Final thoughts (for now) on an Amazon digital music service

3 minute read

After exhausting myself making predictions about Amazon and its possible entry into the digital music market, I’m now prompted to return to the topic, prompted by Hypebot’s latest article on why Amazon doesn’t seem to be doing anything yet. I might quibble with some of Hypebot’s stated reasons (e.g., how likely is it really that Microsoft can make a success out of Zune?), but after thinking about it I do agree that an Amazon/eMusic deal doesn’t necessarily make sense, both because of the difficulty in bringing over the eMusic subscription model into the Amazon environment and because Amazon is probably perfectly capable of matching eMusic’s other features on its own (as I noted in my first post on the subject).

I’m looking forward to reading Hypebot’s thoughts on how Amazon could create a successful strategy for entering the digital music business. In the meantime before the pros weigh in I thought I’d make one more amateur’s attempt to speculate on a future Amazon offering. After re-reading my posts, here’s my current thinking on the key approaches Amazon should take:

  • Amazon should leverage the fact that it will be selling both CDs and digital tracks, including offering cost-effective ways for customers to add digital track or album purchases to a CD purchase or vice versa, whether at the same time or at a later time. For example, a person might buy a digital track, like it, and then come back later and buy the digital album and/or the CD; Amazon should make it easy and cheap to do this.

  • Amazon should offer some sort of pricing scheme that allows it to offer more value to its customers, for example by offering at least some digital tracks and albums for amounts significantly lower than the current standard $0.99 and $9.99 prices respectively, and/or by offering reasonably aggressive discounts on digital music to customers willing to commit in advance to a certain volume of purchases.

  • Amazon should leverage its strengths in the area of back-end infrastructure to create Amazon-specific service offerings that others would find relatively difficult to match. I mentioned a music backup service based on Amazon S3 and a personalized recommendation service based on Mechanical Turk; there may be other such possible offerings as well.

There are also the obvious things Amazon could do, like using buying data for digital music to further refine its existing recommendation engine. (For example, the fact that someone bought a given CD after buying the corresponding digital album should be taken as a positive recommendation for that album.) There may also be other things I’ve neglected to mention; for example, I didn’t address the issue of Amazon bundling digital music with the music players it sells, either by offering coupons for later downloads or by pre-loading music on the devices.

Finally, two areas where I think I went a bit astray in my predictions:

  • In retrospect I think my advocacy of a relatively complicated form of variable pricing was probably a mistake. I think any form of variable pricing is more likely to feature a set price for a given track, based solely upon the identity of the track.

  • I think Amazon would not be able to rely as much as eMusic on people not downloading their maximum entitlements under their subscription plan. In eMusic’s case the subscription plan is mandatory, so casual users are thrown into the same pool of downloaders as the heavy users, with the former subsidizing the latter. In Amazon’s case the casual users would have (and would no doubt take advantage of) the option to buy tracks or albums a la carte, so the pool of subscription users would be much more weighted towards heavy downloaders who’d be more likely to use up their monthly quotas, resulting in corresponding lower discounts.

At this point I’ll conclude and await Hypebot’s next Amazon-related post.