Why the iTunes Music Store is successful

I’m not a big fan of the iTunes Music Store. In fact, there are a couple of things about it that really rub me the wrong way, and I have refused to be a customer of it for some years. But, I had an experience yesterday that showed me why it continues to be such a success in the marketplace.

Yesterday, I got an email from a list I’m on, announcing that musician Mike Doughty has a new live album available on his website for electronic download. I’m a big Doughty fan, and I like buying relatively direct from music artists under the theory that they are likely getting a higher percentage of those sales. So, I went to the web site and checked it out.

The site had two flavors of the album for sale. There was a 128 Mbps mp3 version for one price, and a FLAC version for slightly more, neither with any DRM. Excellent! I tend to rip mp3s at 192 Mpbs, so I was reluctant to explore the 128 Mpbs solution. I did a little googling on flac and found there were some utilities for the Mac that would convert them to mp3. So, it looked like that would work for me. I downloaded one of the flac utilities (MacFLAC) and started my purchase of the flac version of the album.

Problem #1:
On the website, I filled out my customer info and credit card data, and clicked “purchase”. And the cursor spun, and spun… and spun… and finally the operation timed out. Grrr… Reload the page, start the purchase all over again, and this time it succeeded. A minor annoyance, but one that never happened to me back when I bought through the iTMS.

Problem #2:
After the purchase completed successfully, I got a link to a web page where the music could be downloaded. Sadly, the packaging of these links was simply a list of all the tracks. And because this was a web page, the basic solution was to click each song and “Save As…”, or right-click the link and download the link target to the desktop. Song by song, so 26 tracks. Sigh. And because these were lossless tracks, the files were large. I kept this up for a good bit of the afternoon, starting a new download every time I saw the previous one had completed. After far too long of a manual effort, I finally had all of the flac files downloaded.

Problem #3:
I launched the flac utility I downloaded, and right away it told me it had to install some files in order to run. Sigh. It really irritates me when a tool isn’t written to load libraries, frameworks or unix utilities from its own bundle. But oh well, whatever. I clicked OK, and it attempted to install. And immediately returned an error. It wasn’t able to complete the installation it needed. Suddenly I worried that I had downloaded a bunch of music in a file format I wouldn’t be able to use. Some more googling, and I found another utility, one which launched and ran just fine.

Problem #4:
The new flac utility, Switch, would load the flac files, and would export to mp3. Excellent! It even had a button to allow me to control the bitrate of the mp3 files it generated. The problem was, it also had options for a number of other things, some of which I had never heard of before. Should the “Channel Encoding Mode” be set to “joint” or “stereo”? Should I use VBR encoding, and if so, what should i select for minimum and maximum bitrate? Should I check “Include CRC to detect errors”? I don’t even know what CRC stands for! There were simply too many options for me to use it confidently. I finally opted to convert the files to AIFF format, going from one lossless format to another (and hopefully the conversion doesn’t lose any quality. That conversion (which had no additional config options) worked just fine, and from there I imported the aiff files directly into iTunes. Hey, I’m not a fan of the iTMS, but I still use iTunes as a player! Once I had the aiff files into iTunes, I could use iTunes itself to convert the tracks to 192 Mbps mp3 files. Finally!

Problem #5:
If the original downloaded tracks ever had ID tag info, they certainly didn’t after multiple conversions. So, I spent a chunk of time manually entering track names, artist name, album name, etc. Not particularly difficult, but it’s a tedious process at best. Oh yeah, and I did some more google surfing to find album artwork to paste onto the mp3 files.

I finally accomplished what I wanted, but it was far more work and pain than it needed to be. I think the iTunes Music Store has succeeded in large part because it has removed virtually all of the hassles I describe above. You find a song you like, you click “Buy” and you’re done. The tracks end up in iTunes, in the right format, with ID tags and album artwork and everything. It really couldn’t be simpler, and for a consumer product, that is absolutely essential. By contrast, the experience I went through yesterday…any single one of those problems would have stopped my mom in her tracks.

Just to be clear, I do not mean to paint the iTMS as a panacea for everything I describe above. I’m unwilling to buy DRMed music, and the bitrate offered is only 128 Mbps. But those are issues that “my mom” wouldn’t care about in the slightest, and the ease of use story still wins by a mile.

I’m a big believer in not griping gratuitously if you can’t also offer a solution. And I do think the problems I described could have been solved with more work on the part of the music vendor. For instance:

Music format and bitrate
The Russian music site AllofMP3 has a nice solution to this problem. Their online music library is stored in a lossless format. When you make a purchase, they ask what format and bitrate you want the files. When the customer makes their purchase, the music server churns away, encoding the files in the necessary format. It takes a couple of minutes, but it saves the paying customer from having to do the work.

Downloading of files
Rather than providing a list of songs all of which need to be downloaded individually, the music server could easily zip them into a single bundle for download. This would give the customer a single file to download (perhaps a single zip file per album?). Again, the whole point is to save the customer from needless tedium in the process of acquiring the files.

ID tags
The original lossless music files on the server should have all necessary track metadata, and when the tracks are being encoded in the customer’s selected format, that data should be seamlessly translated into the new tracks.

None of this requires anything in the way of innovation; the code for doing all of the above is fairly straight forward and doesn’t require novel research. For a capable engineer, I think it’s less than a week of work. It’s just a matter of thinking the problem through and doing the right thing. Which seems to be a rarity these days.