Description

The risk-reward of owning Alphabet is better today than a year ago.  The stock is up 50% since then and earnings are up 40%, so the valuation hasn’t changed substantially.  I think the upside over the next two years is greater, and the downside lower than I thought it was a year ago.  Many things have changed for the better.  First, new management has shown greater expense and capital discipline.  Incremental EBIT margins have averaged 47% over the last four quarters.  Year-to-date capex is down 9% even while revenue has grown 20% (an acceleration from 2015).  $10 billion of stock repurchases have been authorized, whereas no buybacks had been authorized prior to October 2015.  Day to day execution has improved, Sundar is well liked internally, and he has freed the founders to focus on long term strategy and innovation.  Second, disclosure has improved around Other Bets losses, a nagging source of uncertainty prior to 2016.  Third, new revelations about the size and growth rate of Google Cloud Platform (GCP) have caused me to meaningfully reassess the potential value of GCP.  Fourth, I’m encouraged by the Pixel effort to produce an integrated Android phone.  I didn’t previously model any direct handset revenues, but Pixel appears to compare well with iPhone even on the first attempt, and is inarguably better in important ways.  Like cloud computing, the handset market is vast and valuable and can be a meaningful portion of Alphabet’s enterprise value in the near future.  The value of and moats around Search increase daily, and Alphabet has at least three very large and high probability projects which can generate meaningful enterprise value in the future.

Q3 and the near term

What was universally believed about Q3, including by me, is that Search revenue would slow due to the lapping of the addition of a third mobile ad on highly commercial mobile searches.  This was partially true as the third mobile ad compare was probably a 200-300 bps headwind.  However, global Search growth probably accelerated by 100 bps versus Q2’s growth rate.  Offsetting the difficult compare in the US was the addition of a fourth paid ad on highly commercial mobile searches (1.5% tailwind), and ads in Maps (1-2% tailwind).  Prior to six months ago, Maps was virtually un-monetized, despite being one of the most valuable properties on the Internet.  I’m optimistic about future Maps revenues, but I can’t comfortably quantify the potential and so haven’t modeled it.  Also offsetting the US Search slowdown was a 600 bps acceleration in Rest of World constant currency growth, and a 600 bps acceleration in Other revenue due to GCP and G Suite.

I’m optimistic about the near future.  My advertiser contacts expect Q4 Search growth similar to Q3 and YouTube growth is likely to accelerate in Q4.  Further, two new initiatives will gain traction as the next year unfolds.  First is Maps ads, which is straight-forward.  Second is Expanded Text Ads (ETAs).  I think the market has grossly under-appreciated the significance of ETAs, and this is confirmed by discussions with advertisers.  ETAs were released from beta in late July and began gradually rolling out in Q3.  ETAs had a negligible effect in Q3.  ETAs will become mandatory post-January 31st.  ETAs have expanded the number of characters available in a text ad.  Below is an example of the old text ad format on the left, on the right is an ETA.

The mechanical effect of an ETA is to take real estate from organic search results for highly commercial searches, making it more convenient to click a text ad rather than scrolling through to organic results.  The subtler effect is that it makes text ads more informative and gives advertisers more freedom to write attractive copy, which is particularly useful for mobile users who want to conserve data usage.  

The confusion which has caused the market’s misperception is that early results of ETAs showed no measurable difference in click-through rates between ETAs and standard ads, even for sophisticated agencies.  I believe this is because ETAs require a greater effort at copy writing.  After allowing a few months for optimization, dozens of contacts have measured CTR increases of up to 400%.  It’s hard to pinpoint a median, but it’s clear to me that the direction is up.  I’m waiting enthusiastically to see what kind of fireworks show ETAs will put on over the next few quarters as advertisers master the new format, possibly only a roman candle, but I’m confident it won’t be nothing.

Another stocking stuffer waiting in Q4 is the launch of Pixel.  Indications are that about 3 million Pixels will be sold in Q4 at an ASP of about $700, meaning about $2 billion of handset revenue will be recognized in Q4 where no handset revenue existed in Q3.  The unknown is margins.  It’s believed that component margins are similar to iPhone, but it’s unknown how much Apple’s efficiencies are worth to gross margin.  And initially sales and marketing support will probably consume all the gross margin.

Google Cloud Platform

What everyone knows for certain about public cloud computing is that AWS will be the winner.  But this is inexorably wrong.  Google should be the natural winner.  Google handles 30% of all Internet traffic; they build servers that perform ahead of Moore’s Law; they’re the largest owner of fiber globally; they have the lowest cost data centers; they created the leading container engine (by a factor of five versus the #2); their security is superior to AWS; they have the industry standard speech recognition, image recognition and machine learning APIs.  GCP recently applied Deep Mind to lower their electricity costs by hundreds of millions of dollars a year.  GCP is 30-60% cheaper than AWS, with lower latency.  Where GCP falls behind AWS today is in the software stack, which admittedly is a very large part of the value.  But cloud software is Google’s business, they’ve written two billion lines of code, in contrast to Facebook’s 60 million and Windows’ 40 million.

I was wrong to think that GCP is immaterial today and has merely option value.  The reason I was wrong is because I observed, like everyone, that AWS is ubiquitous.  But GCP’s customers are users with violent storage and computing needs, like Airbus, auto OEMs, Spotify, Snapchat, Evernote.  

In a little-noted Google Horizon presentation on September 29th (link below), GCP SVP Diane Greene revealed the year-over-year revenue growth rates of GCP: storage 10x, compute 4x, containers 25x (and as an aside, G Suite 3x).  Whatever lead AWS has, GCP is bearing down in warp drive.  

https://www.youtube.com/watch?v=XHVAIIeFNcw

Growth % of Revenue Growth Contribution

Storage 900% 30% 270%

Compute 300% 65% 195%

Containers 2400% 5% 120%

Total 585%

Maybe no one watched this presentation, or the market didn’t process the gravity of the numbers, or no one cares.  If GCP revenue was $500 million in 2015 (sell side analysts estimate more), it’s perhaps at a $3.5 billion run-rate, versus AWS’ $11 billion.  If GCP is worth 10x revenue, it has a value of maybe $35 billion today, growing at 500%.  Today no value is ascribed to GCP because it’s unprofitable.  Even if GCP grows by 200% over the next year, that adds $70 billion of enterprise value to Alphabet in 2017.  In a few years, this won’t be your father’s Google.

These analysts think AWS accounts for 40-50% of Amazon’s enterprise value (16-20x TTM revenue, with revenue growing a passable 60%.  I report, you decide.

Valuation

Alphabet trades for 15x my estimate of 2017 core free cash flow, net of cash.  Today Search accounts for about 100% of earnings, and I expect Search to grow at 15%/year for the next few years.  

2016E Core EBIT $33.6 billion

Tax @ 20% 6.7

Net Income 26.9

Diluted Shares .698

EPS $38.50

2019 Core EPS $58.60 @ 15% CAGR

Multiple 20x

Core EV/Shr. $1,172

2019 Cash/Shr. $267

Core Equity Value/Shr $1,439

On top of that, I estimate the value of the high-growth, loss-making businesses in 2019:

Revenue Multiple EV

YouTube $23 billion 5x $115 billion

GCP $20 billion 10x $200 billion

Hardware $22 billion 1.5x $33 billion

Total $65 billion $350 billion

Diluted Shares .698

EV/Share $500

So the stock price could more than double in three years, to over $1,900/share, if the non-core businesses are valued fairly.  If Alphabet repurchases stock aggressively, then the stock price could triple.  But that's just like, my opinion, man.

Catalyst

Growth