Tuesday, August 24, 2010
Gross margin key for EPS estimate
at 12:59 PM
Duh! Ok, maybe some didn't know this. Anyway, I've done some sensitivity analysis on my model by changing one variable at a time by 10%, and looking at the resulting percentage change in EPS.
First, I simply adjusted the variable by 10% and froze all the other variables, noted the EPS change, and moved on to adjust the next variable (resetting the previous one to how it was originally). Here's a couple of charts for that (click to enlarge):
So GM% has almost twice the effect on EPS than iPhone, which in turn has about twice the effect than any of the other major factors: operation expenses, taxes, iPad or Mac revenue, all of which are way, way more important than all the other stuff (iPod, iTunes, SW, Peripherals, and OI&E). I always knew that GM was important and that iPhone was the most important of the product lines, but I didn't know these precise relationships.
But then I realized the way I ran these scenarios treats each product line in isolation to the others, because I froze the revenue on everything else while changing just one product line. This approach goes against a very important aspect of Apple's business, something I took great effort in trying to take into account in my model, which is the inter-product synergies aka halo effects. A 10% higher than predicted iPod revenue could come from higher unit sales or higher ASPs. If it's from more units, then my model has built in some effect on Mac unit sales down the road (and iPads and iPhones to a lesser extent). So, Mac revenue should not have been frozen and the resulting effect on EPS would be further amplified.
To analyze these effects, I let all the inter-connectedness in the model play out when changing each of the product lines, and I even distinguished between changing a product's unit sales or its ASP or its gross margin, with interesting insights. Here's the four main lines:
Hmm... perhaps I need to increase the halo effect from iPhones. But, since it's already the most sensitive lever, I'd rather not exacerbate its effect. So, I'll just observe these relationships for now.