we discuss both model-based optimization and test-and-learn approaches to pricing and revenue management. Both have their place. Furthermore, it is likely that the most successful next-generation approaches will be hybrids of model-based and test-and-learn approaches. Specifically, such approaches will perform price testing and update prices guided by parsimonious but realistic assumptions about the underlying structure of customer response. This is, for example, the basis for the data-driven approach to pricing described in Section 5.5.
Finally, I should reemphasize the importance of the feedback loop in Figure 2.5. Pricing and revenue optimization is dependent on rapid and effective updating and monitoring more than almost anything else. The faster and more effectively prices can be adjusted to account for what is currently happening in a market, the better the business results that will be generated.
2.5 SUMMARY
• For many companies, pricing involves a complex set of decisions that is often poorly managed or unmanaged. In many cases, the pocket prices charged to customers are the result of a large number of uncoordinated and arbitrary decisions. The price waterfall and pocket-price histograms can be useful ways of visualizing the current state of pricing management within a company.
• Traditional approaches to pricing include cost-plus, market based, and value based. Of these, cost-plus is the most common. Each pure approach ignores important aspects of the pricing problem. Most companies are not purists but rely on some combination of the approaches, with the emphasis changing over time.
• The scope of pricing and revenue optimization is to set and update the prices for each combination of product, customer segment, and channel—that is, each cell in the PRO cube.
• The approach used for pricing and revenue optimization must be tailored to the specifics of the underlying market structure. From the point of view of the seller, the key concept is what type of customer commitment is required in each market. Different types of customer commitments require different analytical approaches.
• Pricing and revenue optimization requires a consistent process for decision making, evaluation, and updating. A typical process is illustrated in Figure 2.5. This process includes operational activities that involve setting and updating prices in response to market and cost changes, as well as supporting activities that provide input to the operational activities.
• Many of the approaches to calculating optimal prices in this book are based on a classic forecast-and-optimize approach in which we derive an explicit model of how we believe that demand will respond to the prices we offer and use the techniques of mathematical optimization to find the prices that best meet our objective function. This approach has the disadvantage that it requires choice of a price-response model and explicit estimation of the parameters—if the wrong price-response model is chosen, then the resulting prices will be less than optimal. Test-and-learn approaches have the appealing quality that they are nonparametric—they do not require the choice of a demand model beforehand. However, they have drawbacks as well. Both types of model are described throughout the book.
2.6 FURTHER READING
The different examples of customer commitments listed in Section 2.3.2 are examples of pricing modalities. A discussion of pricing modalities and why pricing modalities are different in different industries can be found in Phillips 2012c. Examples of pricing modalities in different industries and how prices are set and communicated in different industries can be found in Özer and Phillips 2012, chaps. 3 –16.
Some of the common deficiencies in pricing organizations and the challenges of building an effective pricing organization are discussed in Tohamy and Keltz 2008 and in Simonetto et al. 2012. Cudahy et al. 2012 provides more detail on pricing as a closed-loop process.
2.7 EXERCISE
Brain and Company is a consulting group that offers a foolproof pricing and revenue optimization service guaranteed to deliver $2 million in benefit to a customer. It costs Brain and Company $500,000 to provide this service. Unfortunately, Brain has a competitor, Dissenture, that provides a similar, but somewhat inferior, service that only delivers $1.5 million in guaranteed benefit. It also costs Dissenture $500,000 to provide this service.
a. When competing with Dissenture, what price does Brain and Company need to charge to guarantee that it wins the business? Assume that neither Dissenture nor Brain will price below cost and that both of them know each other’s costs and the customer benefits in each case.
b. How would Brain’s price need to change if it only cost Dissenture $400,000 to provide its service?
NOTES
1. Several real-world examples of this effect are described in Cooper and Kaplan 1987.
2. In a 2003 survey, 43.7% of responding companies answered that following the rule “Price is made up of direct cost per unit plus a fixed percentage mark-up” is either “Important” or “Very Important” in setting prices (Rao and Kartono 2009).
3. One possible explanation for the policies being used by the two sellers is that Bordee did not have the book in stock but, if it received an order, it planned to purchase a copy from Profnath, while Profnath had the book in stock and just wanted to slightly undercut Bordee. More details on the multimillion-dollar biology book can be found in Eisen 2011.
4. In the software industry, value-based pricing often refers to pricing policies that are based on usage or metrics other than the type of computer or the number of users. This practice was pioneered by the software vendor PeopleSoft, which priced its Human Resources module based on the number of employees in the licensed company and its Financial module based on the annual revenue of the licensee (Welch 2002). The goal of such approaches is to segment the market according to value in order to capture higher license fees from high-value customers. While PeopleSoft’s experience was widely regarded as successful, other attempts to implement value-based pricing have been less successful, particularly when competition drives fees down.
5. In the economics literature, determining the value that every customer places on a product and then charging them that value is called perfect third-degree price discrimination by a monopolist.