In part 2 of my series “Sympathy for the labels” (see also part 1) I discussed the concern of labels that selling through eMusic lowers overall profitability by diverting sales away from the iTunes Store and other higher-priced outlets. Note that I’m referring here to labels that are realizing at least some profit from eMusic sales. As I wrote in part 2, labels that are losing money on every track sold through eMusic, e.g., due to fixed per-track mechanical royalties, should not be selling through eMusic, period. However even if a label’s eMusic-related sales are profitable, the label might still have a legitimate concern about whether those sales are supplementing sales through other services, or displacing them.

I think the most straightforward strategy for labels to address this concern is to work with eMusic to implement some form of price discrimination:

A seller price discriminates when it charges different prices to different buyers. The ideal form of price discrimination, from the seller’s point of view, is to charge each buyer the maximum that the buyer is willing to pay.

The basic idea is that selling at a single price divides potential customers into three groups:

  • those who think the price is too high (and who therefore don’t buy at all)
  • those who are willing to pay the price (but not much more)
  • those who would be happy to pay a higher price (but end up paying the same price as everyone else)

In a single price scenario the seller (a record label in this case) makes no money at all on the first group, and makes less money on the third group than could be made in theory. The seller could increase overall revenues and profits by offering a lower (but still profitable) price to the first group, and a higher price to the third group.

However doing this is not trivial:

Every seller would price discriminate if there were not two major obstacles standing in the way. First, the seller must be able to distinguish between those buyers who are willing to pay a high price from those who are not. Second, there must be substantial difficulty for a low-price buyer to resell to those willing to buy at a high price.

In the context of eMusic the second obstacle corresponds to people buying low-priced MP3s from eMusic and then reselling them to others at higher prices (a form of arbitrage). Unless I’m missing something, this is a non-problem. If anyone is inclined to violate the eMusic subscription agreement then they’re far more likely to simply offer up their purchased music collections on P2P file sharing networks. While this adds to the labels’ general woes it doesn’t directly affect the eMusic vs. iTunes Store issue, since I’m assuming that the customers of both services are willing to pay some amount for music, no matter how small, and are not obtaining music solely via P2P networks.

I’ll therefore ignore the second obstacle and focus on the problem of distinguishing between price-sensitive and price-insensitive customers and then price discriminating on that basis.

Product versioning

One way to do this is through product versioning, i.e., offering multiple versions of the same product that differ in terms of quality or other aspects. Examples in the music industry include offering albums as CDs (including album art and liner notes) as well as in digital versions at varying levels of quality (e.g., 128Kbps MP3 vs. 192Kbps VBR MP3 vs. FLAC, etc.). Price-insensitive customers are happy to buy the higher-quality version, while others are willing to accept a lower quality product in return for a lower price. (In actual practice retailers typically offer at least three product versions at three different prices, using so-called “Goldilocks pricing.”)

However such versioning doesn’t directly address the eMusic vs. iTunes Store issue, since both are offering roughly equivalent products. In fact, for those releases available on both services eMusic has traditionally offered an arguably higher-quality product at a lower price, since unlike the iTunes Store it offered tracks that were playable on more devices and were free of DRM-imposed restrictions.

Before it left eMusic Tzadik used a particularly strange method of product versioning: It made albums available for downloading on eMusic but omitted tracks that were 15 minutes or longer in length. I strongly suspect that eMusic subscribers are much more interested in downloading albums than single tracks, both because independent music doesn’t tend to be singles-oriented and also because downloading complete albums reduces customers’ mental accounting costs. Tzadik’s policy thus amounted to offering a significantly inferior product on eMusic in the hopes that all but the most price-sensitive customers would go elsewhere to buy it. It’s analogous to a car manufacturer forcing low-cost car dealers to sell its cars without back seats: it might make superficial sense to someone in marketing, but customers would generally think that the manufacturer was taking them for fools.

My conclusion is that product versioning in the classic sense won’t work in the eMusic context, at least given current eMusic offerings. (However it’s possible that in the future eMusic and the labels could cooperate to offer certain premium products, as I’ll discuss in a future post.)

Release windows

Another strategy is to offer a product initially at a relatively high price, and then to later offer the same or a similar product at a lower price. (This can be thought of as a special case of product versioning, with the differentiating product aspect being time rather than quality.) This is the strategy that book publishers use with hardcover books vs. paperbacks: Price-insensitive buyers purchase the hardcover version, while price-sensitive buyers are willing to wait a few months for the paperback to come out. The delay is necessary because the differences between a hardcover and paperback book are not necessarily enough to promote proper price discrimination if they were released at the same time: Most if not all people would likely be willing to settle for the paperback version and ignore the hardcover version, since what’s most important to them is the content of the book (“What happens to Harry Potter?”), not the sturdiness of its covers.

In an eMusic context this strategy corresponds to labels releasing albums to CD and the iTunes Store first, and then waiting a few months before releasing to eMusic. The assumption is that the quality difference between the CD and the iTunes Store digital version is sufficient to promote price discrimination at the time of original release between very price-insensitive customers (who buy the CD) and moderately price-insensitive customers (who buy the iTunes Store digital version), and that the most price-sensitive customers will be willing to wait some time before purchasing the album at eMusic. Of course, some people won’t want to pay the iTunes Store prices and will go to P2P to get the album, but it’s not clear how many of those people would have paid in any case, even through eMusic.

One problem with this strategy is that it doesn’t take full advantage of labels’ marketing efforts at the time of the albums’ initial releases. Such campaigns are to some extent wasted on eMusic customers, especially those customers who predominantly purchase through eMusic, because they can’t buy a newly-released album through eMusic at the time when their interest in the album is at its height. By the time the album becomes available on eMusic customers may put a much lower priority on downloading it, having found lots of other albums to consider downloading in the meantime.

Similar considerations have motivated movie studios to consider reducing the window between theatrical release of movies and release on DVD. However others have claimed that it’s counter-productive to move toward shorter release windows and so-called “day and date” releases to multiple channels. Also, sales of movies are much more driven by marketing campaigns than sales of albums, many of which tend to sell well long after initial release.

Thus it’s likely a viable strategy for labels to have a window between availability in “full-price” digital music services and availability on eMusic. However as I discuss in the next section it’s also possible that labels could find a strategy that protects iTunes Store sales while not totally forgoing the possibility of realizing eMusic-related revenue starting at initial album release.

Behavior-based differential pricing

If a retailer has enough information about its customers then it can use customers’ past behavior to determine whether they appear to be price sensitive or not. Retailers can then implement pricing schemes that target particular purchasers for special treatment. For example, in the past Amazon has experimented with varying prices shown to different customers, a practice also followed by other online sellers. In one application of this strategy, a site may use targeted pricing to grow its customer base, offering special discounts to get new customers hooked while displaying higher prices to loyal customers, on the theory that they are less likely to switch to a competitor.

As the referenced articles indicate, while differential pricing may be economically efficient it is not easy to implement such schemes without running into customer concerns about fairness. The same Internet that makes online commerce possible also makes it possible for customers to compare the prices they are being offered, and to evaluate whether or not they’re being unfairly taken advantage of in particular circumstances.

In the case I’m considering the concern is about eMusic impacting labels’ sales through the iTunes Store and other services. The problem at hand is thus as follows:

  1. Use what eMusic knows about its customers to identify which are most price-sensitive.
  2. Implement a mechanism to allow labels to sell through eMusic only to the most price-sensitive customers, forcing less price-sensitive customers to go to higher-priced outlets.
  3. Do this in a way that is fair and can be justified to all eMusic customers.

In my opinion the most price-sensitive eMusic customers can be identified as follows:

  • They are long-time eMusic subscribers, having recognized early the value of the service as a low-cost way to purchase music.
  • They have signed up to the higher-priced plans (e.g., Connoisseur plans or equivalent) and/or to annual or biannual plans, in an effort to get the lowest per-track price possible.
  • They regularly max out their plans in order to not waste money on unexpired downloads, and make minimal or no use of relatively expensive booster packs.

It may be counter-intuitive that the customers spending more money at eMusic are generally the most price-sensitive, since we typically assume that big spenders don’t care about prices. However in the context of eMusic the highest-paying customers are most likely not wealthy individuals but rather people of average means who are dedicated music lovers and have an almost insatiable demand for new music. They are price-sensitive because even a small decrease in price makes a significant difference in the amount of additional music they can purchase. For example, someone on the $74.99 Connoisseur plan is downloading the equivalent of about 30 albums a month, one new album every day. A 10% decrease in price would equate to three extra albums per month for the same amount of money.

eMusic already has all the information it needs to identify price-sensitive customers according to the above criteria, since it obviously knows which plans its customers are subscribed to and how much they’re downloading. eMusic also has the ability to make particular releases available to some customers and not to others, since it has to do this already in order to implement restrictions on which releases are available in which countries. It therefore should be well within eMusic’s capability to allow labels to designate particular releases as being available only to customers previously identified as being price sensitive. The only remaining problem is to do this in a manner that is arguably fair to eMusic’s customers.

In my opinion the most fair mechanism would be based on either customers’ selected plans or on the length of (uninterrupted) time they’ve been eMusic subscribers, or both. For example, making selected new releases available only to people on Connoisseur plans could be justified as providing extra benefits to eMusic’s best customers. Similarly, making new releases available only to people who’ve been eMusic subscribers for at least six months (say) could be justified as a reward for customer loyalty. Using the third criterion above (maxing out plans and not buying booster packs) is arguably not fair, both because customers may have good reasons for not using all their downloads (vacation, sickness, etc.) and also because customers who don’t download their entire quotas and who buy booster packs are in fact contributing more to eMusic’s and the labels’ bottom lines.

Addressing concerns about selling through eMusic vs. other channels

Based on my comments above, I propose that eMusic work with its labels to allow them to designate certain albums (and associated tracks) as “premium” offerings that would be made available only to certain eMusic customers; these albums would typically be new releases but might also be back catalog albums that were still doing significant business at the iTunes Store and in CD format. Such releases would be made available only to eMusic customers who were currently on a Connoisseur plan or had been eMusic subscribers for at least the past six months.

(I think the ability to download premium releases should be extended to all eMusic subscribers if they stay with the service long enough, both because it’s more fair to existing customers and also because this encourages new subscribers to stick around long enough to become loyal customers and perhaps upgrade to more expensive plans. I chose the six-month window as being long enough to protect labels’ revenues and profits on new releases sold through non-eMusic channels, but short enough not to discourage new eMusic subscribers. Finally, eMusic could also extend the ability to download premium releases to other selected customers as appropriate, for example customers on older high-priced plans, customers on annual or biannual plans, and so on.)

Labels would then have the following options with regard to selling through eMusic, with pros and cons as noted:

  • Not sell through eMusic at all. This would guarantee higher payouts in all cases, but could mean forgoing incremental revenues and profits from price-sensitive customers, who might otherwise purchase music from other labels that do happen to sell through eMusic, or who might simply resort to downloading the label’s releases using P2P networks. I’d recommend this option only for labels whose cost structures are such that they cannot sell any releases profitably through eMusic.
  • Sell all releases through eMusic, without exception. This would maximize the potential revenue and profit obtained through eMusic, but might negatively impact revenue and profits through other channels with higher prices. This strategy would make the most sense for labels whose sales through the iTunes Store and other mainstream-oriented services are relatively insignificant.
  • Sell only back catalog releases through eMusic. This would alleviate the financial impact of eMusic diverting new release sales away from the iTunes Store and other outlets, but (as with not selling through eMusic at all) would mean leaving money on the table by driving away price-sensitive customers who might otherwise have purchased the new albums through eMusic at the time they were originally released. I believe this option would thus be inferior to the next one.
  • Sell all releases through eMusic, but designate new releases and certain others as premium offerings available only to certain eMusic customers (as described above). This would help preserve new release business through the iTunes Store and other services by limiting the extent to which more casual buyers could obtain new releases through eMusic (e.g., by signing up for a free trial), while still getting incremental revenue and profits on new releases from dedicated music buyers who have proven to be price-sensitive (and are therefore unlikely to buy from the iTunes Store or elsewhere).

The major downside of the fourth option is that it could hurt eMusic’s ability to attract new subscribers, since people just joining eMusic wouldn’t be able to immediately download certain new releases (including most likely those from the latest “flavor of the month” artists) unless they signed up to a high-priced plan. However at the same time this option might help with eMusic customer retention, since existing customers would retain their ability to get new releases at low cost only by staying with eMusic and not interrupting their subscriptions.

In the end I think this comes down to a trade-off between eMusic’s desire to grow its own business and the labels’ desire to protect theirs. As I implied in the conclusion of my first “sympathy for the labels” post, I believe it’s in eMusic’s long-term interest to keep labels reasonably happy, even if doing so harms its own business a bit. I believe my proposal above could help address labels’ concerns about eMusic’s negative impact on non-eMusic sales, and would arguably be better for eMusic itself than having labels using eMusic only as a back catalog outlet, or not selling through eMusic at all.

luckybleu (luckybleu@optonline.net) - 2008-06-10 15:40

QTRAX is coming signed sealed delivered