Steve Jobs’ Funky Numbers: a closer look at his essay
Yesterday, I commented (in an extremely long post) on Steve Jobs essay about DRM systems in the music industry. I was impressed with his post, and lauded it. Having slept on my thoughts, I still feel that way, but one thing irks me – Jobs’ numbers were bad.
I noted yesterday that Mr. Jobs numbers were, in a few instances, exaggerated and simplistic but I didn’t say much more than that. Today, partly in reply to an email I received, I am going to focus briefly on his numbers.
What Jobs Said:
Through 2006, 90 million iPods were sold along with 2 billion songs from iTunes. That yields an average of 22 songs purchased from iTunes per store.
Reality:
That math may be factually accurate but it’s misleading. The iPod has been around for many years and any given single owner has probably gone through more than one iPod in that time. (I, for example, have had two break and am on my third. I also have a Shuffle that I received as a gift. That means, I account for 4 of those 90 million, not just 1). Extrapolating song purchases based on total units is far less telling than doing it by unique buyers.
Better Math:
Recalculate those numbers based on the songs purchased per unique iPod owner rather than total units. Something like this example:
Assume 5% or iPods were replaced because they broke. Assume another 20% were upgraded to a new model. That would mean 25% of the 90m ipods sold were duplicates for prior buyers.
90m iPods – 25% = 67.5m iPods
Assume another 10% of buyers might have two different items (a Shuffle and a video iPod, for example) then you’d have 57.5m iPods sold to unique individual buyers. The real number might even be lower, but for guesses, it’s probably not bad.
With 2 billion songs sold, to 57.5m owners that means the average unique iPod owner has bought 34.78 songs. Proportionally that is much more than the 22 songs per Ipod in Jobs’ example but it is still a small number. It’s like buying 3 albums instead of two. Hardly a big deal.
What Jobs’ Said:
The most popular iPod holds 1000 songs and the average iPod is nearly full. He
Then concludes that means 22 songs (35 by my adjustment above) out of a 1000 are DRM protected.
Reality:
Jobs infers the average iPod and most popular model are the same thing. That’s misleading in several ways. First, even if the most popular iPod today probably does hold 1000 songs, capacity is a factor of file size. MP3’s in 320kb format are going to be much larger (and reduce the total capacity) as compared to MP3’s in 128kb format. Further songs don’t come in standard lengths; some will be longer than others. Second, and more importantly, capacity has varied over the past few years since the first generation iPod in 2002 as Apple has expanded the lineup and played with drive size. Since his numbers are taken from across the life of the iPod, using the numbers from the most current generation, and most popular product, is misleading. Third, iPods have been designed to carry more than music (particularly the most recent generations of the Nano and Video iPod). Even if these devices are close to full, not all of the data is likely music but Jobs statement seems to suggest otherwise. Ideally the math would account for sales of the different iPods over the years and provide a number of songs appropriate to the average capacity of the total units used solely for music. For Jobs purposes those calculations were probably unnecessarily complex.
Better Math:
For my purposes getting that data to give true accuracy requires reading through a lot of financial statements and I don’t want to spend the time (though the wikipedia entry on the Ipod does consolidate a lot of the info. Instead, I am going to arbitrarily pick numbers at a discount to Jobs listed data.
I’ll call the average iPod capacity 4gb. That number is drawn from the fact that approximately 80m of the 90m iPods sold were sold after the introduction of the Mini (4gb capacity January 2004) and the later introductions of the Nano (1 to 8gb) and Shuffle (512mb to 1 Gb). Smaller capacity units represent a substantial portion of the iPods in use.
For my math, I will next arbitrarily discount the capacity of an Ipod to 175 songs per gigabyte to account for different song lengths and file formats. (Note: the number is not completely random: 175 songs is about 73% of the maximum capacity Apple lists for a 1gb shuffle).
Assuming Jobs statement that most iPods are nearly full, I’ll say an average iPod (4gb) is 75% filled with music, which is random but seems a fair number considering the different kinds of data that can be on it (as well as the fact that part of the drive is allocated to system operation). With these assumptions:
3 GB of a 4GB iPod is filled with music.
At 175 songs per gigabyte : the average Ipod has 525 songs.
If 35 of those songs were bought by a unique user in iTunes: 35 out of 525 are DRM protected (as opposed to 22 out of 1000). That is approximately 6.5% of the songs on the iPod with DRM. That’s about triple the number Jobs used but it is still a negligible amount.
Conclusion:
In Jobs essay he used the simplest of numbers to make his case. I can’t blame him. From startups to big business people always want to present the data that best supports their case and Steve Jobs was definitely making an argument.
Apple is under fire that the Fairplay DRM system is limiting consumer choices, especially in Europe. Jobs wants the world to know, that Apple is being made a scapegoat for the consequences of the music industries forced behaviors with DRM. Even if iTunes makes users less willing to migrate to another platform of devices (ands that’s deemed anti-competitive), it is the music industry that put Apple in this spot. That is the subtext of his message, as i read it.
Jobs used numbers that show, at the extreme, that many of the complaints being levied at Apple are unrealistic and factually inaccurate. I wish he provided better math (actually, more accurate data) in doing so because I think, in reality, even a conservative crunching of the numbers would still support his claims.
