Monday, October 15, 2012

Long-term Apple Segmented Revenue Visualizations and Projections

Two years ago I made the rather conservative (but somehow controversial back then) prediction that by this last quarter just ended in September and about to be reported next week, the iPad would have passed the Mac in trailing-twelve-months revenue. That post also included a far-reaching projection into the future for all product lines, as well as some specific estimates for FY2011.

As it turned out, iPad revenue has in fact already exceeded Mac revenue on a trailing-twelve-months basis but by the first quarter (Jan) of this fiscal year (at least I got the year right): for calendar 2011 the iPad logged in revenues of $24.9b while the Mac only managed $23b. Regarding the other specific predictions for FY11, I was slightly high for iPod revenue ($8.0b vs. $7.5b actual), iTunes ($6.5b vs. $6.3b actual), and Software ($3.2b vs. $3.0b actual). I came up way, way short on iPhone ($32.7b vs. $47.1b actual) and iPad ($18.0b vs. $20.4b actual), and a tad short on Mac ($21.0b vs. $21.8b actual) and Peripherals ($2.2b vs. $2.3b actual).

About the very long-term projection (shaded hazy part), a careful study of the chart shows some considerable differences, which I'll leave up to the reader to compare visually. Here's the update now with the FY2013 predictions (click on all charts to enlarge even more):

I also had a stacked one without the log-scaled vertical axis. Here's an update of it:

Here's another format which I first tried 2.5 years ago, just before the original iPad launch, in which I show the percent share of revenue for each product segment. Here it is updated with the last couple of years' actual results and this time I'm including the long-term projection (I wasn't this daring back then):

Notice the contrast between the highly dynamical history of self-disruption over the last decade, against the nearly static and boring projection I've drawn. This gives away the main motive of the projection as a mere extrapolation of the current state of affairs, an unambitious scenario that avoids trying to guess what might Apple attempt to displace next, such a guess driving a whole set of estimates which would not provide much reassurance except on some pet theory about the Next Big Thing, and most likely end up corrupting this rather clear and simple picture that faithfully (even candidly) follows along current trends. This past/future contrast is even more stark in the next chart showing the same segmented revenue shares data but from a similar perspective as the very first chart in this post (individual lines and log scale):

Finally, here's a new and pretty interesting one showing the adoption curves and installed base growth for each product line (see my Limits to Growth series for an in-depth explanation of these adoption models), including estimated waves for non-Mac PCs and iOS (I've used all iPods as an early proxy for "mobility" before 2008), as well as my long-term conservative projections for these adoption waves (which are the basis for the revenue estimates above). Enjoy:


CP said...

In the last plot, it looks like there is a negative number of iPods shipped. Is that a mistake?

Daniel Tello said...

Those are not units shipped, but units shipped to users who never owned an iPod before. Same with all the other lines in that last chart. Follow the linked article about adoption/replacement cycles (part 3 of that series has the iPod specifics) to understand how this "new users" metric could be negative (as it has been since late 2010).

pk de cville said...


Fortunately, I reviewed your logarithmic charts again today and found that I understood their magnitude only when I 'decrypted' the impact of the logarithmic scales.

Logarithmic scales highlight the decreasing rate of growth, but arithmetic scales are great for appreciating the ironic impact of the Law of Large Numbers!

(When a $200B revenue company grows by 'only' 20%, that's $40B of GROWTH!)

Would you consider adding arithmetic charts to your article? Thanks for your great work here.

Daniel Tello said...

The second chart is the linear version of the first chart, the third is the linear version of the fourth. The fifth is just linear, there's no logarithmic version of it.

Please don't call that growth notion the Law of Large Numbers, capitalized like that as if it was its true proper name. I call it "lol numbers" to distinguish it from the real one from probability (LLN).

Ovi said...


You are only modeling 33% revenue growth in 2013 over 2012?

Seems rather conservative to me.

Daniel Tello said...


I agree.