Daily Email from August 30th, 2016

The following email was sent to Above Avalon members on August 30th, 2016. 

Every story is written from the perspective of Apple. Each daily email contains 2-3 stories and is of comparable length and detail to the following email.

August 30th, 2016

This is the exclusive daily email for Above Avalon members.

Access the archive, communicate with other members, and discuss today's stories here. View this email in your browser by clicking here.  

Today's stories: Calculating the iPhone Installed Base, Segmenting iPhone Users by Upgrade CyclesNew 2016 and 2017 iPhone Sales Estimates

Hello everyone. Today's update will be dedicated to iPhone financials. I built a new iPhone sales financial model, and I want to go over some new numbers and assumptions for the iPhone installed base, upgrade cycles, and unit sales.

Before we get to the update, there are two other topics worth mentioning:

Apple Invites. Apple sent out invites yesterday for its September 7th keynote in San Francisco. I will be attending the event. The invitation (seen here) contains bokeh, which is the rendition of out-of-focus points of light, set within a very dark background. This likely hints at some pretty interesting iPhone camera upgrades. All signs point to a dual-lens camera for the iPhone 7 Plus. The invite also contains the words "See you on the 7th," where the "See you..." is another nod to cameras. The event looks to be centered on new iPhones (cameras and audio being the two big items) and new Apple Watches. We will have a much more in-depth preview of the event next week.

Apple/Ireland Update. The European Commission just ordered Ireland to recover $14.5B, plus interest, from Apple in back taxes. I am surprised. That total would seem to imply the Commission classified most of Apple's Irish tax structure as illegal (doesn't make much sense to me). As we talked about yesterday, late signals were pointing to a much lower amount. Even up to a few hours ago, the Irish Times was hearing the amount will be "billions of euro," so clearly there are some disappointed Irish government officials. Tim Cook is out swinging this morning (he just published a letter regarding the Commission's decision). At this point, it's unclear if Apple will actually ever end up paying the $14.5B. Since this news just broke, we will discuss the topic in greater detail tomorrow.

REMINDER: This email is from August 30th, 2016. The following estimates have since been updated. My current iPhone estimates are available for members here


Calculating the iPhone Installed Base

A few months ago, it became clear to me that I needed a revised financial model for estimating the iPhone installed base. My old methodology involved tracking changes to the iPhone installed base on a quarter to quarter basis.

For the first few years of iPhone sales, this is a pretty manageable task. However, with the iPhone 6 and 6 Plus being released in 2014 and the resulting explosion in sales, the model began to get very complicated. When it was clear that the iPhone upgrade cycle was also extending, trying to track the installed base quarter to quarter just wasn't feasible.

I developed a new financial model for estimating the iPhone installed base that relies on annual iPhone sales. This methodology makes it much easier to track the change in the iPhone installed base since there are fewer variables to estimate. In addition, I suspect this new model contains much more accuracy as I can segment the iPhone installed base according to when a user enters the iOS ecosystem (more on this shortly).

In order to estimate the iPhone installed base, we start off with simple iPhone unit sales by fiscal year.

  • 2007: 1M
  • 2008: 12M
  • 2009: 21M
  • 2010: 40M
  • 2011: 72M
  • 2012: 125M
  • 2013: 150M
  • 2014: 169M
  • 2015: 231M

Next, we divide these annual iPhone sales totals into two segments: sales to new iPhone customers and sales to iPhone upgraders (customers who were already iPhone users).

For the first few years, these calculations are rather simple as there weren't many iPhone owners in a position to upgrade in 2007 or 2008.

Jumping to 2010, things become more interesting. Apple sold 40M iPhones in 2010. In order to find the portion of those sales that went to iPhone upgraders, we look back to see which group of iPhone users were likely in a position to upgrade. It turns out that the 12M users who bought an iPhone 3G in 2008 celebrated their two-year anniversary in 2010. Running with an estimate of a two-year upgrade cycle, this means that a good portion of those iPhone 3G users likely upgraded to a new iPhone in 2010. Accordingly, Apple sold 10M iPhones to existing iPhone users in 2010. This is less than the 12M users who bought an iPhone 3G in 2008 because of churn. I include a 13% churn rate for the first upgrade cycle. We then subtract 10M from 40M to reach 30M as the number of iPhones sold to customers new to iPhone in 2010.

We repeat this exercise for each year, keeping track of which bucket of iPhone owners was likely in a position to upgrade. For example, iPhone 3GS users were ready for their first upgrade in 2011. In addition, original iPhone owners were ready for their second upgrade in 2011. As the years progress, the calculations become a bit more intense although it still beats the craziness found with quarter-to-quarter gyrations. By time we get to 2016, we have nine different buckets of iPhone owners to keep track of in terms of upgrade cycles.

After running this exercise for each year, we arrive at estimates for the number of iPhones sold to new customers. This data is valuable because we can then add each year together to reach an estimate for the current iPhone installed base.

iPhone Sales (Units) to New Customers

  • 2007: 1M
  • 2008: 12M
  • 2009: 19M
  • 2010: 30M
  • 2011: 54M
  • 2012: 95M
  • 2013: 88M
  • 2014: 82M
  • 2015: 102M
  • 2016: 106M
  • Total: 589M users (this is the iPhone installed base as of year-end FY2016 - the end of September).

One of my previous estimates for iPhone installed base was 554M users as of February 2016. This means my new installed base model is a bit more conservative. Diving into the model in greater detail, I think I am able to capture the full extent of repeat iPhone upgrade behavior much better than before. Since I am including a larger portion of iPhone sales to upgrades, this ends up reducing the number of iPhones sold to new customers.

Do these numbers pass the litmus test? Let's check with Apple's disclosures.

Here's Tim Cook on Apple's 3Q16 earnings conference call:

"[W]e added millions of first-time smartphone buyers in the June quarter, and switchers accounted for the highest percentage of quarterly iPhone sales we've ever measured. In absolute terms, our year-to-date iPhone sales to switchers are the greatest we've seen in any nine-month period, and our active installed base of iPhone is up strong double-digits year over year."

Those comments are extremely helpful. Cook went on to say the majority of iPhone sales in China were to people new to iPhone. My 106M iPhone unit sales estimate to new customers in 2016 represents an all-time high for Apple, which matches Cook's comments. In addition, China Mobile likely led to a boost in iPhone sales to new customers beginning in 2015 with the iPhone 6 and 6 Plus. My estimates reflect a 25% bump in sales to new customers in 2015. My previous model for the iPhone installed base had the number of iPhones sold to new customers falling in 2016.

My estimates also reflect a 22% increase in the installed base year-over-year in 2016, which compares to Cook's "up strong double-digits year over year."

EXTRA: If we bring the topic of the iPhone user base into this discussion ("user base" is different than "installed base" because the iPhone user base includes used iPhones and hand-me-downs), we are probably looking at something around 670M to 690M total iPhones out in the wild as of the end of September 2016. Once again, this is a bit lower than my earlier estimates, but it seems to better match Apple's latest disclosure of more than one billion Apple devices in the wild. The following numbers are my estimates (as of September 2016):

Segmenting iPhone Users by Upgrade Cycles

One issue that many analysts have been having is figuring out what to do with iPhone installed base data. If we assume the installed base is 589M iPhone users, how do we use that data to figure out upgrade cycles? This is important because once we have estimates for the number of iPhones sold to upgraders, we can then combine that data with sales to new customers to arrive at overall iPhone unit sales estimates.

In the past, one very quick method of using installed base figures was to simply multiply the total by a certain ratio depending on iPhone upgrade cycles. For example, if we assume the average iPhone user upgraded their iPhone every 2.5 years, we multiply the overall iPhone installed base by 40% (1/2.5) to reach an estimate that Apple will sell 235M iPhones (40% x 589M) to existing iPhone users each year.

However, there is a big problem with that calculation. It is assuming that all iPhone users are equal in terms of how they view their iPhones and upgrade patterns. In reality, that 589M iPhone installed base includes nearly 200M iPhone users who just bought their first iPhone within the last two years. I suspect these people are less likely to upgrade their iPhones as frequently as someone who bought the original iPhone or iPhone 3G/3GS. (There are much fewer of those early adopters included in the iPhone installed base).

My answer for navigating this trickiness is to segment the iPhone installed base by iPhone upgrade cycle. I run with shorter iPhone upgrade cycle estimates for iPhone early adopters (people who bought their first iPhone in 2007, 2008, 2009, and 2010) while I use longer upgrade cycles for those users who just recently bought their first iPhone. In addition, I included a lengthening process across the board to reflect a maturing iPhone product.

These are the iPhone upgrade cycle estimates that I used when calculating the iPhone installed base, grouped by the year a user purchased their first iPhone: (As an example, if my first iPhone was a 3GS purchased in 2009, I would be part of the 2009 line item.)

  • 2007: 2 years (the average user upgrades their iPhone every two years)
  • 2008: 2.2 years increasing to 2.4 years
  • 2009: 2.2 years increasing to 2.5 years
  • 2010: 2.2 years increasing to 2.5 years
  • 2011: 2.2 years increasing to 2.8 years
  • 2012: 2.5 years increasing to 3.1 years
  • 2013: 2.6 years increasing to 3.1 years
  • 2014: 3.1 years
  • 2015: 3.4 years

Notice how I have users who bought the original iPhone in 2007 as remaining on an average two-year upgrade cycle. However, for recent new iPhone buyers (2015), I assume an average 3.4-year upgrade cycle. Previous calculations that I did for the average iPhone upgrade cycle pointed to an overall average of 33 to 34 months. My current assumptions reflect an overall iPhone upgrade cycle of 36 months.

One caveat: The 106M new customers to iPhone in 2016 end up playing a very crucial role in this subject as they represent 18% of the entire iPhone installed base. If these users upgrade to a new iPhone much sooner than my 3.4-year average, the overall iPhone upgrade cycle will be more like 33 months.

New 2016 and 2017 iPhone Sales Estimates

Four months ago, I published my revised 2016 and 2017 iPhone estimates following Apple's difficult 2Q16 earnings report. My iPhone estimates were:

  • FY2016: 205M (down 11% from FY2015)
  • FY2017: 174M (down 15% from FY2016)

At the time, I stated that these estimates represented what I thought to be sales numbers that Apple had a very good chance of reaching (i.e. they were conservative).

After looking at my revised iPhone sales model, I think my FY2017 sales estimate was too conservative.

My new iPhone unit sale estimates are:

  • FY2016: 211M (down 9% from FY2015)
  • FY2017: 200M (down 5% from FY2016)

For my FY2016 estimate, the iPhone SE is helping to offset very weak iPhone 6s sales. I raised my FY2017 unit sales estimate by 14% due to one reason: I underestimated the number of customers new to iPhone in 2016.

I am now expecting Apple to add 75M new customers to iPhone in 2017. Previously, I was assuming Apple would add 40M new iPhone users in 2017. That 35M user difference was the primary driver for my higher iPhone unit sales estimate in 2017. The underlying reason for this discrepancy is that my revised iPhone installed base model has Apple adding significantly more new users in 2016. If Apple is adding more than 100M new people to the iPhone installed base in 2016, there is no reason to assume there will be an implosion to just 35M new users in 2017. Instead, I am running with a 30% drop to 75M new users. That is still a substantial decline, representing the steepest decline in new iPhone users that Apple would have ever experienced. The reasoning behind that can be found in my article "iPhone Warning Signs."

In terms of iPhone upgraders, back in May 2016, I estimated 135M to 145M iPhone users will upgrade to a new iPhone in FY2017. My revised iPhone model pretty much arrives at the same estimate for iPhone upgrades. This would be an improvement from 2016 when Apple did not do as well with iPhone upgrades (the 6s and 6s Plus just didn't connect with consumers as management expected). The reason I am including an improvement in upgrade patterns in 2017 is due to the strong growth in the iPhone installed base. In 2015 and 2016, I estimate Apple added close to 200M new people to the iPhone installed base (Apple can thank China). While a majority of those users will not upgrade their iPhone in 2017, even a small portion (20-30M people) can move the needle.

Looking out beyond 2017, the overall theme for iPhone will likely be weakening new user growth trends offset by improving iPhone upgrade patterns built on the back of iPhone installed base growth. While slowing new user growth will eventually catch up with upgrade rates, that likely won't happen for a few years. As long as iPhone users continue to upgrade their devices periodically (my current assumption is 36 months, on average), an implosion in iPhone sales over the next few years is off the table.

Each daily email is approximately 2,000 words and contains 2-3 stories (10-12 stories/week). 

Story topics include:

  • Strategy and business analysis
  • Financial modeling and estimates
  • Perspective and observations on current news events, competitors, earnings, and keynotes

To become an Above Avalon member, visit the membership page.