Davud Kim’s Forage Capital letter to partners for the third quarter ended September 30, 2016. Also check out David’s www.scuttleblurb.com
“Longing on a large scale makes history.” – Don DeLillo, Underworld
[drizzle]From inception on July 14, 2016 to quarter-end, Forage Capital returned -0.6% (net of all expenses but gross of fees1) vs. +0.9% for the S&P 500 Total Return Index. Amerco (UHAL, -11.1%) and Oaktree Capital Group (OAK, -4.8%) were the two largest detractors from our returns, while MRC Global Inc. (MRC, +26.9%) and Alphabet Inc. (GOOG, +5.8%) were the two largest contributors.
We owned Alphabet for a few weeks before I sold it at a modest gain, which gain I ascribe entirely to luck. Sometimes I’ll buy a stock that seems compelling at first itch while looking for reasons to sell it. In Alphabet’s case, I found a few. I mentally organize Google’s Ozymandian sprawl by visualizing a layered ecosystem that looks, reductively, something like this (you’re looking down on a layered cake):
Platforms proffer apps/utility to users and simultaneously capture and convey user data to deep recurrent neural networks, bolstering machine learning algorithms for purposes of offering relevant ads and experiences to users, whose increasing engagement a) attracts more ad revenue and b) prompts third party developers to enhance and program more apps, drawing further user engagement, pollinating neural nets with still more training data, and so on. Most of Google’s internal development efforts and acquisitions enhance some layer of this flywheel.
Machine learning is so hot right now, but it’s actually been a central and purposeful part of Google’s narrative from the very beginning2 (the Company hired its first Director of Machine Learning in 2001), enabling the scale automation required to someday blanket interstices of daily life with an invisible and ubiquitous search utility that will function as the “third half of your brain,” as Sergey Brin once put it.
Charlie Munger once remarked: “Google has a huge new moat. In fact, I’ve probably never seen such a wide moat.”
Few would dispute Google’s dominant advantage in PC-centric search; but, a far more pertinent concern is search’s continuing relevance in the increasingly curated digital lives that alternative platforms have enabled. Smartphones now account for 65% of all digital media time after having stolen significant share from desktop over the last three years.3 Nearly 90% of that mobile time is spent in self-contained apps (which increased usage informs customization, making for still better experiences and more usage…), with Facebook and messaging apps accounting for nearly a third of all mobile consumption vs. just 7% for Google’s mobile Chrome browser and YouTube.4 Consider that, according to a recent survey commissioned by BloomReach, 55% of US consumers now start their online shopping on Amazon.com vs. 44% a year ago, while the percentage that begin their product searches on Google (and other search engines) has dropped from 34% to 28% over the same period.5 Amazon product searches are particularly pernicious for Google because, in terms of purchasing intent, they are about as bottom-funnel as you can get. App usage cannibalizes time that would otherwise be spent on Google search – we can impute that the average person conducts fewer searches per month on a smartphone than she does on a PC/laptop.6 Accelerated search revenue growth over the last three quarters has been fueled by loading mobile search results with more ads, a tactic that has natural limits and has now anniversaried. We can probably assume, based on management’s refusal to answer basic questions related to mobile query growth, that growth in search queries is materially below that of search revenue.
Facebook is increasingly challenging Google’s dominance at the bottom of the marketing funnel. On its last earnings call, Priceline (OTAs are major keyword spenders on Google) – whose performance advertising ROI (mostly keyword purchases) has compressed for several years – mentioned that, while it has traditionally used Facebook for brand advertising, it was looking to work with Facebook on more “performance-oriented” placements. Meanwhile, Expedia’s engineering team is collaborating with Facebook’s on dynamic ads,7 with spend “now getting to real significance.” [For what it’s worth, I ran a little uncontrolled experiment of my own, launching dual campaigns through AdWords and Facebook for my blog (www.scuttleblurb.com – SUBSCRIBE NOW!), and achieved far superior click-through rates on FB]. Furthermore, Google faces formidable challenges competing against Microsoft and AWS – who have nurtured capabilities and developer/user ecosystems up the stack – in enterprise cloud, without even considering the cultural incongruence of an anthropoid-phobic organization meeting the high-touch service expectations of an enterprise.
There’s also a broader, more ambitious point that merits some consideration (more than I will give it here). A review of US information industries – telephony, radio and television broadcast, film, and other media – since the turn of the 20th century reveals a Schumpeterian cycle in which innovative communications technology or anti-trust action dismantles a competition-stifling incumbent industry structure, catalyzing a period of chaotic industrial openness and competition that inevitably and eventually assumes the same expansive, monopolistic posture of its overthrown predecessor.
Despite its decentralized origins, the online ecosystem is traversing a similar path. While there’s nothing centrally planned about it, the self-reinforcing nature of demand-side scale economies almost demands ineluctable convergence towards winner-take-most systems.
And what began as rapid consolidation around relatively siloed chokepoints – e-commerce (Amazon), search (Google), social (Facebook), messaging (FB/WhatsApp) – is transitioning to territorial battles over ever-expansive platforms that blur previous competitive boundaries.9 10 That infantile scrawl of mine above could just as easily apply, with few edits, to Facebook, Apple, Amazon, Snapchat (really) and a host of others. Given the concentrated benefits of platform economics and the tendency for new platforms to cannibalize the ones that came before, it’s unlikely that everyone will capture equivalent value.
Advances in machine learning empower cross-silo competition. While early advantages in processing might and unrivaled access to data hoisted Google to the forefront, a sudden surge in ML enabling technology – parallel graphics processors used for gaming are now low-cost commodities (embedded with ever-increasing software functionality) coopted for AI applications; sensor-enabled mobile objects have dramatically proliferated – over the last 3 years, has narrowed Google’s lead in massive data capture,11 and enabled competitors to catch up in ways not possible just 5 years ago, complexly amplifying attendant industry structure permutations and value capture uncertainties therein.
Anyways, the point is that while Google’s incumbent advantages might position it well for future platform dominance, I am less confident in that outcome than I was before I put on the position, and I don’t believe that ~28x earnings (ex. stock comp)12 offers the right odds relative to other opportunities I am more capable of understanding.
I guess I should talk about a stock that we actually still own.
[Aside: The pithy way to express an investment idea in a letter like this is to highlight only the upside potential. I think many investment managers do this because they understand that readers prefer strength and consistency over diffidence and nuance. Plus, writing can be a pain, and it’s easier to frame an