The Old Psychology of Internet Pricing

Yes, the Internet can act as a leveler because it provides a great deal of information. But why do sellers jockey for base-price positioning when full price is what shoppers will sometimes pay? The race to the bottom has been exaggerated. The Internet supports multiple price points.

Approach a merchant in a traditional bazaar or souk anywhere in the Arabic world, and you will be treated as unique—indeed, so much so that you will be quoted an exclusive price. The next customer (perhaps wearing a fancier suit or a more expensive watch) will be quoted a higher price for the same item, and still the next (maybe one who has been visibly prowling the neighboring discounters) a lower price. In bargaining cultures throughout history, a seller's key strategic step has been not the back-and-forth dickering process, but the price from which that process starts—a price that differs for each buyer.

You could call it adaptive pricing. For centuries, the world's best salespeople have achieved status and wealth by learning about their customers. To beat their competitors, these salespeople have had to translate customers' tiny signals—such as clothing, jewelry, tone of voice or body language—into judgments about buying behavior, then adapt their prices to match that individual customer.

Market developments have pushed Western cultures away from these models. In pursuit of operational efficiency and fairness, most bricks-and-mortar stores have moved away from one-to-one selling, instead labeling items with a single price tag. Additionally, as commerce moved to the Internet, e-tailers may have tinkered with the price levels of their bricks-and-mortar counterparts but have rarely questioned the one-price-fits-all philosophy.

Why? The utopian rhetoric that surrounded the beginning of the World Wide Web may help explain. Here's an example from BusinessWeek in 1998: "The Internet is a nearly perfect market because information is instantaneous and buyers can compare the offerings of sellers worldwide. The result is fierce price competition, dwindling product differentiation, and vanishing brand loyalty."¹ The Internet, pundits proposed, would (finally) achieve the economists' ideal of a perfect marketplace.

Figure 1: Price for a purple Fujifilm FinePix Z37 10 MP digital camera varies by 25%

Somehow, rather than being taken as hopeful postulates, such notions became pricing axioms. Because Net-based markets were perfect, companies had to lower their prices. Because information was instantaneous, "sales excellence" had to be equated to "price-matching competitors." Because the Internet would enforce a single, static price point for all, sophisticated pricing strategies had to be surrendered. The only path available was a race to the bottom. What if these postulates were wrong? What if the Internet, rather than eliminating the advantages of sophisticated pricing strategies, actually empowered them? What if companies could use the Web to return to the bazaar —and generate huge profits by delighting a larger number of customers?

Look at online stores today: It is clear they can bear multiple price points with large variances (see figure 1). One day in March 2010, major U.S. Internet retailers priced the Fujifilm FinePix Z37 camera at anywhere from $84 to $101 ($85 to $107 after taxes and shipping). This represents price variability of more than 20 percent on the base price and more than 25 percent on the total price. Why do sellers jockey for base-price positioning when full price is what shoppers will sometimes pay?

We can speculate about answers: Perhaps some shoppers are confused by unfamiliar sites or overwhelmed by too many choices. Or perhaps some outlets are hiding their discounts. Regardless of the reason, our point remains: The race to the bottom has been exaggerated. The Internet supports multiple price points.

Adaptive Pricing in the Internet Age

As companies think about pricing in the Internet age, they face a new context. Mountains of data accumulate, and some companies are better than others at exploiting it. The explosion of data eclipses that well-known Internet phenomenon: communities of interest. Certainly, people may join fan clubs to watch old movies or to cook with truffle salt, but film- and salt-sellers no longer need to wait for customers to self-identify. It's just as easy to identify such potential customers through their "data exhaust" (the trail left by their online interactions). In the most obvious example, they may previously have rented old movies or bought truffle salt. That information—once unavailable or incomprehensible—can provide great value in targeting consumers.

It's not just about identifying potential customers, though. It's also about adapting your sales techniques to them. Imagine if each of your retail outlets had a camera at the door that could take a picture of every entering customer, use the picture automatically to make judgments about their expected buying behavior and then instantaneously change every price tag in the store. That, effectively, is what the Internet can do: enable adaptive pricing. Empowered by Web analytic solutions, sellers can learn a great deal about potential customers (see figure 2). Data can help them predict the likelihood of a sale based on factors such as a customer's zip code, the search activity that brought the customer to the site or a combination of the two.

Which Model Is Right for Me?

Many of today's successful companies use some degree of adaptive pricing. We can characterize their strategies in two dimensions: pricing dynamics and customer micro-segmentation (see figure 3).²

Figure 2: Web analytic data can serve as the basis for adaptive pricing

First, let's look at pricing dynamics, the x-axis in figure 3. Basically, this refers to changing the price of a product over time. Today, most companies link their pricing models to product life cycles in a predictable pattern: They mark down prices according to a fairly static schedule, regardless of product movement. They often do so using leading price-setting methods (such as "cost plus" or "price leadership in group of peers") that are simple to implement but generally unprofitable in the long run. More sophisticated pricing dynamics also consider product availability, a technique perfected by airlines over the past 15 years, in which seat prices are raised as the plane fills. Similarly, sophisticated e-tailers have blended forecasts of consumer demand and product inventory to achieve better promotional planning. The most sophisticated level of pricing dynamics, however, taps into the agility of the Internet to gain a real-time demand estimate.

Some companies are introducing real-time pricing for everyday items. For example, the Chicago-based utility ComEd has introduced a residential real-time pricing program that changes the price a homeowner pays for electricity from hour to hour, based on the wholesale price paid by the company. As consumers increasingly become exposed to such real-time pricing approaches, we can expect companies to warm to them.

The other dimension of adaptive pricing is segmentation by customer type. Again, its basic form has a long history. For example: Seniors pay 30 percent less for movie tickets, and corporate customers with negotiated rates pay 25 percent less for hotel rooms. Today the Internet can provide a wealth of potential pricing criteria to micro-segment customers into thousands of small islands.

What does this look like in real life? Consider this example documented in our research: One customer (whom we will call Mindy) types "discount car rentals" into a search engine and clicks on the top result, www.rentalcarmomma.com. From that site Mindy clicks on a coupon that takes her to one of the major car rental companies, but logs off without taking any action. The following week, she goes back to the rental company site (directly, without using the coupon), enters her time and place criteria, and is quoted a price of $300. Another customer (call him Matt), operating in the same time frame, goes directly to the rental company site without using any coupon. He similarly logs off and returns the following week, enters the same criteria Mindy did, and gets a price of $500.

Why? Because based on her browsing history, the company has identified Mindy as a price-sensitive shopper. She came from a site that offers deals, and likely would not have engaged with this company if it offered a flat price of $500 (or even $400). So it offered a targeted discount to expand its customer base.

Figure 3: Two-dimensional adaptive pricing strategies

In our experience, however, most companies in a variety of industries don't work this hard. They may do a little bit of this kind of customer segmentation, and a little bit of pricing dynamics—but they don't coordinate or manage the approach. Indeed, a quick survey of several industries suggests that most languish in the lower-left reaches of figure 3.

There is, therefore, room for a disciplined application of adaptive pricing models, and a large opportunity for first movers in any industry. Although the technique involves a great deal of number crunching, success requires as much art as it does math. Even with software churning through millions of variables, there's significant judgment involved. Given your company's business model, should you focus more on pricing dynamics or customer micro-segmentation? How well can your conversion model predict shoppers' behaviors? What can you change to achieve the greatest increase in conversions? How big a discount do different types of price-sensitive shoppers need? And which prices keep loyal customers loyal, ensuring there is no backlash? (See sidebar: Price Discounts, Not Price Discrimination).

Secrets of the Amazon

Among e-tailers, the company with the most sophisticated adaptive-pricing strategy is Amazon.com. Amazon gains significant competitive advantage from its pricing optimization scheme; for that reason, precise details of Amazon's strategy can be known only to insiders. However, consumer prices are an open window to a business' strategy, and one that deserves a serious look. To get an idea of the sophistication of Amazon's pricing strategy, we collected and analyzed Amazon's prices for bestsellers in the camera and video category every hour for three months. And without revealing any insider secrets, we can say that the picture is breathtaking. Even we were surprised at the extent to which Amazon uses pricing dynamics.

Figure 4: Amazon.com uses pricing dynamics that change by the hour

Amazon changes its prices constantly. For example, over a 72-hour period, half of the top 25 camera and photo products changed prices at least once. Two products had four different prices during that short interval. The number of price changes fluctuated at the level of the stock-keeping unit (SKU), not the category or brand. So did their variance: One SKU varied by 10.4 percent, a remarkably high figure for such a short time period. Another varied by just 0.07 percent, going from $198.86 to $199, as if Amazon was probing the sales effects of rounding cents to dollars.³ (There is even a website, www.thetracktor.com, that tracks the price history of Amazon items.)

Amazon demonstrates surgical precision in adapting prices. Although competing e-tailers such as Best Buy and New Egg also vary prices, these firms tend to keep a price constant for longer periods—usually one week. Amazon sometimes changes prices for a single hour, as shown in figure 4. Clearly Amazon's prices are set by automated pricing engines that experiment continually, aiming to maximize profits by product and category to refine the answer to the question "Are we optimizing the category profitability?"

In addition to such pricing dynamics, Amazon uses product characteristics to help segment its customer base. It sells different colors of the same camera for different (and differentially fluctuating) prices. It may be using odd colors to attract price-conscious consumers, while justifying higher prices for more popular colors. It may even have customer research relating some colors to a willingness to pay higher prices.

Perhaps the most fascinating example of Amazon's pricing agility involves the relationship between an item's price and its position on Amazon's bestseller list (see figure 5). Amazon uses low prices to make a product "hot." Then, as the product gains prominence on the bestseller list, its price is raised to maximize profits. For example, a shopper newly arriving on the site late on January 15 would have seen the Kodak Zi8 HD Pocket Video Camera (Raspberry) ranked #40 on the bestseller list. That shopper would (perhaps subconsciously) see a $220 price tag as justified for such a desirable item—not realizing that Amazon had built that item's popularity with prices of $179 to $200. For two days, Amazon reaped huge margins from such customers. Not all shoppers are equally affected by the bandwagon psychology, however. As a result, over those two days the item dropped in popularity (to the mid-1990s on the bestseller list). So, late on January 17, Amazon resuscitated demand by again dropping the price and repeating the cycle.⁴ With its continual price fluctuations, Amazon emulates the merchants' skills at manipulating buyer psychology.

What's Next?

As online markets grow, companies (perhaps your competitors) will use adaptive pricing to seduce an industry's most profitable customers while still finding a profit-maximizing point for the others. If you don't develop a more sophisticated pricing strategy soon, you will face challenges not only in retaining your best customers, but also in making a sufficient profit from those you do retain. This applies not only to purveyors of today's hot Internet B2C items, but also to automobile manufacturers, financial institutions and telecommunications companies —any industry whose output could conceivably be exchanged through methods that allow sellers to accumulate data they can use to segment customers.

So what's a company to do? We recommend a four-step plan:

1. Recognize the situation. Embrace the need to change—if not to increase profitability, then at least to defend current markets.

Figure 5: Amazon uses shopper psychology (bandwagon effect) to maximize profits

2. Challenge your own pricing-strategy paradigm. What is the best pricing model for your industry and market position? Along which axis do you want to move your adaptive pricing? There is no best pricing model. Rather, among the universe of pricing models, some will be better than others. You need to recognize first the opportunity and establish the goals, starting with understanding your customers. Most companies know this, but some are surprisingly hesitant to embrace data-driven analyses that can enrich customer understanding.

3. Treat customer information as a real asset. Any asset requires investment, maintenance and strategy to maximize its return. Although this article has all sorts of Web analytic data theoretically available to all, not every company has the strong technical architecture (including servers, CPU, memory, bandwidth and algorithms) to receive, store and process the data, then to use them when necessary. You must organize your systems and your people in order to identify the different islands of customers.

4. Continually adapt. Adaptive pricing creates a more complex and faster-changing environment. You must account continually for a variety of factors, especially if you sell in multiple channels. Managing relationships among internal channels and external competitors requires constant attention.

Wisdom of the Ancients

With adaptive pricing, no company will get it right overnight. The shift will not guarantee a huge flow of profits in the next quarterly report. Indeed, because adaptive pricing is all about experimenting, you will make plenty of missteps. (If you don't, you're not experimenting enough.) The point is, selling in the Internet age is not that different from selling anywhere—it requires a customer-centered psychology to develop sophisticated pricing strategies that lead to meaningful profits. Shoppers will pay full price. You just have to give them the option.

Consulting Authors

Stephane Remy is a partner and head of the marketing and sales practice for Europe. Based in the Paris office, he can be reached at This e-mail address is being protected from spambots. You need JavaScript enabled to view it .

Laurent Guerard is a partner and head of the travel, transportation and infrastructure practice for North America. Based in the Chicago office, he can be reached at This e-mail address is being protected from spambots. You need JavaScript enabled to view it .

Rasvan Dirlea is a principal in the automotive, travel, transportation and infrastructure practices. Based in the Detroit office, he can be reached at This e-mail address is being protected from spambots. You need JavaScript enabled to view it .

The authors wish to thank Vincent Cotte, Martin Windle, Ozgun Ataman and Tuvan Sencalis for their valuable contributions in writing this article.

¹ Robert Kuttner, “The Net: A Market too Perfect for Profits,” BusinessWeek, 11 May 1998
² In certain cases, companies add a third layer in the form of a parallel-pricing structure directed at competitive dynamics in pricing. For simplicity's sake we will ignore these dynamics, but they are fully discussed in our article "Covert Pricing: Expanding Profitable Business Behind Enemy lines" in Executive Agenda Volume XII, Number 2.
³ It is also possible that Amazon made frequent microscopic price variations as a way of changing the webpage content to improve search engine rankings for the product.
⁴ We can’t say for sure. First, we checked prices only once an hour, and Amazon may have changed them even more frequently. Second, it’s likely that Amazon could detect this “site visitor” was not an “interested” shopper—and so may have offered us a series of micro-segmented prices.