Executive Summary: This post went a little out of hand on me, so I want to sum it up here. Why did Apple launch its upgrade plan? My best guess is that they just realized this is something customers want, and Apple lopped off that little piece of friction. In addition, the upgrade program is going to get a lot of people to enroll in Apple Care, and insurance plus statistic means that is a nice little profit kicker. Finally, the upgrade program has the added benefit of further eroding the wireless operators’ relationship with their customers.

Details plus the numbers below –

I immersed myself in pricing details for the iPhone over the weekend (and here), and as a result ended up reading a lot about Apple’s new upgrade policy. Ben Evans wrote a good piece on the subject and I exchanged some ideas with him.

Apple watchers have been trying to figure out why Apple launched the policy and where the company may go next.

My take is that Apple has about a half dozen good reason for launching their upgrade program. It makes sense that a good strategy would have multiple underpinnings. On the other hand, maybe I am giving Apple too much credit and they just pulled this idea out a hat and only in hindsight does it look smart.

First and foremost, I think the upgrade plan is good for Apple customers – it is a very simple plan at a time when smartphone plans are anything but simple. Apple seems to be on a mission to remove friction form its customers interactions with its devices, and choosing a cell phone plan has become a process full of friction. Below, I am going to list an array of financial and strategic reasons for the plan, but this friction removal could be the sole justification for their action.

A second motivation for the plan is to further erode the relationship between the mobile carriers and their subscribers. Many of us in the US are tied to their carrier by lengthy contracts. Moving that contractual relationship to Apple gives us one less reason to interact with the operators, and moves us one step closer to complete ambivalence about choice of operator. Ambivalence being another word for commoditization.

I also think there are some interesting financial implications for the plan. The upgrade plan is 0% loan to buy the iPhone. Apple currently earns 1.5% interest on its $203 billion cash balance (calculated as $766 million in interest income in the last quarter, annualized). The NPV math on that means that Apple is foregoing $15 per phone, a tiny amount when their profit per iPhone is several hundred dollars.

Ben Evans, among others, also wondered what happens in a year when everyone on the plan upgrades. Apple will suddenly have a huge inventory of used phones. It turns out that the used phone market is much bigger than most people realize. Apple will now effectively control that market by having the biggest inventory.

Going back to my earlier math, the upgrade payments for a 128 GB phone for one year are $489 (12 times $40.75). In a year, Apple will let people on the plan trade in the old device for a new. Assuming the new device is full priced at $849, than means Apple is effectively paying users $360 for their old device ($849 minus $489). Today, a refurbished iPhone 6 costs $550 on eBay. Assuming this math stays constant, Apple could resell its traded-in phones and make $190 profit. (Or more likely, it will cost them about $40 to refurbish the phone, so call it an even $150.)

This upgrade plan also signs everyone up for Apple Care. Having been a student of the Godfather of Behavioral Finance, Richard Thaler, I know that in-store insurance programs are a terrible purchase decision for consumers. Nonetheless, I somehow always end up buying Apple Care. (Maybe it’s the Apple Reality Distortion Field in full-effect.)

Warranties historically have been a very lucrative product for electronics companies which is why it is so bad for consumers. Apple’s program is a better than most, but the math is still on their side.

Apple Care costs $129. There is no hard data on claim rates, but we can make some guesses based on Apple’s Financial Data. First, a good primer on the subject is this article in Warranty Week Magazine. They do an interesting analysis. (Note that their estimate of AppleCare revenues have a big accounting flaw in them, so I am not including that data.)

Then I took a look at Apple’s Financial Statements and unit shipments. Apple reports how much they reserve for warranties, meaning dollars that they put aside to save for each device sold  in case they have to pay for defective phones or other warranty costs.

Apple’s warranty data is all over the map, and they changed their accounting practices on the matter in 2007. So treat all of this with a grain of salt. I am talking very rough numbers here, and I am sure that there are all kinds of accounting nuances that I missed because they are either buried too deep in the financial statements, or just not there at all.

But a couple things are clear. They reserve less for iOS devices than for OSX devices, which makes sense since these devices cost less. Intuitively, the cost of replacing a $500 iPhone should be less than the cost of replacing a $2,000 MacBook. I try to separate those two with some calculations at the bottom of my data here.

So rough math, Apple reserves about $20 per iPhone/iPad for warranties. The average sale price of an iPhone is about $600 which means that Apple is reserving 3.3% ($15/$600) for each iPhone.

Now let’s take this a step further. Only 3.3% of iPhones are returned for defective parts, it is reasonable to assume that claim rates on Apple Care are in the same range. To be conservative, let’s use some round numbers, and say its 10%. That is 10% of Apple Care customers come in and ask for a new phone. Apple will have a steady stream of upgrade-plan iPhones coming, and let’s assume the average cost of those is $300 per device (see the math above for the calculation and then average across different size phones). So 10% of $300 is $30 per device, plus $20 warranty reserve, plus $15 in that 0% loan they are providing consumers. All this adds up to $65 in costs, set against an additional $129 in AppleCare income, for a profit of $64. Not bad work.

Now before anyone starts drooling over all the profits Apple is going to make from this upgrade plan, let’s keep one thing in mind. By my math, only 10% of used iPhones will be used as refurbished insurance replacements. There is probably an equal amount that will go to other, outside insurance schemes from the carriers or others. That still leaves a big pile of used iPhones coming to the market.

I will not go into all the math here, but in rough numbers, the profits from selling all that extra Apple Care, plus the ability to re-sell refurbished phones to other insurance providers (i.e. the carriers) as well as direct to consumer sales of refurbished phones (which is something all the US operators offer) could cover about 70% the costs of the upgrade program. Which means that if they can sell about a third of the total traded-in phones into emerging markets then the upgrade program reaches breakeven. All of my numbers strike me as pretty conservative, so it looks like the upgrade program should be comfortably break even in its own right. Now, for any publishing sell-side analysts reading this, if you can, see if Apple will tell you how many people sign up for the plan.

So where does all this get us?

The key underpinning of this upgrade is that it is going to get a lot of people signed on to Apple Care, and through the beauty of statistics and insurance, that should be a nicely profitable business. Add in a few more revenue streams, and the upgrade program in total will probably be reasonably profitable for the company. Not a whole new product line, but a nice kicker on an already profitable iPhone business. And then to top it all off, it essentially bakes in customer loyalty (which is a minor point since iPhones already enjoy something like 95% retention), and scores a strategic advance against the carriers.

But like I said at the outset, maybe they are just doing it to make our lives a little simpler.