Much to my consternation, Silicon Alley Insider has become a must-read for me, but its analysis of a potential Apple digital music subscription service is a bit off the mark.

…But if that’s really the only thing keeping this from happening, then this is a done deal. Right now Apple sells about 20 iTunes tracks for each iPod or iPhone it sells. So the two sides just need to replace that revenue in a subscription model.

For argument’s sake, let’s say there’s a digital music boom coming, and the tracks-per-device number jumps 50% — to 30. How much would each side need to compensate for lost sales over the life of the device? Very little.

* Apple collects about a third of each 99 cent track sold: 30 x .33 = $9.90
* The labels keep the rest: 30 x .66 = $19.80

Two things — contrary to the headline, that’s not an “all you can eat model,” unless I’m missing something. And replacing potentially lost iTunes sales aren’t what the labels are trying to do. They’re trying to replace lost CD sales.

According to Jupiter surveys, somewhere between 5% and 15% of the songs on anyone’s iPod are purchased from a downloads store, the average user has 1500 songs on his iPod (though that’s skewed by big collections: only 20% have more than 1,000), and the average iPod user spends $20 to $35 per year — not over the lifetime of a device — on downloads. (The average paying downloader actually spends at least $10 more, but a lot of iPod users don’t buy any downloads.)

UPDATED: Oh, duh, I get it. SAI is focusing on the difference b/t all you can eat for $80 vs. $20. Apologies for being dense. Well, my other data is still useful for the calculations, and it shows why the labels would want a way bigger number than $20.