¡@

In this paper, I develop a profit maximization pricing model named the incentive compatible pricing model that will better off both suppliers and customers for firms that offer digital products. As we will see in this paper, the point of the incentive compatible pricing model is to get the consumers to sort themselves into different product offers according to their willingness to pay. All that the producer needs to do is to choose the product offers and set the prices appropriately so as to induce the consumers to self select into appropriate product offers. I also develop several supporting strategies that will further increase the firm¡¦ s profit at no expense of consumer surplus. These strategies include those that will reduce transaction costs, reduce total first investment costs, enhance consumer¡¥ s swiching costs and strategically use of network effects.

In this thesis, I will develop a profit maximization pricing model named the incentive compatible pricing model that will better off both suppliers and customers for firms that offer digital information products and online services. Besides I will also develop several supporting strategies that will further increase the firm¡¦ s profit at no expense of consumer surplus.

Whinston, Stahl and Choi (1997) think that digital products are goods that can be digitized and transferred over a digital network, such as the Internet. In the electronic marketplace today, digital goods consist primarily of information products. Information products include those products that are first produced in digital format and then printed on physical media, such as books, magazines, newspapers, journals, photographs, maps, sound recordings, and computer games as well as those that are distributed and used in digital format such as databases, television programs and some computer software. In addition to information goods of all kinds, ¡§ concepts¡¨ and various kinds of services all have the potential to become digital products, as Whinston, Stahl and Choi (1997) indicate many products are simply a token or a symbol whose physical form is not an essential requirement, for example, a virtual flower (whose main purpose is symbolic) sent over the network could embody the gesture of greeting, consolation, affection, or any other emotion. Whinston, Stahl and Choi group digital products in three broad categories shown in table 1.

But until now, there is still a lot of confusion about what¡¥ s the better pricing strategy for digital products. Some think that digital products should be paid for by the byte or one piece, some by advertising, some by subscription.

The primary incentive for sellers to differentiate is the reduced substitutability between products as differentiated products become imperfect substitutes for each other (Whinston, Stahl and Choi, 1997), thus, product differentiation gives a firm a certain power within its own market. Although product differentiation is observed in physical markets, it will be more widely practiced in electronic commerce because the transmutability of digital products makes them highly customizable. Futhermore, detailed data on consumer preferences are more abundant in computerized market environments.Besides, the cost structure of digital products involving high fixed costs and low incremental cost is naturally associated with price discrimination (Varian, 1996c), and when we put together the possibility of an ¡§ information relationship¡¨ with the online delivery, we are immediately led to ¡§ mass customization¡¨ of information goods-- there will be several different forms of each information product and consumers can choose the one they prefer (Varian, 1996c). All these factors make product and price discrimination more efficiency and effective in electronic commerce.

In practice, the most significant problem in implementing price discrimination is to determine different consumers¡¥ willingness to pay? One method is to gather more detailed information about customers¡¦ preferences, but consumers usually don¡¥ t willingly reveal their true willingness to pay, thus, pricing needs to based on something else. Here are three such ¡§ something¡¨ :

3. Consider the specific hardware or software(such as browers) that needs to ¡§ view¡¨ the information goods.

4. Goldilocks pricing: In the absence of any additional information having three versions rather than two may be attractive due to ¡§ extremeness aversion¡¨ on the part of consumers.

According to Bakos and Brynjolfsson (1997), various pricing schemes can be thought of as either aggregation or disaggregation along some dimensions. Strategies that aggregate consumer utility across different goods (bundling), different consumers (site licensing), or different time periods (subscription pricing), respectively can be thought of as aggregation strategies, while using micropayments for rental of software ¡§ applets¡¨ or other discrete units of information, can be thought of as disaggregation. They also indicate that reductions in marginal costs made possible by low-cost digital processing and storage of information will favor aggregation, and reductions in distribution costs made possible by ubiquitous networking tend to make disaggregation strategies more profitable.

The key intuition behind the benefit from Bundling is that in many situations, consumers¡¥ valuation for a collection of goods has a probability distribution with a lower standard deviation per good compared to the valuations for the individual goods. That is, the high and low values for individual goods tend to ¡§ average out¡¨ so that consumers¡¦ valuations for the bundle include proportionately more moderate valuations, and the shape of the demand curve changes, flatter (more elastic) in the neighborhood of the mean price and steeper (less elastic) near either extreme. (Bakos and Brynjolfsson, 1997). As Schmalensee (1984) has argued, such a reduction in ¡§ buyer diversity¡¨ typically helps sellers extract higher profits while reducing the deadweight loss from non-zero prices, as more units were sold than if the goods were offered separately.

We can see the benefit of bundling through the following simple example. Assume there are two goods, and suppose that each is valued between zero and one dollar by some consumers, generating linear demand curve as figure 3, and a consumer¡¦ s valuations of each is independent uniform distribution ( figure 1). When the producer sells the goods individually, he gains the profit of (0.5-c)*0.5*2 = 0.5-c (indicated as the shaded area in figure 3). Now if the seller sells the two goods as a bundle, what will happen? First, the probability density function for the consumer valuation for the bundle of two goods is the convolution of the two uniform distributions, which will be shaded like an inverted ¡§ V¡¨ , as in figure 2 (Bakos and Brynjolfsson, 1997). Because the change of the probability density function, the demand curve changes accordingly, as we can see from figure 4, the demand curve is flatter in the middle and steeper in the extreme. Under this demand curve, the seller can gain the profit > 0.5-c (indiicated as the shaded area in figure 4) and can reduce deadweight loss.

¡@

The law of large numbers implies that the average valuation for a bundle of goods with valuations drawn from the same distribution will be increasingly concentrated near the mean valuation as more goods are added to the bundle (Bakos and Brynjolfsson, 1997). That is, as more and more goods are added to the bundle, the sum of valuations becomes more concentrated around the mean and the shape of the bundle¡¥ s demand curve is more favorate both for the seller and for overall economic efficiency.

Bakos and Brynjolfsson (1996) note that the bundling strategy is particular attractive when the marginal costs of the goods are very low, when the correlation in the demand for different goods is low, and when consumers valuations for the individual goods are of comparable magnitude. If consumers demands remain hererogeneous even after bundling, then a mixed bundling strategy, which offers a menu of different bundles at different prices, will dominate pure bundling. However, when consumers¡¥ valuations for the goods in the bundle are not correlated, the profit advantages of mixed bundling over pure bundling diminishes as the number of goods in the bundle increases (Bakos and Brynjolfsson, 1997).

5. Current Pricing Practices for Information Goods Distributed over the Internet

• Products:

• Cost: Assume the cost function of a firm is C(£UOi, Q), where O is the product offer, summation of O is the total combination of all product offers available at the firm, and Q is the capacity level of the firm.

• Customers:

Max { ƒÃ [(Pj)*Qj] ¡V C(£UOi, Q)}

Pj : price for product offer j.

P = WTP-CS - TC + SC + £`

Max profit = { ƒÃ [(Pj)*Qj] ¡V C(£UOi, Q)}

Transaction costs of customers on the Internet includes:

• Single password: Currently, users must register for most of the sites they are interested in and try to remember all the username and passwords that might have different formats. The value can be created through using only one password to access large amount of content that the provider can have through all means such as licensing from other content providers or creating by itself. And the larger the content base, the more valuable the value provided.
• Single billing: If there is no single billing, each user may receive a dozen of bills which may have different due days or formats from different content providers or businesses. This is really a great burden for customers trying to keep or check all these bills, so single billing can be more convenience to customers. Besides, single billing is also beneficial to the vendor because it raises the average bill to a level which is economical.
• Guaranteeing quality of service or products: There are thousands of businesses on the Internet, how can a user decide the best information source or vendor, how can a user trust the vendor, fear, uncertainty and doubt has been one of the primary limiting factors in the developing of Internet commerce. So a huge benefit that a huge provider can provide is to guarantee quality of service or products through all means available and then build a certain reputation that the user will trust.
• Intelligent agents: Providing intelligent agents which are designed to learn a users¡¦ needs and unique characteristics and to facilitate users¡¦ interactions over the Internet can largely reduce users¡¦ transaction costs and building another sort of switching costs.

4. Strategic Use of Network Effects - Enhance WTP and Q-

Max [ ƒÃ (Pj*Qj )¡V C(£UOi)]

P = PH1 = tH*SH profit = PH1*£k- C(SH) (1)

P = P0 = tL*S0 profit = P0*1 - C(S0) (2)

Profit: [PH1*£k- C(SH)] + [PL*(1-£k) - C(SL)] (3)

If C(SH) ¡V C(S0) „T £k*PH1 ¡V P0

or C(SH) + C(SL) ¡V C(S0) <£k*PH1 + (1-£k)*PL1 ¡V P0

then versioning the product :(PH1,SH) and (PL1,SL)

PH2 = PH1 ¡V SL(tH ¡V tL)

Profit: [PH2*£k- C(SH)] + [PL1*(1-£k) - C(SL)] (4)

and C(SH) + C(SL) ¡V C(S0) „d £ktHSH + tLSL ¡V £ktHSL - tLS0.

In this paper, we develop an optimal pricing model: the incentive compatible pricing model for those that provide digital information goods and on line services on the Internet and the necessary steps in doing the incentive compatible pricing model, they are 1.trying to bundle as many information goods and services as possible. 2.differentiating the products. 3.deciding how many product offers should be included on the site. 4.deciding the materials to be included in each product offer and its price. Though not finish yet, the incentive compatible pricing model has the potential to solve the problems of how to charge customers and market failure on the Internet while it may have less effects in physical world. The model must be supported by other strategies in order to make users and vendors both better off. These strategies include those that will reduce transaction costs, reduce total first investment costs, enhance consumer¡¥ s switching costs and strategically use of network effects. At last, we make a comparison among some major digital information goods or online services providers and indicate the potential providers that might use the incentive compatible pricing model.