Tuesday, March 19, 2019

UPDATE- Apple (AAPL): Number of Shares Sold Short Explodes

In the last Financial Alchemist report, I wrote about the massive increase in the number of AAPL shares sold short according the 2/28/19 report. I speculated that it had to be a single entity, likely a hedge fund that was at the center of such a large change in short interest. However, what did not occur to me was the possibility that this was linked to Apple's accelerated share repurchase program. This involves an investment bank essentially going short the stock as part of the process. Kudos to genius Daniel Tello at AAPL Model for pointing this out.

The ASR process:

  1. Apple enters into an ASR agreement with an investment bank and pays full cash amount upfront.
  2. The bank borrows AAPL shares from portfolios of its brokerage clients and/or other brokerage houses, using the cash from Apple as collateral. The investment bank is now effectively SHORT.
  3. The investment bank delivers Apple the borrowed shares, equal to 85% of the payment, and outstanding shares are reduced immediately.
  4. The investment bank goes out to the open market and buys AAPL shares to cover its short position returning shares that it borrowed. 
  5. When the ASR is complete, the bank settles with Apple the net remaining shares owed, further reducing outstanding shares. 

We should expect to see a significant reduction in Apple's share count when it reports 2Q19 results. The effect on EPS is less known since the share count used in the calculation is a weighted average of outstanding shares over the quarter. However, it will be quite lower, and the ASR does not reflect the shares Apple might be buying back on its own using a 10b-5 plan.

Sunday, March 17, 2019

Apple (AAPL): The Number of Shares Sold Short Explodes According to the Latest Report

The number of AAPL shares sold short as of 2/28/19 increased by 56.9M or 143% from the previous report date on 2/15/19. Short interest stood at 96.8M shares representing 2.1% of the float versus 39.9M shares or 0.9% of the float 2 weeks prior. This is an unprecedented move for as long as I can remember. For the 21 two-week periods prior to the most recent report, the average absolute change was 3.2M shares sold short with a standard deviation of 2.4M shares. The average. absolute percentage change is 7% with no two-week period every exceeding 20%. Thus, for the last year there the period to period changes have been quite small. 

So what happened in the last two weeks of February that led to the short interest more than doubling? Even when Apple announced it weak guidance and that it would no longer report unit sales causing the stock to tank, short interest only increased 15% or 5.2M shares. The stock has been on the rise and sentiment has seemed to improve. Thus, a change of this magnitude in AAPL short interest doesn't make any sense.

What could make sense is that a hedge fund, or a few people with a lot of capital placed a massive short bet. It's highly unlikely that large number of typical investors just all of sudden simultaneously decided to go short on AAPL. That would imply that overall sentiment had soured which doesn't seem to be the case. There have not been any negative news or significant developments, and the stock has been rising. It is hard to fathom that this increase was caused by a large number of investors. Most likely a person(s) is making a huge bet that AAPL is going lower. Perhaps as high (or even higher) as 50M shares, or $8.7B in value. If they haven't already covered they are enjoying a $600M paper loss. If they have covered, they don't have anything to show for their efforts since AAPL has not been lower in the days since the last report. It will be interesting to see where AAPL short interest stands when the 3/15 report is released. 

Saturday, March 9, 2019

APPLE (AAPL): Estimating iCloud Revenue and Paid Subscribers

For some of Apple's services, details such as revenue and number of users are better known than others. iCloud is one such service that Apple has provided little disclosure making the number of paid subscribers and revenue a mystery. Using various data points and tidbits of disclosures, I estimate iCloud revenue was $3.3B in FY18 and ended with 155M paid subscribers. Paid subscribers has grown to ~170M as of 1Q19.

Apple’s Services segment has grown faster than 20% annually for the last 3 years. While still a relatively small portion of total company sales, Apple Services generate predictable, high margin revenue streams. Apps and licensing are the two largest components, generating about 50% of segment revenue. AppleCare, Apple Music, and iCloud round out other major services. Revenue from Apple Music and AppleCare is not difficult to approximate. Apple disclosed AppleCare revenue for FY14, FY15, and FY16 ($3.2B), and indicated revenue was flat in FY17. In FY18, AppleCare growth ignited as Apple expanded distribution to resellers, as well as bundling with iPhone financing plans. For Apple Music, management occasionally reports the number of paid subscribers and Y/Y revenue growth rates. Taking the monthly price for a Music subscription and multiplying by the number of paid subscriptions, Apple Music revenue can be estimated. 

Apple has never provided any meaningful hints as to the potential size of its iCloud business, revenue-wise or number of paying users. It's likely that iCloud's contribution to total company revenue was so small it was never worth mentioning. That changed a couple years ago, when management regularly touted iCloud's performance on conference calls. Since 2Q16, almost every quarter TC has stated iCloud revenue grew very strong double-digits, or very very fast, and a couple quarters growth was over 50%. However, without a reasonable estimate of base revenue for a starting point, that commentary is not of much help.   

Unsurprisingly, estimates for Apple's iCloud business are all over the map, ranging from less than $2B to more than $5B in FY18. Based on the calculation of paid subscribers, I estimate for FY18 iCloud generated $3.3B from an average of 136M paid subscriptions. Apple ended FY18 with 155M paid iCloud subscribers. That number will rise to $4.7B in FY19; and by 2025, iCloud could potentially generate over $15B in revenue annually. 

iCloud revenue is the product of the number of subscribers and the amount they pay (for a monthly plan). That monthly amount users pay is one of three possibilities (plans) which in aggregate forms a blended amount- the ARPU (average revenue per user). The challenge is coming up with a reliable estimate for paid users.

Estimating iCloud revenue is dependent on first arriving at a solid estimate of paid subscribers. For paid subscribers, there are not any concrete numbers known. Two years after its introduction, iCloud reached 300M users (paid & free). In early 2016, an Apple executive said there were 782M total iCloud users. Today that number could be as high as one billion. The number of total users is of little use since the portion of paying users could be anywhere in between zero and the combined total, presenting a wide range of error.

Taking information that Apple has disclosed, such as total number of paid subscriptions (includes Apple Music, iCloud, 3rd-Party), along with Y/Y growth rates for iCloud revenue and Y/Y growth for number of 3rd-party subscriptions, we can back into a number of iCloud subscribers.

Estimating iCloud ARPU is more art than science. There is not any hard information to incorporate; surveys and developing a logical narrative are the only available guidance. iCloud plans are offered in 3 monthly tiers - $0.99 / $2.99 / $9.99. Using intuition and data from surveys, we can calculate ARPU based on estimates of the mix (subscription tiers). 

Every iCloud user receives 5GB of storage for free, a paltry amount that Apple has never increased. This pushes users to pay for extra storage. There are 3 tiers- 50GB, 200GB, and 1TB for $0.99, $2.99, and $9.99 per month. The vast majority of paid users fall in the low tier. Generally, users start at the free tier and then move up the tiers as they require more storage. 5GB is not ample for hardly anyone. Users either pay for more storage or choose to store only partial amounts of their data. 50GB is sufficient for most users. However, over time, users who take lots of photos and videos will find themselves needing 200GB. The 1TB tier is the least popular; it represents a single digit percentage of paid users and is typically includes family plans and photography/graphics enthusiasts. I estimate ARPU is approximately $2.00/month which is based on a percentage mix of 65/30/5 for low/mid/high tiers.

Apple reported in Jan-17 (1Q17) that it had 150M paid subscriptions which includes Apple’s own services and 3rd-party content offered on its stores. Therefore, the max that iCloud paid subscriptions could total is less than 150M at the end of CY16. Factoring in Apple’s paid music streaming subscribers of 20M, that number has to be less than 130M. In addition, knowing that there were many 3rd-party subs at 1Q17, iCloud subs would be a good bit less than 130M. How much less? 20M, 50M, 80M? It depends on the number of 3P subscriptions, which like iCloud, is not known. We only know that as of CY16 end, neither iCloud nor 3P subscriptions exceeded 130M, individually and collectively. 

In 3Q18, Apple PR stated that the number of 3P subscriptions increased 95% Y/Y.  In addition, the company stated on conference call that iCloud revenue grew by more than 50% that quarter. Additionally, Apple reported total paid subscriptions increased roughly 62% from 185M to 300M. Using iCloud revenue growth as a proxy for subscription growth, we solve for the two base amounts by finding the proportions A + B  which total 158M in 3Q17 (185M including Music) which increase 95% for 3P (A) and 43% for Cloud (B) to amounts (A)2 and (B)2 in 3Q18, which total 255M (300M with Music). 

A + B = 158
(A * 1.95) + (B * 1.43) = 255
(A+B+Music) * 1.62 = 300

-Total Subscriptions & Y/Y Growth Rate
-3rd-Party Subscription Y/Y Growth Rate
-Apple Music Subs & Growth Rate

-iCloud Revenue Y/Y Growth Rate

-Number of iCloud Subscriptions (3Q17)
-Number of 3rd-Party Subscriptions (3Q17)

By using 95% for 3P growth and 43% for iCloud, we can solve for the base amounts in 3Q17 that would produce 62% total growth for paid subscriptions. The result is 103M iCloud and 55M 3rd-Party subscriptions in 3Q17, increasing to 148M and 107M in 3Q18, for iCloud and 3P, respectively. 

For periods other than 3Q17 and 3Q18, I use management's comments regarding iCloud growth to estimate the other periods in FY17 and FY18. Essentially, iCloud has grown 50%+ in some quarters, and others were colored as “very strong double digit growth.”

The growth rate for iCloud subs is estimated from iCloud revenue growth rate. The sub growth rate is less than the 50% revenue rate to account for increases to ARPU. It is highly probable that the bulk of iCloud revenue growth stems from sub growth opposed to a considerable increase in ARPU. Even as some users switch to higher-priced plan, the effects on ARPU are diluted by the addition of new users on the low tier plan. Sub growth likely contributed ~40% and increased ARPU added ~10%+ to the iCloud revenue growth rate.

3rd-Party subscriptions exploded after Apple began allowing subscription billing across all app categories. Originally 3P subscriptions were mostly limited to apps with an existing subscription business for content or services outside the App Store. Today, subscriptions are replacing in-app purchases for the monetization apparatus in many apps. Instead of a one-time payment to unlock premium features in-app, the features are only unlocked for a set period of time (subscription period). The ARPS (average revenue per subscription) is much lower, typically a few bucks for a subscription period of several months to a year, opposed to traditional subscriptions that usually cost $10-$20 per month.  

Apple reported that users spent $1.55B and $2.7B on 3rd-party subscriptions in CY15 and CY16, respectively. This information is not particularly of great use since ARPU is unknown and using a rough estimate (based on average subscription prices) is imperfect since the purchase volume number covers the entire year thus contains subscriptions that were only active 1,2,6, etc. months. One-half of that $2.7B theoretically could have been earned in the last 3 months of the year. While imperfect, if we can infer a run rate from using the annual amount as an average- gross up $2.7B to $4B, and assume annual ARPU of $100, the result is 40M 3P subscriptions as of Jan 1, 2017. However unreliable, it is a starting point. The ballpark estimate of ~42M 3P subs (based on CY16 purchase volume) at the start of CY17 fits reasonably as it equates to 88M iCloud subs at start of CY17. When possible, I try to use various methods and calculations to test the reasonableness of the estimates.

I believe iCloud could grow to a $17B annual revenue business by 2025 as the number of paying users and ARPU increases. In the future, cloud services could potentially become as important as internet or wireless service. As an individual's life becomes increasingly more digital, there is a need for storing that digital information. What we think of cloud services today will different significantly from cloud services of the future. As Apple's installed base continues to growth with more users converting to paid users coupled with paying users upgrading storage plans, iCloud revenue will keep rising. One possible headwind to revenue growth could be sluggish ARPU growth stemming from price reductions and/or increased storage allotments. I expect Apple will increase the size limits on its tiers, but storage needs will be increasing as well muting the overall impact on ARPU growth. 
Apple has over 1.4B active devices, including 900M iPhones in its installed base which is growing ~10% annually. I would put the number of unique users around 850M. With 170M paid iCloud users at the end of 1Q19, penetration equates to just 20% of unique users. This provides a long runway for iCloud revenue growth and the key to reaching $17B in 2025.