When Apple founder Steve Jobs revolutionized the digital music world with the introduction of the iPod in October 2001, he faced a dilemma: How much is a song actually worth in a new marketplace in which production and distribution costs, like the manufacturing and shipping of compact discs, are no longer an issue?

Over the last decade, much has changed for digital music consumers — including the growing array of mobile devices that can play music, and the rise of new rivals to Apple’s iTunes store. But one thing has been remarkably steady — the price of a song, long held at 99 cents for the majority of individual tracks (with some hits costing as much as $1.29 and a few less popular tunes priced at 69 cents).

New research by Wharton marketing professor Raghuram Iyengar suggests that the major record labels, which have not always been convinced of the fairness of the pricing structure established by Apple’s iTunes, could be making higher profits from their music catalogues — not by raising their prices, but by lowering them.

Iyengar, who specializes in issues of pricing and consumer behavior, recently surveyed about 600 digital music consumers with the goal of developing optimal pricing structures for the music industry, which has been losing revenue as it struggles to switch from an era of records and CDs to the world of digital downloads. His research looked at both pay-per-song plans — such as the one offered by iTunes — and also subscription models like those from companies such as eMusic, in which consumers pay a regular monthly fee for a fixed number of song downloads.

Iyengar was surprised to learn that not only was the pay-per-song model more popular with the digital consumers, but that their music purchases would increase sharply at a lower price — so much so that the record labels would actually raise their overall profits by dropping the cost per song.

Although somewhat constrained by lack of data on the actual cost of producing a song, Iyengar says his research suggests that the optimal price a record label should charge a retailer is probably closer to between 30 cents and 40 cents rather than the 60-cents-or-more that is reportedly the amount the labels currently charge iTunes. This will have downstream effects in that the cost to the customer would come in at around 60 cents to 70 cents per song rather than the typical 99-cent price common on iTunes.

The music labels “seem to be charging much higher [fees] than they should,” Iyengar says. “If I compare what their profits are when a record company charges a retailer 60 cents a song, I find that [the current] overall profits for the entire channel, which is the label and the retailer, are almost 50% lower than what they could optimally be when the record label charges lower wholesale prices.” Iyengar’s research is presented in a paper titled, “Optimal Pricing for a Menu of Service Plans — An Application to the Digital Music Industry” (PDF).

From Cell Phones to Digital Music

Iyengar had already conducted extensive research on optimal pricing plans for mobile telephone companies, but he wanted to know more about pricing in digital music because of the significant differences between the two industries. “With a cell phone contract, it’s not only how much you use it, but how much people call you, so there’s much greater uncertainty when you think about choosing a plan,” Iyengar says. At the same time, music consumers are likely to behave differently than cell phone customers in that they are choosing songs they plan to listen to repeatedly. Also, more ups and downs may occur over time in music purchases depending on which artists are releasing new material — a factor that could make buyers less likely to sign up for a subscription plan with a monthly obligation.

The issue of pricing has long been a thorny subject within the music industry. Music labels became accustomed to large profits in the late 20th century because barriers to entry — including control of distribution and the high costs of manufacturing records, cassettes and compact discs — gave the industry the ability to extract more revenues from recorded music. But now music labels are concerned that the 99-cent model developed by Jobs and his iTunes store is actually too low.On the other hand, some experts also note that either higher prices or complicated pricing changes could drive more customers back toward the practice of illegal free downloading, the bane of the industry.

The pricing issue is taking on critical importance because of the music industry’s recent slump. According to data from the Recording Industry Association of America, overall revenues from music sales — including sales of compact discs as well as digital music — were $11.75 billion in 2006 but dropped to $10.37 billion the following year. In 2008, during the middle of the economic recession, revenues plunged even lower, to just $8.48 billion.

Conjoint Analysis and Pricing Plans

According to Iyengar, it is difficult to take existing information from the digital music industry and conduct meaningful analysis on the relationship between price and demand because most companies rarely change their prices. Instead, he decided to use the well known marketing research technique of conjoint analysis, a practice developed at Wharton by marketing professor Paul Green. In conjoint analysis, customers are offered a menu of plans with varying features in order to determine which features are most influential in their purchasing decisions.

Respondents were shown several choice sets, each consisting of three different pricing plans that included both pay-per-song and subscription-based alternatives. In each set, respondents could choose one of the three plans or decide not to choose any, if all of them were unappealing. Iyengar says that looking at which plans consumers chose helped explain what factors influenced their choice of plans, which, in turn, helped to design optimal pricing plans.

Iyengar worked with a digital music retailer to develop the survey data from a sample of both existing customers as well as individuals who had expressed interest in the service. A separate control group of 350 customers showed that the most important attributes for a digital music service are access fees, the maximum allowable number of monthly downloads, the per song price, and features such as audio quality and the type of available music.

One of the key issues Iyengar was seeking to determine was the price the record labels should charge the retailers. While the record industry does not make this wholesale price public, news reports have indicated that the price the music companies now charge to iTunes and other retailers is about 60 cents a song, as noted earlier.

To determine how sales would be impacted, Iyengar calculated optimal wholesale prices for record labels and the downstream retail prices at several levels of production cost to the record labels. For instance, at a zero per song production cost, the model results indicate that the record label should charge a wholesale price of 23 cents a song to the retailer, and the cost to the customer would come in at 54 cents per song. But because of the increased downloading by consumers at this lower price, the profits for both the label and the retailer would actually increase significantly. Iyengar calculated that the combined monthly profits from each consumer would be about $5 higher, an increase of roughly 50% more than a scenario where the record label charges a wholesale price of 60 cents per song to the retailer.

“The record companies are overcharging,” Iyengar says, based upon the data that he collected. “Even if you consider a higher production cost of 15 cents per song, the model results indicate that record labels should charge a wholesale price of around 40 cents per song to the retailers. Thus, the current wholesale price of 60 cents per song is clearly suboptimal.”

Iyengar’s research also found that customers tend to have a preference for pay-per-song plans as opposed to a monthly subscription option, and that the former approach leads to higher profitability. As a result, it makes more sense for a retailer offering just one type of plan to provide only the per song pricing alternative. The research also suggested that the pricing decisions made by a record label play a significant role in the popularity of the so called “bucket plan,” which allows a certain number of downloads for a set fee. For instance, if the music companies keep the wholesale price at a low level, Iyengar found, this makes it attractive for retailers to offer a bucket plan to their customers. He compared this to restaurants that find it profitable to offer an “all-you-can-eat” buffet to increase demand, but only if the cost of providing that food remains low.

“If the costs go up,” he notes, “all-you-can-eat may no longer be optimal. What pricing plans retailers offer their customers clearly depends on how much record labels are charging them.” 

Iyengar believes that the overall analysis suggests that lowering prices could be a path to greater profitability for an industry that has been in trouble in recent years. “If the price goes down, consumer demand will go up, so much so that overall profits will be higher” — and that could have added benefits beyond higher earnings. For example, “it has the potential to reduce piracy because lower prices will entice consumers to download songs legally” rather than be tempted to steal them, he suggests.