Wednesday, December 22, 2010

Apple Financial Floor Plan

I've made the following assumptions:
- Approximate individual product GMs: Mac, iPod, iPad, and Peripherals around 30%; iTunes about 10%; SW and iPhone about 60%.
- Tax and operating expenses by individual product track company-wide ratios of GM.

Most likely mine isn't "totally self explanatory" either.


Horace Dediu said...

Keep it coming. I will say in my defense that I used only Numbers. Was your chart done with Excel?

Daniel Tello said...

I used Excel to derive some pixel dimensions from the financial data, and just used Preview (believe it or not) to render it into the rectangular blocks. It involved heavy copy-pasting with the Rectangular Selection tool which gives real-time pixel-level measures. I'm sure Photoshop or some such would have made it a breeze.

Anonymous said...

I understand the size of the billion box, but your vertical and horizontal axis are inversely related, are they not? Is there any way to attach labels or metrics to each axis, so to speak.

Daniel Tello said...

Anon, it's just the surface area that corresponds to any of the dollar amounts. So both axis measure the same thing (square root of dollars if you want to get technical). The only distinction between horizontal and vertical is the concept by which the dimension is aggregated: as the different revenue sources vertically, and as the P&L line items horizontally.

Anonymous said...

What I don't get is the meaning of the height of the cells. The 2010 cells for iPhone and iPod are about the same width, but very different heights. Why not make all cells the same width? Or the same height? or square?

For instance, you chose to have the iTunes cell very thin, and very wide. Why not twice as thick and half the width?

Daniel Tello said...

I see what you mean. The width was set based to the 2010 net income, so that all product lines had the same width for net income (and thus the same width for pre-tax and gross income). This (net income in chocolate brown) was set at 60 pixels wide, the same as the $1b square in the legend. Because of my assumption mentioned above that taxes and opex for each product track company-wide ratios of gross margins, these values also align in width across product lines in 2010. This way you can tell the different amounts of net income in 2010 by the heights of each horizontal bar. Obviously GM for each product is different, so the width of the COGS (light creamy stuff) to complete the whole revenue for each product varies. This further helps with visualizing the growth in net income for each product in 2011 by keeping the same height and looking at the increased width (I included a reference line in the 2011 net income side at 60 px where the 2010 net income would fall).

The full iPod cell made of all four colors (total revenue) in 2010 is much wider than iPhone, showing the higher gross costs for iPod, but it's a fraction of the iPhone height due to the much higher sales (and profit) volumes in iPhone. In 2011 the total width in iPhone nearly doubles to almost the same as iPod while iPod remains at the same width (flat growth), and yet the iPhone height is still multiple times that of iPod so the total surface covered by iPhone (revenue but profitability in dark colors even more so) is multiple times that of iPod.

If I had chosen to make iTunes twice as thick, then the net income in brown would have been half as wide (30 px), but since I decided to make all products' 2010 net incomes at the same width, then the iPhone bar would have been twice as tall (as well as all the others but iPhone is the extreme in height), and the whole chart would look too tall and too narrow (except for iTunes), and it would make the net income comparison with 2011 less apparent. It would be the same chart, in essence, except for a change in width/height ratio. Notice the ratio I chose is not perfect as I had to reposition some of the iTunes stuff. But making the whole thing narrower to try to fit all of iTunes in it's proper horizontal bar would imply a much taller/narrower chart for all other products.

Alternatively, I could've chosen to normalize all the widths to 2010 revenues instead of net income. Then the different product profitabilities would probably show up more clearly, as well as the revenue growth in 2011 (instead of profit growth). I just preferred to focus on profits, as that should be a more direct measure of stock valuation.

Hope this explains my somewhat arbitrary choices in widths and heights. Keep asking if you still have questions.

Anonymous said...


Thanks thats the data and answer, I was looking for, and now understand how and why you did it that way. Might be curious to see what "normalization" might do, but respect the fact that you've got other things to do. BTW, we've e-mailed before, I am a big fam of your efforts. Any chance, you're familitar with Tufte's work at Yale, e.g. Visual Display of Quantitative Information? I am a big fan of his concepts as well.

Daniel Tello said...

I'm just a little familiar with Tufte's work (read some of it a long time ago). Wish I had the time to study more. Tough choice, beach, study, beach, study... hmm. Skying, big boring book... aahhg knowledge is overrated anyway ;)