iTMS Downloads Surpass 300 Million

[rogo]

I am inclined to believe a few things:

(1) The data is not real. I, too, track this and did indeed have one forecast that put 300 million at this date. I rejected that forecast in favor of one that had it 9 days ago for a number of reasons... Anyway, I can believe the data only if there is an error in the 230-250 data.

(2) The rate of growth should not be slowing. They are selling ridiculous numbers of iPods. While new countries would be nice, they should increase further a rate of growth that should already be increasing. The more users, the more potential customers.. Even if this is a novelty, they should be selling more songs.

(3) The next data point will help make sense of the graph. April 8 would be bad news. April 9 and beyond would be terrible new... Arpil 5-7 would show a tiny rate of increase... April 4 or better would be appreciably more impressive.

(4) Absent a confirmation of the next milestone at <34 days, I don't see how they can possibly reach 1 billion this year. The rate would need to be ~2.3 million per day when it's currently running -- allegedly -- below 60% of that. Think about it this way... There are 10 months... If we ignore the 230 for a second, there appears to be a gain of 100,000 on the average for each month that passes. If that continues, then the average will be more like 1.8 million / day this year for a total of ~550 million additional leaving the year at 850 million sold to date.

(5) Whether this business is "tracking" will be crystal clear by summer. Apple is either on it's way to selling 5% of the world's recorded music in 2007 or it isn't. If the former, the music arm of Apple will ultimately be worth much more than the rest of the company. If the latter, look for Apple to come out with a Napster-to-Go-like service by early next year. (That might happen anyway. Subscriptions are a fine complement to the selling business and Jobs already knows this... "Pick whichever method you like, Ms. Jane Q. Customer."

 

[wdlove]

500 Million by year's end...any takers?

Going by the graph, it looks like a real possibility. So yes, I will. I think that this is just another amazing milestone by iTMS downloads.

 

[Lacero]

I'd say a billion songs cumulative by the end of this year. The shuffles, lower priced iPod photos and new iPod minis will make sure 2005 is the year of the iPod. Then more cross-over promotions with Star Wars, and god knows what else, and you have the making of a blockbuster. The only caveat to this projection if the song prices remain at $0.99.

 

[Mike Teezie]

Well, I just recently came around to the iTMS, and I've bought two albums this week.

It's quite addicitive, so I can totally see 500 million this year.

 

[spaz]

iPods means iTunes sales

don't forget that every iPod sold grows the potential for iTMS sales. a ton of people got iPods for Christmas (about 4.5 Million, in fact) and even if people don't run to the iTMS and start buying right away, the iTMS is where they're going to be going if and when they start purchasing music online. the buy rate has almost nowhere to go but up, up, up.

 

[xsnightclub]

HOLIDAY SHOPPING SEASON

It should be painfully obivous to people that the slight decrerase in rate of growth after mid january can be attributed: to not just the holiday shopping season season ending with an uptick for when people are redeeming the gift cards. But ALSO remember the record companies flood the market with new music for the holidays, leaving little of interest for the january-february releases.

This is done to increase sales and keep music released closer to the Grammys for awards.

So the fact that sales still increased is damned impressive.

 

[FeralCat]

Dec 16 - Jan 24, 50 mil songs = 1,28 mil a day
Jan 24 - Mar 2, 50 mil songs = 1,35 mil a day

I think it's slowly stabilizing to around 1,3 mil a day. For a next bug jump we would have to wait for new stores.

Also, since when is a revenue jump of 6.9 mil a day over a month not that much?
(Completely ignoring the invisible relatively in your statement)

-Hertz

Hertz, I don't understand your post at all. 1.35 million - 1.28 million = 70 thousand per day, or $70,000 more per day, not $6.9 million. It's ~$2.2 million more revenue over a 31 day period, so even if you mean additional revenue over the entire month, your number doesn't make any sense.

Anyway, congrats, Apple. I say that the rate of growth can't be sustainable, but as people move from buying CDs to buying digital music over the next 3-5 years, iTMS is going to be a major factor in Apple's brand recognition & consumer impressions. I can only hope that this translates directly or indirectly into more Mac sales & that Apple continues to be an industry leader in design for both computers and consumer electronics.

 

[runninmac]

(4) Absent a confirmation of the next milestone at <34 days, I don't see how they can possibly reach 1 billion this year. The rate would need to be ~2.3 million per day when it's currently running -- allegedly -- below 60% of that. Think about it this way... There are 10 months... If we ignore the 230 for a second, there appears to be a gain of 100,000 on the average for each month that passes. If that continues, then the average will be more like 1.8 million / day this year for a total of ~550 million additional leaving the year at 850 million sold to date.

I agree with you there is no way that they are going to reach 1 billion downloads by the end of the year even if they do a massive campaign with and insentive like they did for the 100,000,000. Im thinking that 500 is most accurate and 600 at most. sorry i hate being negative but i think that thats the truth

 

[samiam]

If we ignore the 230 for a second, there appears to be a gain of 100,000 on the average for each month that passes. If that continues, then the average will be more like 1.8 million / day this year for a total of ~550 million additional leaving the year at 850 million sold to date.

That's assuming a linear dependence, but if you look at the graphs posted earlier it's much closer to exponential than linear. The doubling time is decreasing a little over time, but it seems to be about 5.5 months. This works out to an increase in the daily rate by about .05% per month. Right now that's about 150,000 songs/day per month, but if the trend continues, it will be increasing by 500,000 songs/day every month at the end of the year.

So, if the current exponential slope continues, it will hit 1B this year.

But, if the exponential slope continues to decrease slowly, as is more likely since the number of stores is saturating, it will fall a little short.

And if Napster takes off, the rate could start to decrease.

 

[ipodmann]

thats great!
i bought my first songs through iTMS last night, J Los new album so I like to feel I have contributed towards that

Well you must be the one contributing big time to the increase in music sales, because I have not bought any songs for over a month.

 

[BlueDjinn]

Two More Expanded Charts

OK, call me crazy but I decided to plot out two revamps of the earlier graph: the first one is an enhanced look at total iTMS sales:

The second one adds total iPod sales into the mix:

 

[samiam]

there appears to be a gain of 100,000 on the average for each month that passes. If that continues, then the average will be more like 1.8 million / day this year for a total of ~550 million additional leaving the year at 850 million sold to date.

That's assuming a linear dependence, but if you look at the graphs posted earlier it's much closer to exponential than linear. The doubling time is decreasing a little over time, but it seems to be about 5.5 months. This works out to an increase in the daily rate by about .05% per month. Right now that's about 150,000 songs/day per month, but if the trend continues, it will be increasing by 500,000 songs/day every month at the end of the year.

Not that any of this matters a whit, but I should say that by "linear" I meant the time dependence of the rate, which would mean the actual number sold would have a quadratic dependence on time.

For fun I plotted the square root of the number sold, which would be a straight line if the dependence were quadratic. Note that the rate is increasing faster than quadratic, but 900M (30^2) does seem pretty close to the continuation of the trend.

I also tried a cubic dependence by plotting the cube root, and this gave the best straight line yet, with the extrapolation ending at about 800M.

For reference the semilog plot is also attached with extrapolation to 1B. Taken as a whole, this looks like the worst fit, but considering only the last 9 months or so, it looks pretty exponential.

 

[TheMasin9]

yay

This means more money for my stoxx. woot

 

[Xtremehkr]

A billion is not impossible. It could also depend on the quality of music that comes out in the next year. Releasing music to ITMS seems to be in style these days and enough artists chose to debut their work that way it could increase sales.

There are some big names releasing music soon too. Hip hop seems to hold a permanent position on the most downloaded list (for singles), and there are some big releases coming this year. 50 Cent is expected to sell an incredible amount of copies of his album that is out pretty soon. He's at the top (and bottom) of the list for today on the singles downloads.

I'm not a big fan of hip hop/rap/r&b, but a good year in music would affect the number of songs that I download. It's been a while since anything new and exciting came along.

 

[eazyway]

subtle

Someone correct me if I'm wrong, but is there a subtle difference here between the old numbers reported and the new? I thought the old numbers pertained to number of songs purchased, and new pertained to both purchased and downloaded. Obviously there's a lot of downloads of the Tuesday freebie songs. It'd be important to separate these numbers when estimating the health of the iTMS.

Dave

purchased and downloaded implies both for the same song...thus 300M purchased.

"purchased or downloaded" would mean the those purchased and downloaded plus those only downloaded

 

[rogo]

That's assuming a linear dependence, but if you look at the graphs posted earlier it's much closer to exponential than linear. The doubling time is decreasing a little over time, but it seems to be about 5.5 months. This works out to an increase in the daily rate by about .05% per month. Right now that's about 150,000 songs/day per month, but if the trend continues, it will be increasing by 500,000 songs/day every month at the end of the year.

So, if the current exponential slope continues, it will hit 1B this year.

But, if the exponential slope continues to decrease slowly, as is more likely since the number of stores is saturating, it will fall a little short.

And if Napster takes off, the rate could start to decrease.

The average per day still needs to be as high as I said, whatever the function. And while the total has been growing, the rate of increase is now approaching zero. It's hardly exciting to go from 40 days to 37 days to sell 50 million. This is especially true since the last two periods were 2/3 of the prior, approximately (40 vs. 62, 62 vs. 94). The last period was >90% as long. Not good for the curve.

I need the next data point to reset the models, because the 200, 230, 250, 300 data is way too lumpy to be useful.

 

[Poff]

And we still don't have it in Sweden...

Wonder when it will come to us scandinavians..

 

[sworthy]

How is everyone making these graphs? Care to walk me through it in excel (emailing your file would be nice too) - sworthington @ ups . edu (remove the spaces of course)

 

[samiam]

The average per day still needs to be as high as I said, whatever the function. And while the total has been growing, the rate of increase is now approaching zero. It's hardly exciting to go from 40 days to 37 days to sell 50 million. This is especially true since the last two periods were 2/3 of the prior, approximately (40 vs. 62, 62 vs. 94). The last period was >90% as long. Not good for the curve.

True, but data points tend to fluctuate, especially around Christmas, and differences fluctuate more, so it's never wise to extrapolate from two or three points. The total numbers don't deviate from the plotted trends by very much, and all the plots seem aimed at near 1B give or take a couple of hundred million.

And 37/40 is not an insignificant reduction: There are 14 intervals of 50 left to 1 billion, and (37/40)^14 is .33, so if the time continues to decrease by this amount for each interval, the last one will only take about 12 days. The total number of days for 14 intervals, based on this assumption, is 303 days, which puts us right at the end of the year.

 

[MOFS]

beatles songs should not be allowed to be covered. they turned a beautiful song into trash, just like a perfect circle did to Lennon's imagine.

worst. covers. ever.

What about Joe Cocker's (I think) cover of "With a little help from my friends" - ya know - the one from the Wonder Years?

Or the cover of "Across the Universe" at the end of Pleasantville?

Still - great news for Apple. Now we need more 60s, 70s and 80s stuff on the site and less latino stuff. And no more latino free downloads!

 

[MacDanny]

16 free iTunes songs

I FINALLY got my iPod Shuffle 1GB total. When setting it up you get 16 free iTunes songs if you open a new iTunes account. I hadn't heard about that before but that is a great way to introduce iTunes to new users!

Was anyone else aware of this???

 

[sworthy]

You actually get the songs with an existing account too... They even had a couple of good ones

 

[johnnyjibbs]

Good stuff. I've bought about 50 of those.

 

[d.perel]

Good stuff. I've bought about 50 of those.

50 iPod shuffles or 50 free songs?

 

[rogo]

True, but data points tend to fluctuate, especially around Christmas, and differences fluctuate more, so it's never wise to extrapolate from two or three points. The total numbers don't deviate from the plotted trends by very much, and all the plots seem aimed at near 1B give or take a couple of hundred million.

And 37/40 is not an insignificant reduction: There are 14 intervals of 50 left to 1 billion, and (37/40)^14 is .33, so if the time continues to decrease by this amount for each interval, the last one will only take about 12 days. The total number of days for 14 intervals, based on this assumption, is 303 days, which puts us right at the end of the year.

Samian, that's solid math at the end there. I agree, most fundamentally, that extrapolating from 2-3 data point isn't helpful. In fact, that's why when I came up with 37 days for this interval, I rejected it... It "felt" wrong. So I went with the aggressive number and was mistaken (there's a post somewhere hear where I went for 27-28 days).

I think it's the "give or take a couple of hundred million" that's the salient point. If it's 800 million, OK, but not as fabulous as Jobs walking into MacWorld SF next January and saying, "Since we launched a music store to work with these babies (he points to a rack with all versions of the iPod sold to date lined up in it), we've sold a billion songs."

So you see my interest in the next data point. Whether it's a 50 or a 100 (and I think they may stop announcing the 50s anytime), it'll be fascinating to see. Anything >34 days to the next 100 million would be a continuation / acceleration of the trend. So let's see what happens come March 19... Is there an announcement of another 50 million? Or do we wait until April to learn about 400 million....