Why Q3FY08 iPhone revenues won’t be up very much

This excerpt of material intended for MDJ and MWJ readers was sent to a subscribers-only mailing list last week, but we wanted to get it out in the public in the hopes of avoiding the kind of uninformed Apple-bashing cycle that seems to start with every quarterly release. Apple’s recognized hardware revenue for iPhone sales in the June 2008 quarter won’t be that much higher than in the March 2008 quarter, and it’s not because Apple didn’t sell many iPhones (well, people might have been waiting for iPhone 3G, but surely many thousands of people bought the original iPhone in the quarter anyway). What’s more, Apple explained three months ago why this would happen. Here’s the scoop.

From Peter Oppenheimer’s statement

Among many other things in his prepared statement during Apple’s financial conference call of 2008.04.23, chief financial officer Peter Oppenheimer revealed this news concerning Apple’s accounting for iPhone revenue:

Because we announced the specific new features to be included in the iPhone 2.0 release, and plan to provide them to iPhone customers as a free upgrade in late June, we will delay the start of revenue recognition for all iPhones sold on or after our March 6th announcement date until the iPhone 2.0 software is delivered. The revenue and cost of sales associated with these iPhones will be recognized ratably over the remaining terms of their respective 24-month estimated lives. Revenue recognition of handsets sold prior to March 6th and payments from carriers has not changed as a result of our announcements.

The analysts inquire

Naturally, during the conference call, analysts wanted to know much more about this concept. Keep in mind that at this time, Apple had said the new software would arrive in June. As you know unless you’ve been on another planet, the iPhone 2.0 software shipped simultaneously with iPhone 3G on 2008.07.11.

Mike Abramsky of RBC Capital Markets asked for more information about why Apple would delay revenue recognition of iPhone units until the 2.0 release, and how analysts should think of that. Peter Oppenheimer responded by repeating most of the material on this subject from his prepared remarks, and emphasized (twice) that he was repeating it, so we’re not going to repeat it here. We will note one sentence, though: “Once we deliver the [iPhone 2.0] software, we will resume … revenue recognition of the phones on what we sold between March 6th and that date, and begin to recognize, as we had in the past, iPhone revenue on a go-forward basis.” When Oppenheimer finished repeating his prepared material, Abramsky asked for confirmation that “really, what we are looking at it just the exclusion of hardware revenue next quarter … from iPhone.” Oppenheimer confirmed: “Until the [2.0] software is delivered.”

This is slightly tricky, but a later question by Piper Jaffray’s perpetual Apple bull Gene Munster helped clarify it. Munster asked if Apple would be recognizing the costs and expenses related to iPhone sales between 2008.03.06 and the date that iPhone 2.0 appears, even if it would not be recognizing revenue from handset sales during that period. In this context, recognize is an accounting term meaning “we’re now listing this on our books as revenue.” If you purchase a five-year magazine subscription, the publisher can record your entire payment as “sales” in the month when it received the money, or it can put the money in the bank and only list 1⁄60 of it as “sales” each month for the next five years. In the first situation, the publisher “recognizes” all the revenue upon receipt. In the second one, the publisher “recognizes” the revenue ratably defined here to mean “in equal portions for the term of the subscription.”

Apple could recognize the cost of iPhones sold (the materials used to make the handsets, the packaging, the shipping) and the administrative cost of getting them sold (product development, advertising, licensing fees for light and airy guitar music) at the time the phones were sold instead of ratably over a 24-month period. Doing so would mean Apple’s expenses would stay high while iPhone revenues went down during the June quarter, and Munster wanted to know if that was the plan.

Peter Oppenheimer said, “No, Gene. We will delay recognition of both the revenue and the product costs on the iPhone from — beginning on March 6th until we release the [iPhone 2.0] software, but we will continue to expense our operating expenses around selling and advertising and engineering as period costs, as we always have.”

Munster then asked if Apple would be charging customers who owned iPhones before 2008.03.06 for the upgrade. Same answer from Oppenheimer: “No. As I said in my prepared remarks, we will be providing that software to the then-current iPhone customers for free.”

With all other possibilities ruled out, that led Munster to the real question: “If everybody is going to get it for free, why are you deferring [revenue for iPhones sold] after March 6?” Oppenheimer replied, “Because the customers that purchased the phone on or after March 6th presumably were aware of the software that we told them that we were going to come out with, and that they were going to get it for free. And as a result of our announcement, the way you properly account for that is to defer the revenue, and that’s what we’re going to do.”

This argument goes back to early 2007 when Apple revealed (as take-apart fans had already discovered) that late 2006-model iMacs were capable of draft 802.11n wireless networking, not just the 802.11b/g networking the company had promised in the specifications. All you needed to turn on the faster networking was new software — and Apple charged US$2 for it, or made it free with the company’s own 802.11n AirPort Extreme Base Station hardware. You may remember the outcry.

Why the charge? Apple had just spent much of 2006 refuting charges of shady bookkeeping in the options backdating matter that’s still the subject of civil litigation today. Now eager to avoid even the appearance of impropriety, it turned out Apple had another lawsuit magnet on its hands with previously undisclosed features. If the customers knew the hardware was supposed to provide 802.11n networking but Apple didn’t implement it until later, then foes could argue that Apple should not have recognized revenue for however much the 802.11n networking feature was “worth” until the enabling software shipped. If the feature was worth US$10, then Apple should have taken out US$10 from each iMac sold before releasing the AirPort 802.11n Enabler, and recognized that chunk only when the software was delivered.

Failure to do so is arguably “fraud” — accepting money from customers for “promised” features that weren’t implemented. The argument can then be extended to say that if Apple knew about features it intended to implement, it should similarly delay recognizing that revenue. This, of course, is an accounting nightmare, one Apple chose to avoid by charging an extremely small fee to deliver the software (and by building it into Leopard), eliminating any basis for a claim that Apple had “pocketed” money from non-existing features. Technically speaking, this is not a “rule” of corporate accounting, but failure to cover tracks like this in the 21st century just invites class-action lawsuits, and no company wants that. It’s why Apple’s thinking on Mac OS X updates seems to have shifted to “no new marketable features, only bug fixes and very minor improvements.” If Mac OS X 10.2.2 had been released in the current climate, it’s not at all clear that Apple would have included Journaled HFS Plus — the company might have waited to ship it in Panther to avoid accusations of shenanigans.

But now we come to products like Apple TV and iPhone, products that Apple always intended to upgrade for free with new features during the product’s useful life. That opens the same can of worms from the other end: how much of a US$399 iPhone sale can you “recognize” if you know you’re going to add an iTunes Wi-Fi Store and better geolocation and, eventually, third-party software — someday? Apple punted: it treats both products as “subscriptions” over the estimated useful life of 24 months, recognizing revenue ratably over that period.

Note that Apple does not do this with the iPod Touch, which is in the same technological and accounting situation. That’s why this January’s iPod Touch software upgrade cost US$20, but was included free on new units, even though iPhone owners got the same new features for free. Apple probably treats the iPod Touch differently because to treat it as a subscription product, the company might have to separate and announce iPod Touch unit sales numbers, and it does not want to provide that kind of information to its competitors.

And now, Apple finds that having closed the accounting can of worms, it’s open again. Even though the iPhone is already booked as a subscription, the company announced major new features coming for free in June. That’s exactly the same problem as before — selling a product today that doesn’t provide all the features that Apple has said it will have. Rather than try to determine “how much” of a current iPhone’s value is tied up in expectations of iPhone 2.0 software, Apple is not recognizing any iPhone revenue (or costs of sales) between the date of the iPhone 2.0 announcement on 2008.03.06 and that magical future date when the software ships.

Since the schedule slipped, and iPhone 2.0 software did not ship until after the end of the June quarter, then Apple therefore booked no new iPhone hardware sales revenue during the June quarter, since 2008.03.06 was in the March quarter. It also means that while Apple listed the unit sales of iPhones in the last 23 days of the quarter (2008.03.06 through 2008.03.29), it did not recognize any revenue from those sales. If you assume (for simplicity) that iPhone sales were linear throughout the quarter, that’s right at 25% of iPhone sales “revenues” that don’t appear on Apple’s data summary or financial statements for the March quarter.

Because of this, Apple’s iPhone hardware revenue for the June quarter will be just slightly higher than it was for the March quarter. The number of iPhones that Apple was using for its daily hardware revenue calculation increased every day (because Apple sold more iPhones every day) until 2008.03.06, when Apple froze the number for the above reasons. The number stayed frozen between 2008.03.06 and 2008.07.11 when iPhone 2.0 and iPhone 3G shipped. Since that frozen number of iPhones used in the daily calculation is higher than on any day before 2008.03.06, the overall recognized June quarter revenue will be slightly higher than March quarter revenue — but nowhere near what it would have been had all this not happened.

What happens now? On 2008.07.11, Apple started recognizing all that suspended revenue ratably — divided evenly over the remaining 24 months of the handset’s expected useful (for accounting purposes) life. If, for some bizarre reason, you purchased an iPhone on 2008.07.01, then Apple will recognize your phone’s revenue on a daily basis over 720 days: two non-leap years (730 days) minus the 10 days of limbo. If you purchased an iPhone before 2008.03.06, Apple’s still recognizing your handset revenue on the same basis as it was before.

Presumably, although no one asked, Apple did the same thing with Apple TV sales between Steve Jobs’s announcement of Apple TV “Take 2” at Macworld Expo on 2008.01.15, and the software’s release on 2008.02.12. Isn’t accounting fun?