Microeconomics: E pluribus tunum (Oct 22nd 2009) The Economist
Question: Why did Apple introduce price discrimination on iTunes, and how can it optimize price discrimination to further maximize profit?
Perfect price discrimination is the goal of many businesses because it maximises revenue and minimizes consumer surplus. There are many challenges in achieving perfect price discrimination, but when Apple first introduced price discrimination in April 2009, it took a great step forward despite using a fairly basic approach.
Before Apple introduced price discrimination, it used a single-tier pricing model where every song cost $0.99 (or an equivalent in local currencies). Because of this uniform price, Apple presumably lost out on purchases and profits in some cases.
In terms of lost purchases, it is reasonable to assume that more purchases would have taken place if songs were offered at a price high enough to cover costs, but low enough to attract more buyers. In terms of lost profits, it is also reasonable to assume that some customers would have paid more for some songs.
These two cases show the advantage of selling goods as close as possible to the price buyers are willing to pay, as long as costs are covered.
iTunes and other electronic distributers have an interesting cost structure in that incremental supply costs are very small, meaning the supply curve is very elastic. Since iTunes is already established, it costs very little to provide any extra song to consumers. The ongoing costs in terms of network, software support, hosting, marketing, and administrative costs are borne regardless of sales volume, and the royalty costs for media sales are paid per sale.
This creates a challenge for iTunes in terms predicting the frequency of purchasing different songs at different price levels and optimizing the price to download ratio to maximize profits. This calculation increases in complexity as the number of price levels increases, but so do the benefits of discrimination, as seen in figure 1: Each new level of discrimination reduces consumer surplus and increases revenue.
Even if we assume all songs available for sale on iTunes are the same quality, there are many reasons consumers would pay different amounts for different songs. Assuming a normal distribution of musical preferences, two thirds of people would probably enjoy music in the same range which would probably include artists like The Beatles and Michael Jackson. On the “tails” there would be demand for increasingly obscure artists and genres. Songs in the middle distribution would likely not deviate widely from the $0.99 cent range due, but hard-to-find music on the tails appealing to limited musical tastes would likely sell with higher prices. Similarly, other music that would have been passed over by consumers would likely sell at lower prices that still cover costs. The downside includes the analysis, implementation, and opportunity costs of the price discrimination model itself (i.e., the opportunity costs of mistakenly pricing a song at $0.89 when $0.99 would have yielded more revenue).
In order to implement price discrimination, analysis must first decide the type that would be most appropriate. There are three types of price discrimination .
First degree price discrimination bases prices on the amount each type of consumer will pay. In microeconomic terms, the price is a function of demand elasticity. This type of price elasticity could be employed by iTunes because the demand for different songs is a function of consumer preferences. I.e., musical beauty is in the ear of the beholder.
Second degree price discrimination discounts prices proportionately with quantity. This type of price discrimination is used in certain instances on iTunes when an album is priced less than the sum of prices of the individual songs. However, it is not in place for unrelated songs. I.e., iTunes does not have a model where 100 individual songs could be purchased for the price of 90.
Third degree price discrimination varies price specific to the seller based on segmentation. I.e., customers living in Monte Carlo or Beverly Hills would likely pay more for songs than customers living in Harare. Other examples of market segments could be Queen’s MBA graduates or people living in subsidised housing. This is the type of price discrimination being referred to in the article when it stated “Person-specific, rather than song-specific, pricing would be more efficient. But sellers’ data are not refined enough to set different prices for different people.” Although it is true that this type of sophistication is not yet achievable with iTunes, personalized pricing will likely be possible in the future based on advancing technology. In fact, companies like Dell computers are able to achieve some level of this type of price discrimination using purchase histories, registration information (indicating demographics), clickstreams, search histories, and other methods.
Other pricing models could also be profitable for iTunes. According to the article, “The most revenue, according to the 2009 survey data, would be generated by charging the students $21.19 for entry and 37 cents a song. This would raise consumer surplus by 30% compared with uniform pricing.” Interestingly, this is a model similar to that employed by Costco wholesalers – charging a membership fee, then offering reduced prices based on high-quantity. Costco is not an electronic distributor and relies on completely different models of economies of scale and economies of scope, and price discrimination in retail stores would not be practical since every shopper pays the same price as every other shopper at Costco regardless of the quantity purchased. This contrasts with iTunes because of the low incremental cost of electronic distribution - in microeconomic terms, iTunes is able to leverage economies of scope by using the distribution model already in place to sell incremental songs.
In conclusion, iTunes implemented pricing to optimize revenue and minimize consumer surplus. The revenue growth and consumer surplus shrinkage are proportional with the number of price levels. iTunes currently leverages type 1 price discrimination based on consumer preferences and willingness to pay, and an aspect of type 2 discrimination based on quantity sales in terms of albums. It will likely be able to introduce type 3, ‘person-specific’ pricing in the future to further optimize profit.
Figure 1: Price discrimination at 3 price levels on iTunes
Before price discrimination, a single price (P) was offered. The revenue is bounded the pink box with P, D, A, and 0. The consumer surplus was the area above line P and the demand curve.
After price discrimination, 3 prices became available, represented as P1at $1.29, the former single price at $0.99, and at $0.89.
The revenue from iTunes after price discrimination is equal to the sum of the following 3 boxes:
1. Bounded by P1 at $1.29, C, (low) D1 and 0
2. Bounded by P1at $0.99, A, (med) D1 and (low) D1
3. Bounded by P1at $0.89, b, (high) D1 and (med) D1
Each new price level introduced reduces the amount of consumer surplus and increases the amount of revenue.
Price Discrimination and iTunes
 Pricing with market power – Chapter 7 (Included in class notes): Personalized pricing