Some Things I’ve Learned over the Years

First some housekeeping. I will be launching a Delaware-based fund/partnership early next year with more details to come over the following several months – so this blog will either turn into a private one exclusive to members within my own network or become a much more casual platform. Needless to say, I am fortunate to have built such a network of like-minded individuals who I highly respect and are much smarter and better investors than myself. I certainly look forward to continuing the dialogue!

An Obsession with Studying the Superlative Business:

Over the years my approach to sourcing ideas has evolved. By far my favorite approach to sourcing ideas now is to identify the very best companies in the world (the top 0.01%), whether public or private, study them in-depth, and subsequently draw the most fruitful lessons I can from them. I love this approach as I have discovered that studying the few companies I admire the most can teach me much more about business than your average run-of-a-mill company. My evolution in sourcing is also a function of compounding my knowledge base of different businesses over time, and slowly realizing that most ideas, no matter how optically inexpensive they may look, simply aren’t investable for a highly concentrated, high hurdle-rate oriented investing approach. In my mind if you’re truly a highly rational concentrated investor, your opportunity cost for owning any name should be absolutely massive. Another way to think about this is to invert the common “low P/E/multiple approach” to sourcing ideas: find the companies with the best chance of maximizing long-term earnings power first, then figure out the right price to pay to ensure the desired margin of safety. For the people who are still stuck in Ben Graham’s “valuation-first/cigar butt” world, my valueless opinion is that the faster you get to Munger’s approach, the better off you’ll be long-term as an investor. Hint: You’ll be much smarter.

What makes a superlative business? To answer this question in a paragraph would be a disgrace to this blog. But here are a few things to think about: they are managed by exceptional people who are obsessed about winning, and have no other major distractions in their life other than the company they manage. You can perhaps think of them as similar to Nietzsche’s Übermensch – a small subset of the human race – and in this particular case the business world – who are making serious progress in advancing humanity forward and are driven by their own independent set of values (not those of others a la the herd). These values are typically formed by unique personal experiences.

Culturally, some of these companies share common traits. They don’t shy away from taking risks and view failure as an option; strictly speaking, they play to win, not to lose. To say that they have an entrepreneurial culture is an understatement. Having such a risk-taking culture helps mitigate risk aversion bias which plagues most companies and people. Just from the realization of this bias alone, I have always approached decisions on assessing the risk of not taking any risks.

What else? Speaking from a numbers perspective, they have an exponential return on invested capital when measured over a multi-decade horizon. This is a function of the business’ competitive moat growing wider over time which can yield a non-linear long-term return on capital profile. If you are an exceptional business, you are by my definition a super durable business that can withstand the test of time and also enjoy an open-ended runway of growth. As for the classical economists who promote competition as the best form of value creation for all – they are all wrong. Like Peter Thiel said, competition is for losers and monopolists are for the studs. Since most businesses do not fit this profile, this is where being forced or at the very least having the intention to hold a business for 10 years or longer should greatly limit your range of investing opportunities.

Let’s think about all the companies that went out of business over the past century or so. How many of these companies that were in business 50 years ago actually stood the test of time? I think the ones that have survived had the leadership and a company culture able to withstand and adapt in rapidly changing competitive environments. Having an innovative culture, with the open-mindedness for business model evolution goes a long way towards staying in business. I don’t think at this point I even need to point out that these companies have a long-term orientation. Without these traits, no matter how dominant a business may be now, it lacks the qualitative characteristics that enable it to stand the test of time. You can see examples of these businesses all the time: once dominant companies that are now having their moats being eroded away by superior, more nimble organizations, and will soon become a shadow of their former selves. Think IBM, Oracle, Viacom[1], and the US department stores.

Take Oracle for example, the upside case expressed by some is that this is a rather cheap stock with an attractive near-term FCF yield. While their on-premise database and application businesses are in secular decline, this is a slow and measurable rate of decline, as switching costs are substantial and enterprise relationships are very sticky; and as the Oracle installed base continue to migrate their workloads to the cloud, Oracle should be able to retain the customer relationship despite cannibalizing itself. What I think the bulls fail to appreciate is that the economics of Oracle’s on-prem business (which is bound to shrink significantly) is much more attractive than what will likely become of their cloud business (which is bound to become the main value driver of the company). Oracle was extremely slow to transition their product to a cloud environment, where switching costs are much lower as customer uptimes are way shorter and having a proprietary/unique tech stack is not as important as offering a lower price tag to enterprises at scale. In fact, the most successful database cloud provider in the future will likely win by offering a much cheaper product based on open-source technologies. Hint: I’ve written up this company on this blog.

Leading disruptors that had no highly profitable legacy business attached such as AWS and Workday in infrastructure/platform cloud services and HRM applications, respectively, already know the endgame scenario in enterprise software, and are investing all their resources and efforts towards positioning themselves for it. What Workday is doing specifically is disrupting Oracle and SAP in HRM & Financial Management applications. In classic innovator’s dilemma fashion they attacked the SMB/smaller enterprise market first with a largely price-driven strategy and a product incrementally “good enough” for customers that outweighed performance-based concerns. So Oracle is fighting a multi-front war and even as it manages to grow their cloud business at a rapid clip (no one disputes this fact) to help offset on-premise attrition, the overall company margin profile of a lower tier cloud software business will be much, much lower than their licensing/maintenance software model – that I can almost guarantee you. It’s not a coincidence that there’s a similar industry narrative for the Pay-TV/the cable networks and department store industries: despite transitioning their business models they all face the same threat of being largely dis-intermediated and thus becoming much less profitable. In fact, we’ve actually seen a similar movie before in the classified section of local newspapers as content and ad dollars largely shifted to digital/online channels. Ask a millennial if they check the local newspaper/classified section when they are looking for a local gig and you’ll know what I’m talking about.

Another common theme among these long-term losers is that they typically opt to use share buybacks/financial engineering as a pitch to shareholders in a desperate attempt to drive outsized short-term EPS growth. This is all but a mirage in value creation – especially for those that elect to lever up to buyback overvalued stock like what some of the cable networks have done over the past several years. Why are they doing this? Hint: it helps to read proxies; sometimes, you’ll figure out that either the CEO/controlling shareholder is a super wealthy empire-builder or will likely retire in a few years. Financial engineering can only take you so far, and scaling value creation successfully via M&A and buybacks over a multi-decade period with a secular declining core business is extremely rare in my view. What’s typically more common is self-dealing and/or integration train wrecks. As a result, it would be an absolute nightmare to hold a share of a business like this for the next 10 years. OK, you may have guessed now that I despise IBM and Oracle and don’t think too highly of their leadership. For good examples of companies that have been successfully navigating these software industry changes, I would look to Microsoft and Adobe Systems.

One final thought on enterprise software is that the technology stack along with how it is sold is evolving over time. The reality is that no new software engineer today looking to build the future is getting trained in .Net, C# or the Java tech stack. With the rise of AWS, a public cloud platform that sells software infrastructure basically as a utility, we are now transitioning to an era where software pricing is based on volume/usage instead of subscription. As a brief history, Software 1.0 was the era of on-premise solutions based on the licensing/maintenance contract model, Software 2.0 was of course the rise of the Software-as-a-Service (SaaS) model with Salesforce as the industry pioneer, and Software 3.0 as mentioned is evolving into software-as-a- platform. AWS is undoubtedly the industry pioneer in this model but there are some smaller, emerging cloud platforms such as Twilio, Atlassian, Shopify and New Relic. Once a platform becomes nearly ubiquitous amongst third-party application developers, an application (such as Uber, for example) can outsource certain platform-as-a-service functions to a platform such as Twilio (in this case communication-based functions such as texting/SMS and voice), and be billed on a volume basis by Twilio. Utility-based business models that become ubiquitous are extremely attractive, and AWS is at the center of Software 3.0.

Something else I’d like to talk about is investor herding. As you may have noticed I’ve written up some pretty well-known “hedge fund hotels” on this blog. Some investors appear to automatically assume that a hedge fund hotel must be a crowded trade. As a fundamental investor I don’t think one can draw any conclusions on what a business is worth based solely on knowing the shareholder base. I think my friend at one of the well-known Tiger cubs said it best: In general hedge funds (at least the top-tier, fundamentally-oriented, research-intensive ones) are better stock pickers than the average market participant. Theoretically they have the better resources, talent and a sensible incentive structure for the investment team to generate sustainable alpha. However there may be periods such as recently where a redemption cycle in full swing can cause outsized short-term volatility in shares with a cap table heavily skewed towards hedge funds. This doesn’t mean the intrinsic value is impaired as the market price reflects the consequence of forced selling.

This broader issue appears more like a psychological issue to some and is certainly not a rational stock picking process. Guys that actively avoid hedge fund hotels might have what I like to call the contrarian syndrome where they are biased to shun from ideas that are widely-held by savvy investors. The opposing psychological bias of herding is certainly a notable risk as well.

The beauty of the market is that there is always a winner and loser in every transaction and almost every participant has a different reason for transacting. I’m sure there is a subset of “value investors” who look at Amazon stock and think it’s grossly overvalued trading at >160x PE. Then there is a subset of “momentum” oriented investors that must own the stock and chase performance at any price. Bottom line is market consensus is objectively captured by the price offered and whether there is a large margin of safety is dependent on one’s estimate of intrinsic value. To their detriment, people seem to over-complicate things due to certain innate tendencies.

JD and Amazon: Scale Economics and the Customer Experience

I’ve talked about both JD and Amazon quite extensively on this blog, but I can’t resist talking enough about them.

Some define these businesses as internet companies, whereas others may label them as online-retailers or e-commerce companies. Honestly, I think these are customer experience companies. It’s the same way Tony Hsieh defined Zappos (the online shoe retailer that Amazon acquired). Zappos’ company culture was highly customer-centric. Zappos call centre representatives would be more than happy to spend hours speaking with lonely bored customers who just wanted to talk to someone. To reinforce the company’s culture and values around exemplary customer service, every new Zappos hire would be required to work in the 24-hour call center for a minimum period of time. JD for example makes sure that each call center rep is actually physically smiling when they say hello to a customer on the phone.

Why are the best consumer-based internet companies, and online retailers in particular, obsessed about the customer experience? The Stratechery blog, in my view, is the best blog online focussed on TMT industry analysis and strategy that talked about an interesting theory called Aggregation Theory. My own take on this theory is that since the marginal cost of distribution or the user acquisition cost approaches close to zero over time for many internet platforms, only the companies that are able to provide the best user experience at scale are able to capture these exceptional unit economics, and in turn commoditize suppliers (product exclusivity is a non-factor) and thus become the long-term industry winner. This theory aligns with the conventional wisdom in e-commerce (all the industry leaders agree), that the best way to maximize value in mass-market retailing is to provide the best customer experience.

Some value investors avoid investing in technology-based shares due to their admitted inability to predict long-term changes in such business models. If we invert this view and focus on what’s not likely to change, we can build conviction around certain long-term drivers. For example, there is almost no question that 10 years from now, customers would want a wider selection of products from Amazon, a better customer experience, and even more competitive prices. Within its major markets, no other company on Earth will be able to add more value to the customer long-term around price, selection and the customer experience than Amazon. Whether a T-shirt is ordered on a virtual reality platform and then delivered to the customer’s door by a robot or drone, or whether an apple is purchased at a no checkout Amazon Go store 10 years from now are frankly quite irrelevant as long as the business has experimented to figure out what provides the best customer experience at scale. In other words we don’t know what the precise economics of the business will look like 10 years from now but what we do know is that the moat is primed to grow wider over time and the business will be much more valuable than it is today. And that, to me is enough reason to invest at the right price. Think of this as a channel-agnostic or you can even say business model agnostic approach to long-term shareholder value creation, and more of an enduring business recipe for success. The companies that understand this are bound to win.

I think the only e-commerce competitors that will survive the ever-expanding Amazon retail onslaught and also perform well will be the very specialized, pure-play online retailers that are already the market leaders within a very particular niche or specific product category (they’ve already built substantial scale and market share in such category). As an example, AO World is an extremely customer-centric, UK-based retailer that specializes in selling 3C and home appliance products online. AO is a structural share gainer in the UK-based market for large home appliances and electricals. The barriers to entry in selling large home appliances are higher than the average retail product due to a specialized logistics and fulfillment requirements for this category. The logistics infrastructure required for bulky “white appliances” such as refrigerators, washers, and dryers requires 2-men delivery crews who are specially trained to install such products at the customer’s home and the company has to provide high-quality post-sales service. This is simply not a piece of the value chain you can outsource to third-party delivery firms and expect high customer satisfaction. In fact I am not aware of any reputable third-party logistics firms that provide large home appliance delivery services. On top of being able to operate the physical element of fulfilling orders more efficiently than your largest direct offline and online competitors, the widest selection of brands of SKUs within the specific category at competitive prices are required for completing the flywheel (achieving sustainably attractive ROICs long-term). In terms of the customer experience selling appliances online over brick-and-mortars yields another considerable advantage as legitimate product reviews/online video reviews are easily assessable.

Some other European-based pure-play e-commerce companies that should do well despite Amazon include: Zooplus which is an extremely well-managed online retailer specialized in selling pet supplies, Yoox-net-a-Porter sells luxury apparel online, and ASOS is an online apparel retailer (with little customer overlap with Amazon) that sells fast-fashion apparel largely to the millennial demographic. What these e-commerce companies have in common are a product selection unrivaled within their specific category which no other retailer can match – including Amazon. This is in part due to the massive SKU count required and the more “long-tail” nature of such categories – especially the apparel category which I will discuss in more detail below – which is not easy for a subscale competitor to replicate. Despite this, there is no question that the business of online retailing will always have the “Amazon overhang”, and likely some competitive pricing and margin pressure will exist no matter the category, but I’m a believer that there is space for multiple category winners and Amazon won’t dominate every single vertical. There’s a big difference between business durability and profitability… it appears that if you have the right time horizon, prioritizing the former over the latter could be a very wise decision, but most investors and CEOs don’t have the right time horizon.

JD is the only e-commerce company in the world that I know of that’s successfully building a multi-category last-mile delivery network that includes both bulky home appliances and fresh groceries. In fact from my knowledge only JD, AO World, Suning and Gome sell large appliances at scale with a specialized self-owned last-mile delivery network; Ocado specializes in selling groceries online but I am unsure whether they operate their own last-mile network. These two categories (the aforementioned large appliance category) – and the cold-chain logistics/fulfillment network required for fresh produce delivery are extremely difficult to replicate as it is not simply a question of having the capital required to build such a network at scale but also how efficient the network is over time. For example for Alibaba to fully match JD’s cost advantage on a per order basis it would require them to in-source all the product inventory themselves and build out a high-quality, modern warehouse network requiring close to 50 – 100mm sqm of GFA if we take the average SKU or order count per GFA for JD and apply it to Tmall’s order volume (this is a rough calculation but given we know that Tmall’s total order volume is a multiple of JD’s – the gross floor space that will be required is safe to say massive). We know this scenario is basically near impossible for several obvious reasons. I will not spill it out in detail here but very briefly Ali has a very software-centric culture, there are already existing 3PL firms that Ali’s merchants rely on which are improving their own efficiency over time, local government zoning laws near major cities are a pain in the ass and China likely doesn’t have 50 – 100mm sqm of GFA of strategically located real estate for modern warehouse buildouts (in fact it might not even have 20 – 30mm and of the total modern warehousing stock available most of it is already taken). Let’s not forget that JD is relentlessly improving their fulfillment network every year and will likely get to 2 – 3 hour delivery windows for most of its network within the next 5 – 7 years due to its huge lead. Taobao sellers often evade the VAT tax and a significant number of them experience chronic losses in the chaotic Taobao marketplace. The fragmented nature of Taobao/Tmall means those small sellers can never benefit from the same type of scale advantages that JD’s 1P operation could. Instead they have to rely on BABA to improve its platform/logistics each year to catch up with JD but even those improvements aren’t going to structurally alter the cost structure of the sellers. In short if you do not own a hub-and-spoke fulfillment model yourself, replicating a direct-seller (1P)’s cost advantage such as JD’s is nearly insurmountable.[2] Data and software can only take you so far without getting your hands dirty with the physical. The lesson here is if you want to secure the best customer experience, you have to in-source the entire value chain. An interesting note: Bezos has apparently said that one of his biggest regrets when starting Amazon was that FedEx and UPS already had the network in place for highly reliable and express parcel delivery services.


Thinking Deeper about Online Direct-Sales (1P) vs Marketplace (3P) Models

Two of the largest retail categories (globally and in China) are fast-moving consumer goods “FMCG” and apparel; these categories are growth opportunities for JD as the former is under-penetrated and the later JD is under-indexed to. The former I believe JD is likely to win given the nature of the category and JD being able to leverage its 1P strengths: low return rate, time-sensitive, high SKU count but a bit more standardized on average than apparel, high inventory turnover, high order frequency, specialized logistics required (for fresh groceries) and some health-sensitive products. Recent professional third-party independent surveys support a roughly 50/50 incremental share split between JD and Ali supermarket. I believe whoever has the superior logistics network wins in this category long-term as functionality and price appears to be the major customer purchase considerations. More standardized items and low-SKU count categories (more leverage with suppliers WRT volume rebates) should also attract a more diversified traffic which should help fuel the 3P business grow its selection of merchants and products. Unlike electronics and home appliances, greater expansion into FMCG will capture a broader customer base which JD can use to cross-sell a greater selection of higher-margin apparel, cosmetics, household goods etc in a virtuous flywheel.

Apparel is totally different animal. Many apparel products are driven by fashion where emotion is a larger element in the purchasing decision; this is thus a slightly different type of user problem for online retailers to solve. It is no surprise that the marketplace model works more effectively for apparel merchants vs. selling wholesale as long-tail traffic and demand is dispersed across many different SKUs (with tens to hundreds of millions of SKUs potentially in the apparel category). Tmall’s marketplace as opposed to JD Mall is simply designed/tailored for a better/more user-friendly experience for browsing hundreds of different products/SKUs within the same category such as apparels (takes an hour to find the perfect pair of shoes?) and as a result is a fantastic property for monetizing purchasing intent via ads as engagement is high[3]. We are already seeing Ali’s management switch the narrative from GMV growth to rev growth which is driven by ad revenue streams. This is in contrast to JD’s and Amazon’s site/app which you could label more as “functional utilities” as their user interfaces enable speedy customer checkouts.

JD has had only limited success in selling apparel inventory directly aside from success in more standardized apparel products. This is not to say that they can’t expand their apparel marketplace GMV longer-term if they work on positioning their fashion merchants/sites effectively (such as investing in developing separate, customized sites for individual merchants a la Tmall so that certain apparel merchants have less brand dilution if they list on JD, improving the overall tech stack for their 3P merchants via tools and features, and increasing customer traffic to the overall JD platform which will make it hard for apparel merchants to ignore longer-term. There is also evidence from professional third-party surveys that suggest that JD’s lower overall traffic conversion rate compared to Ali is due to their more limited selection. I don’t see why this can’t change as their product selection expands. Amazon Fashion for example started off targeting the low-end of apparel brands, and continues to work its way up to the higher-end.

So FMCG category is super strategic to building a one-stop shop online shopping platform -> JD must win, Apparel -> not a deal breaker to the long-term thesis but provides additional upside optionality if category expansion is successful.


Revisiting the JD Thesis

“When the facts change, I change my mind, what do you do, Sir?”

– John Maynard Keynes

I’m starting to get lazy writing this post 😦 so let’s use point-form notes for this section:

-Remember Bezo’s quote about there being two types of companies: the ones that always try to charge more and the ones that lower prices over time. The latter are applicable to many high-turnover, structurally low-margin retailers who maximize NPV by reinvesting efficiencies into lower price given the rather price elastic demand profile of their category. Examples include many supermarkets, ultra-deep discount retailers such as Costco, Aldi, Lidl, and major categories on Amazon and JD. People often focus on the first-order effect of lower prices = lower margin, and higher volume, but forget to take into account higher volume rebates from suppliers/better mark-ups over time and higher cost efficiencies/inventory turnover = some margin recapture longer-term. As a result such pricing decisions adhere to the goal of maximizing the long-term absolute dollar free cash flow per share, not long-term margins.

On operating leverage:

-Marketing and advertising expense leverage (which is not often talked about) kicks in over time as less traffic will be sourced from third-party traffic and more directly from organic sources such as new customers from either the WeChat/QQ Mobile channel or JD’s app/website. 5 – 7 years from now this will largely be a purely discretionary expense for a dominant franchise such as JD.

-Additional upside from leveraging logistics infrastructure such as unmanned warehouses, self-driving vehicles, and drones to enhance their long-term structural cost advantage. The automated warehouses could decrease their labour costs greatly as this is the biggest component of overall fulfillment costs. I have already noticed that their Asia #1 warehouse in Shanghai has increased its order processing efficiency over time from originally 16k parcels per hour to 20k now.

-If there is a hard-landing in China, offline retailers (mall-based or otherwise) will get absolutely crushed. Suning and Gome, in particular, are much less efficient retailers than JD on every key retailing metric such as inventory turnover, and OpEx as a % of direct-sales. In such a scenario JD should benefit even more as the secular shift to online sales accelerates.

-The overall margin expansion story may be delayed by a more few years (due to aggressive price investments into the FMCG category) than what I originally thought before they reach a steady-state level, and their steady-state margin profile may be lower than what I originally assumed (4 – 6% of Net GMV). Mid-case, at a 3.5% operating profit margin as a % of Net GMV by 2021, with the earnings power capitalized at a very conservative 18x – 22x exit EBIT multiple (or less than 25 – 30x FCF with an implied terminal FCF growth of ~6% assuming a 10% cost of capital), and adding up all the other hidden assets and FCF generated over the interim years, JD shares could potentially be worth between $137 – 166 by 2021YE. This valuation assumes zero upside optionality from all the FCF that will be generated in the interim. I think these exit multiples could potentially be very, very conservative as there’s nearly a non-existent risk of long-term disruption (commercialized teleportation is unlikely) for the dominant retailer whose main competitive advantage derives from scale. By this time they should also at the very least be larger than the current size of Amazon’s entire retail business. As additional downside protection in the absolute worst case scenario, it appears Walmart and Tencent have taken a keen interest in JD and have been buying more stock recently around these levels. In the worst case scenario, could they be a liquidity provider of last resort, or even acquire the entire company? It still gives me a bit of comfort that there is some sort of valuation “floor” here…

Amazon: Still the Best Business in the World

To reiterate the original thesis: Amazon is the best company in the world managed by one of the greatest investors in the world and is one of the most compelling multi-decade compounding stories today.

Who would have guessed 10 years ago that AWS was going to dominate public infrastructure computing and be worth several hundred billion dollars or multiples of Amazon market cap at the time? Not even Bezos. Even without AWS, an investment in Amazon shares would have yielded a fantastic 10-year return. The lesson here is that upside optionality is almost always underpriced for the exceptional business. As for the investors that always have a bias for overvaluing predictable cash flow streams[4] and over-rate mean reversion-centric theses, the people with this mindset will almost always mis-understand Amazon. Exceptional businesses are exceptional businesses for a reason – their returns on invested capital do not mean-revert to the average company over time.

So if we model nearly $500bn of net GMV by 2021 or a 5-year high-teens CAGR, an operating profit margin of mid-single digits on Net GMV (50% contribution profit margins for 3P sales, 5% contribution profit margins for 1P sales), and assuming the Prime business remains largely a cost-center or is around break-even levels on a rough contribution profit basis, there is a viable path to $30bn of retail operating profits by 2021.

Under this scenario and assigning a 20x multiple on 2021E EBIT, the retail business alone could be worth ~$570bn by 2021 on a future earnings power basis, or slightly above $1,000 per share. I remain a firm believer in the multi-year margin ramp-up of the retail business given all the potential leverage below the gross margin line that should kick in from increasing asset utilization rates from the fulfillment centers, increased warehouse and sortation center automation (which would reduce labor as major input cost), and potential mainstream breakthroughs in driverless vehicles and drones. With 3P GMV potentially consisting up to two-thirds of total company GMV 5 years from now, the higher-margin contribution profit mix from this segment, along with higher product margins for the 1P business from volume scale, category expansion (as Amazon is currently under-indexed to higher-margin categories), should drive overall consolidated retail margins higher substantially. Another perspective on long-term margins: as customer acquisition costs approach 0 long-term for the online retailer, marketing expenses alone which is currently around 5% of total retail revenues, will be purely a discretionary growth expense.

On the terminal value considerations, Amazon’s market share of total US retails sales excl. gasoline is only ~3.4%; their online share is closer to 35% of all US e-commerce sales. I’m honestly not sure if these figures are on a GMV or net sales basis, but that’s inconsequential and is a rounding error in a multi-decade DCF model. Obviously, not all transactions globally will shift online but we don’t even have to assume 50% end-state penetration in the US to create a massive TAM here. This is true today and will remain true even 5 – 10 years from now. With that said, assuming we discount our capital at 8%, I think an implied 5% perpetual growth rate on 2021 projected cash flows will look very conservative 10 – 20 years from now.

On AWS, the public cloud computing infrastructure and platform-as-a-service TAM is at minimum $500bn globally and likely closer to over $1 trillion if we include SaaS applications. Assuming base-case scenario 45% EBIT margins (incremental EBIT margins are already well above 40% as per the latest Q) on a 40% revenue CAGR to 2021, AWS could be earning around $30bn of EBIT and capitalized at 25x could be worth $750bn or over $1,500 per share 5 years from now. As for Google Cloud I think they have some room to grow and they are the leader in cloud-based AI features but aside from that I think their company culture sucks and their moonshot track record has been horrific. Hello drones, Hello Google Fiber. Despite Google cloud trying to differentiate mainly by price on core compute and storage cloud services, we are already seeing the leaders AWS and Azure operate in a highly oligopolistic manner with pricing cuts passed on to customers shallower than the implied Moore’s Law cost savings. Quite frankly, it is pretty embarrassing for a company like Google to have ceded to Amazon such a massive lead in public infrastructure computing when they already had the largest server infrastructure available globally. I think it just goes to show much big of a difference company culture can be to creating shareholder value.

Finally, we have the $140bn of FCF (~37% of Amazon’s market cap today) that could potentially be generated in the interim 5 years. To assume all this cash will be reinvested at a 0% rate of return is extremely conservative in light of management’s approach to allocating capital (which is to continually invest in high IRR projects on a 7 – 10 year horizon to the dismay of the vast majority of public market investors) and the widening moat which continues to offer attractive incremental reinvestment opportunities over time.[5] Jeff Bezos is by far one of the best and most under-rated (in my mind) investors in the world; allocating capital to someone smarter and a better investor than myself with such an exceptional track record of value creation is almost never an unwise decision I think. I believe my numbers have not priced in a more bullish scenario that accounts for additional upside in promising opportunities such as India e-commerce, higher monetization of FBA, Prime video streaming/Amazon studios, potential last-mile logistics for C2C, B2B, B2C commerce, Freight forwarding/logistics, at-home personal/virtual assistants, E-Sports via Twitch, O2O services and countless other adjacent markets that have yet to be tackled and that no one has yet to appreciate.


The Personal Growth of an Entrepreneur

I’ve been hearing from pretty reliable sources that over a recent period Richard from JD hasn’t been working as hard as before, he’s been pleasure tripping around the world over the past year or so with his wife, and he recently purchased a private jet and a few large homes.

Despite how reliable my sources may be, I’m not sure to what extent they are true, or how much it actually matters. It is also important to discern between fact and rumour. There were rumours that the company was going bankrupt right before it IPO’d in 2014. There were also rumours that Richard was shacking his former personal assistant while still dating his current wife. The lesson is we have to be very careful when sourcing news from potentially fake or unreliable sources.

I think to some extent, someone’s personal life can be “walled off” from their business life, and thus should be off-limits for judging a person’s character. I’m sure the range of opinions on this topic is very wide, so whatever. I personally have a more “laid-back” view than others. Common sense is if someone is snorting cocaine every day, they are not fit to be CEO. They don’t have to be a saint or the perfect man who’s going to marry daddy’s little girl. As for the private jet and mansions, let’s be real here: most of us are not going to stay in the same home like Buffett did if we became billionaires. Richard came a long way from being dirt poor and surviving largely on a potato diet. At this stage money shouldn’t really be the primary motivator for him since he’s worth over $7bn. With that said, he’s recently moved his bed into his office and is essentially living at the company now.

My own hypothesis on his recent behaviour (and I am making no excuse for it) is that his management style might have changed after he got married late last year. I think he might have went through a phase – call it a mid-life crises. This is a man in his early 40’s that didn’t get married until last year because he dedicated the vast majority of his adulthood towards building JD. As far as I’m concerned, he deserved a long honeymoon.

After getting married, and seeing JD become a much more successful company post-IPO and the Tencent Partnership, perhaps he just wanted to put in less facetime in the office, and at the same time tried implementing a more decentralized management structure. Steve Jobs for example who is another legendary tech CEO was a giant asshole before falling in love with his wife. He would often yell at his employees for minor mistakes. Richard was a bit similar during JD’s early days; according to many JD employees, he was not an easy boss to work for. What’s interesting is that he’s been physically absent from the company for long stretches of time in the past. For example, he studied overseas for over a year prior to JD going public, and the company was still firing on all cylinders during this time.

It’s obvious he didn’t like how the marketplace was being managed by Haoyu Shen. Now that he’s axed him, he’s taking greater direct control of the business again. So I do believe leaders can change over time, and their motivations can certainly change – this is all part of growing and learning as a person and entrepreneur.[6]

In terms of the company’s performance over the past year or so, the transitory issues surrounding JD’s marketplace segment was a bit concerning, so some criticism is warranted. However, on the most important operating metrics/key drivers such as net GMV, 1P and consolidated gross margins, operating cash flow, the incrementals, and the cash conversion cycle, business is actually looking quite good. I think the investment set-up looks quite favorable going into 2017 as margins begin to inflect materially and as we lap the transitory GMV issues.

One final point is that even if you believe there is a risk of an empire-building mindset, keep in mind that SBC is a major part of employee comp, similar to Amazon. Richard may be a billionaire but he absolutely needs his employees to work hard to fulfill his long-term vision, and if the JD stock price does not go up over time, employee productivity will likely be affected. This in my mind is basically an insurance policy for Amazon and JD shareholders or any investor invested in a similar company with a super wealthy CEO and like-minded employee comp structure.

Big picture: Richard has accomplished what few have (no unambitious or unmotivated person builds one of the best e-commerce companies on the planet, period), he’s still a very young CEO at 42 and has the potential to remain an effective CEO for the next few decades as long as his motivation remains intact. To the interests of long-term JD shareholders, a young benevolent dictator is a thing of beauty.

As much as you can admire someone for his past accomplishments and for what they say, actions always speak louder than words. So if I find out he’s partying every day I would dump the entire JD position immediately.


Inductive Reasoning

I take a management-first approach when it comes to investing in emerging market companies so I spend quite a bit of time conducting due diligence on the people behind the company. On a deeper note around the topic of Richard’s integrity (central to the JD thesis), I’ve summarized my thoughts below.

There are many frauds in China and one obviously has to be weary of investing in stock scams and promotions. My own experience of studying frauds is that the vast majority of them are typically run by highly promotional management teams that rely on deception in order to enrich themselves and sell stock at a highly inflated price. There is no solid evidence that I’ve come across to hint that this is anywhere remotely close to the case of JD.

My main points are as follows:

-As mentioned in my JD paper, Richard hasn’t sold a single share post the IPO event, he’s paid an annual salary of Rmb1 per annum, and has a 10-year option vesting plan. I also mentioned about the accounting differences between JD and Alibaba regarding purposely defining GMV in a much more conservative manner than Alibaba. As per footnotes in the filings, if defined on an apples-to-apples basis under Alibaba’s definition, JD’s reported GMV #’s would be at least 45% higher (~$200bn in gross GMV for 2017E, or ~80% of Tmall’s projected 2017 GMV size). We know Alibaba’s accounting is sketchy and Tmall’s disclosed GMV is overstated – but by 2018-2019 JD’s GMV could be higher than Tmall’s if measured on a like-for-like basis. Also worth noting is that JD’s largest direct retail competitors such as Suning, GOME, Amazon China, and Dang Dang are struggling and have no incentive to report bad numbers. I find it highly unlikely that Richard would have this type of compensation scheme and have the co’s numbers reported so much more conservatively than their largest peer if it was a fraud.

-JD use to deploy a very unique recruiting process which is very telling. They only hire fresh graduates with no previous work experience post-graduation for their management trainee program. From my knowledge, the focus of their recruiting process is not centered around in-person interviews because they are looking to avoid hiring people that are very good at selling themselves in person, which would potentially disadvantage other applicants that may be a better fit for their company but who are not as good at interviewing. Instead, there is a “final written exam”, where the topic is unknown beforehand. Before all the applicants sit down for this exam, a person from JD approaches the applicants, and essentially offers a bribe where they will reveal the topic of the exam for a cash compensation. This is actually the real test and there is no written exam. If the applicant takes the bribe, they failed it. I am not aware of any other company that tests potential applicant’s integrity this way.

-The recently released “corruption memo” revealing the names of the 10 employees who were arrested by police – this was not a “leaked” memo as reported by Western press sources which probably screwed up on the translation. It was released publicly on purpose likely as a warning to future employees on the consequences of acting illegally such as taking bribes from suppliers. Richard’s early life experiences of having employees embezzle money out of his first restaurant business that went bankrupt has driven him to a zero tolerance policy towards any corrupt behaviour at JD. Delivery personnel are not allowed to receive gifts from customers such as water, for example.

-JD’s largest shareholders were was a sign to me, but not a given conclusion. It is really not so much that these investors are among the largest shareholders of JD that gives me any comfort, but how Richard has dealt with them in the past which is telling. The Tiger Global deal was a particularly telling one; the company was in desperate need of cash and weeks away from bankruptcy, but Richard still did not renege on his verbal deal with Tiger (which was less attractive than the new competing offer). In terms of investing alongside some of the best investors in the world, I think a critical lesson can be drawn from the Valeant episode. I suspect due to massive commitment bias they were all unable to clearly see the red flags emerging in Mike Pearson’s behaviour, the flaws with the VRX business model, and how aggressively he ran the company.

-Sidney the CFO is the one who typically meets with groups of large investors. Sidney is quite non-promotional and I personally find management’s long-term expectations for the business (20% consolidated gross margins, low double digit 1P margins) very reasonable. As for Sidney’s 8 – 9 months at Longtop Financial – the massive accounting fraud – after speaking with a few people familiar with “ramp-up periods” for new CFOs in major companies, there is typically at minimum a 1 – 2 year transition period for a new CFO to get very familiar with a company’s books and controls. Since Sidney was at Longtop for less than a year, it’s possible he figured out pretty quickly that the company was sketched and left shortly after finding a new CFO role. Perhaps more importantly, his name was not brought up in the subsequent investor lawsuits.

-The company has the best fulfillment service in the world, period. These are some incredible achievements that I think required real human ingenuity, talent and hard work. Generally speaking, I don’t think people with this kind of vision/ability would commit fraud – it wouldn’t be worth the risk of losing it all. Also, the founder-CEOs that committed massive public company frauds were all super promotional (think Mike Pearson, Bernie Ebbers and Eike Batista) yet JD hardly has an IR effort! This is not a business model centered on paying for friendship/companionship such as a MLM scheme, or price-gauging on life saving drugs, this is about passing on as value to the end customer as possible via supply-chain efficiencies.

-Third-party data supports JD’s numbers. The best data is from the State Post Bureau which is the state regulator that monitors all express activities in China. They publish stats on total express delivery revenue which historically has tracked BABA/JD’s combined GMV almost perfectly (historical correlation > 99%). From this data point, we can deduce that as long as BABA’s GMV has been roughly right, then JD’s GMV would have to be at least roughly right. For JD’s numbers to be fake, both JD and BABA would have to be fudging numbers in concert which is highly unlikely given their animosity toward each other. In addition, Tencent has perfect visibility into JD’s WeChat-related sales data. I think they would have sold their stake a long time ago if there was something fishy here. Instead, they have been recent buyers of stock at these levels.

-Last but not least, Richard’s personal motivation portrayed through his company speeches and Chinese interviews where he says he wants to sell only high-quality authentic products to every Chinese citizen and take good care of all of JD’s employees by allowing them to retire in dignity if they work hard. I find it pretty unlikely that a fraudster would publicly announce these types of admirable goals, essentially live in his office now, and not sell any of his paper stock tomorrow worth around $7bn in aggregate and just quit. JD is one of the most customer-centric companies I’ve studied – similar to Amazon – it would be highly unusual for a fraudulent company to pass on so much value to the customer. If I had to create a fraud, I would first try and make sure that it appears very, very profitable… is this not common sense? And honestly, if I wanted to create a get-rich quick scheme, I definitely wouldn’t compete against Alibaba.

On Investing in China

Is China an investable market? I think studying the history and development of China in the 20th century and analyzing long-term trends provides us with some insights here. We know post the Mao era/Cultural Revolution – roughly from the start of Deng’s leadership – that China has been reforming its economy. Deng normalized diplomatic relations with the US, instituted major market reforms including the development of a private sector and market economy, opened the country up to global trade and foreign investment; this subsequently lifted the majority of the population out of poverty over the past several decades. Deng was a very practical leader, and ridiculed the notion that it was better to be poor under socialism than rich under capitalism – a common mindset among China’s leadership during the Cultural Revolution when untold numbers of people died from famine. This was in stark contrast with the Mao school of thought which valued egalitarianism above all else (a value aligned closer to Stalinist and Marxist policies). We know Mao’s intentions were good but his policies were a great failure.

Let’s face it, democracy has its issues as demonstrated by the recent US election, and is not the best fit for every country in the world. There is no one more credible or smarter about the issues surrounding Asian economic development and politics than the late Lee Kuan Yew (LKY), or the father of Singapore. LKY valued a clean government free from corruption, with high standards in ethics, integrity, competence, and intelligence; he wanted a society driven by meritocracy, a highly educated workforce, and of course respect for the rule of law. For example, he made sure that the compensation of senior government officials were closely tied to the compensation packages of private sector CEOs in Singapore, which helped reduce corruption. On the social stability front, there were no poor ghettos in Singapore despite there being visible Indian and Malaysian minorities who lived harmoniously alongside ethnic Chinese Singaporeans. In each individual neighborhood in Singapore, Singaporeans of different ethnicities were forced to live alongside each other proportional to the national ethnic population mix. This ensured that there would be no minority ghettos like we see in many countries in Europe, where resentment can easily breed into violence as seen today.

LKY believed in social Darwinism, but he was primarily a pragmatist with a largely non-ideological approach towards governance. Singapore was essentially a real-life experiment of the Hobbesian model (a social contract between the people and the State, where some liberties and freedoms are given up in exchange for security and social stability) which became extremely successful. During LKY’s era as the unequivocal benevolent dictator of Singapore until his retirement, Singapore’s GDP per capita grew from a third-world country level into now one of the highest levels on the planet.

The miracle of Singapore did not go unnoticed by the Chinese. Circling back to China: it is now a middle-income country, on its way towards becoming a high-income country over the next several decades. Some can say China is actually the most capitalistic country on earth that still labels itself as a communist one.

Ever since Xi took office, consolidating power and eliminating government corruption has been his main priorities. Some factions of the military were loose cannons under Hu Jintao’s administration, and were dangerously anti-American; Xi basically threw all of these crazy generals in jail. It is also important to note that Xi’s de facto second in command, Wang Qishan, (who’s in charge of the corruption crackdown) was a student of history and the Singapore model, was greatly praised by LKY as one of the most capable people of China, and has no offspring who would benefit from him enriching himself.

Here are LKY’s own words commenting on Xi:

“I would put him in Nelson Mandela’s class of persons. A person with enormous emotional stability who does not allow his personal misfortunes or sufferings to affect his judgment. In other words, he is impressive.”

 – Lee Kuan Yew

Xi is smart and what he understands is that conflict with the US will derail China’s long-term growth. As a result, cooperation is the only viable path the country can take towards its goal of becoming a 21st century superpower. As for the VIEs issue specifically, an action by the Chinese government that effectively wipes out foreign investors is an action that goes against all of what the Chinese government has been trying to achieve over the past 4 – 5 decades: long-term economic development and global cooperation. What the political figures who have the most credibility on the US-China issue (Deng, Xi, LKY, and Henry Kissinger – arguably the greatest statesman in the 20th century and the person who laid the groundwork for the normalization of diplomatic ties between the US and China in the 70’s) all have in common is they simply value pragmatism in their statecraft over ideals.

As for the Taiwan issue, the island can be wiped off the face of the earth by the PRC overnight, so I don’t think it’s wise for Tsai Ing-wen to fuck around here. We also know Trump is an idiot and needs a bit of time to be lectured/advised on international relations. I think what’s driving current Taiwan – PRC relations is largely a very complicated history and social animosity between the two peoples. Sadly these things do exist like other common prejudices, no matter how much some people hate to admit. For the record in Asian circles the unofficial social pecking order among Chinese people in East Asia are Hong Kong > Taiwan > Mainland China. It’s also not a coincidence that the GDP per capita of these three “states” follows the same descending order. Over the next several decades, as China outgrows both special regions economically, this will likely change.

As for the event risk pertaining to a massive banking crises/currency devaluation, if you have an anti-fragile organization with ample liquidity and a quality LP base then you’ll likely benefit from such an event. Being the incremental buyer of last resort/provider of liquidity in times of great market distress has often times yielded very lucrative gains for the prepared… Most importantly, I’m not smart enough to know what the macro developments will be like over the next several years; I’ll admit it. Short-term political and economic based speculations are worthless and futile anyway. However, on the longer-term considerations, such as over the next 10, 20, 30, years, I’m confident that income per capita in China will grow enormously.

With all the added uncertainty that comes with investing in a foreign market – and one that is so culturally unfamiliar for many Westerners – why even bother? Quite simply, the exceptional return potential available at large scale. The number of long-term oriented institutional investors in China remains close to non-existent today. Honestly, investing is never easy, but structural advantages do exist for those open-minded enough to exploit them. The few smart, long-term oriented investors focussed on China such as Hillhouse and Capital Today got in really early, simply picked the best companies levered to the Chinese consumption and internet theme,  and as a result, accumulated a few of, if not the best investing track records over the past 10+ years. It didn’t take a genius to figure this out, but a disciplined approach to long-term investing.

Liberty Global / LILA/K

No post on this blog is serious without a discussion of a Malone-related company. It’s safe to say that ever since LILAK became a tracker, the story has been a complete shit show.

Has it historically been wise to invest alongside Malone? For the most part, yes. But now with dozens of related entities within the complex, we have to be very careful with picking the best story.

Let’s start with Global. First, the failed Vodafone transaction in the spring of 2015 caused the event-driven funds to sell en masse. The unofficial explanation provided by management was a valuation gap that couldn’t be bridged. This sounds reasonable. Vodafone is primarily a wireless company, and wireless companies in general deserve a lower multiple than higher-quality fiber/coaxial-based fixed assets. Soon thereafter, Fries starts giving out “long-term OCF” guidance of 7 – 9% OCF growth for the next few years (for the first time in the company’s public history, I might add) as the competitive situation in the Netherlands continues to deteriorate and the Ziggo integrations are becoming more challenging than originally estimated. Even up to this day, there is no indication that the KPN Fiber JV in the Netherlands is slowing down its FTTH rollout or its aggressive promotional pricing.

Unfortunately, the competitive position in Switzerland appears to be deteriorating as well. Most of Global’s issues in its most competitive markets is a problem of high population density making the economics of overbuilding a lot more attractive for European Telcos. The Netherlands for example, with a super dense pop/sq km of 400+, makes FTTH a very viable economic proposition over large parts of the country. On top, Gfast technology makes it possible to further upgrade copper networks in these shorter local loop markets; Gfast can provide downstream speeds up to the several hundred mpbs range with a relatively cheap incremental cost of capital compared to FTTH. Forced unbundling regulation of the local loop supported by many Socialist regimes in Europe also adds further competition by resellers. As the minimum speeds being offered by the Telcos continue to be well above the industry demand curve, pricing power for both duopoly operators and resellers will likely continue to be soft. Europe, in contrast to the US, is simply a much more competitive broadband market.

With these issues at hand, I don’t see a path towards 7 – 9% OCF growth. No one believed management’s announced guidance and for good reason. I think even 5% OCF growth including project Lightning could be a stretch. Now they’re saying 7 – 9% excluding Ziggo, just lol.

So why has Mike Fries suddenly become so promotional in such an “unliberty-like” fashion? First of all he got most of his options and RSUs struck earlier this year when the stock got absolutely crushed. Secondly, there’s the added incentive to sell/merge with Vodafone if your operating performance is deteriorating. It’s obvious a stronger share price brings with it a stronger bargaining chip to the table with Vodafone.

As stated, management’s rationale for demanding a discount on wireless assets given their lower quality is reasonable but what I don’t understand is the disconnect between paying 11x post-synergy EBITDA for an Latin American asset with a considerable wireless mix vs. not striking a deal with Vodafone at likely better terms and massive synergy benefits.

Now with the deteriorating results of CWC along with the accounting adjustment from IFRS to US GAAP for CWC/Columbus (shouldn’t they have adjusted their purchase multiple for these changes beforehand?), it’s obvious they overpaid.

The LILAK thesis for many Liberty followers was largely based on trust in the management team given their multi-decade success investing in and operating cable assets globally. This becomes even more important when it comes to investing in a business operating less familiar/foreign regions. But just the credibility I think Mike Fries lost over the past year or so with his sketchy behaviour should warrant some caution… I like to think as the management team you get the stock price and the shareholder base you deserve as results come in over time. Fries should seriously just focus on operating the business well instead of revising the guidance every single quarter.

LILAK a narrative parallel to TCI or early Global/UPC?

I don’t think so. TCI was an actual monopoly in the burgeoning US Pay-TV industry when wireless and high-speed data services were not around. TCI was a major financial backer of upstart cable networks which allowed the company to distribute exclusive programming in its pipes. Market power can be pretty insane if you control both the content in the pipe and the distribution. On the M&A front Malone was rolling-up local cable systems run by mom-and-pop entrepreneurs at valuations probably closer to 4 – 5x EBITDA. At peak, TCI had a 30%+ market share of all US Pay-TV homes. When cable and telecommunication assets were near/at peak valuation levels at the turn of the century, satellite distribution became an emerging threat, interest rates were heading up, and the balance of negotiating power was shifting to the cable networks, Malone sold TCI to AT&T. The timing was impeccable.

Will the CEO of LILAK be capable of executing such a successful playbook in Latin America? Maybe… but I seriously doubt it. Seems like it’s been dealt a much tougher hand.

Charter Communications / Liberty Broadband – Still the Best Special Situation Compounder that everyone has heard of

Houston, we have a cost problem:

Excluding programming costs, the average US cable operator has an OpEx per resi sub/month cost of ~$50. For European peers, this is closer to $15 per sub. Certainly household density, competitive intensity and labor costs play a major role here, but the gap is startlingly wide. It’s likely that for the longest time US cable operators were run by complacent monopolists. This may soon be about to change.

Patrick Drahi is still a stud:

Last year Altice entered the US cable industry with their purchase of Suddenlink and CVC, along with announcing massive cost cuts and synergies. Believe it or not, Altice’s $900mm synergy target for CVC is likely achievable and potentially conservative. Excluding programming, CVC had an OpEx cost per resi sub/month of $60 – 65. Assuming Altice fires all of the family friends the Dolan’s hired as middle managers earning a $300K+ salary, and stripping out the Dolan’s/senior management compensation, among other bloated HQ expenses, we could potentially get to roughly $100 – 200mm in overhead cuts. That leaves about $700 – 800mm left for resi sub-related cost cuts. The implied target OpEx synergies management are looking to shave are then around $20 – 25 per resi sub/month. If these targets are met, CVC’s OpEx per resi sub/month will be $40 – 45, which is still well above its European peers at $15 per sub! Numericable for example has an OpEx per sub/month of $15, and let’s not forget, France is a much more competitive market than FiOS territory. That’s just on the OpEx side, as we haven’t factored in any programming cost synergies with Suddenlink, or their plans to move towards more skinny Pay-TV bundles and outright drop certain networks. This could shave another few hundred million off. Finally, the CapEx per passing for CVC is much higher than European peers due to the many duplicate hardware in their network. If upgraded and maintained under Altice, this could shave another $100 – 200mm off long-term.

So on Charter, I think the real, and still rather conservative, cost-synergies will be closer to $1.4 – 1.6bn on full run-rate numbers by 2019. Just the programming synergies alone could be $650mm in a few years if we assume Charter moves to TWC/BH’s lower rate card and if they achieve a pro-forma programming rate card in the mid-$50s per resi sub/month by year 2 post-merger. If we model just an OpEx per resi sub/month for Charter of ~$50 just from the scale benefits of merging multiple systems in a contiguous combined footprint, we can get to over a billion in OpEx synergies easily. Just from the OpEx and the programming synergies alone we can get to $1.6bn in cost savings or double management’s $800mm guidance. Also, this figure doesn’t account for the duplicative overhead and executive compensation costs for BH and TWC, nor any CapEx savings from the benefits of leveraging a larger sub base for purchasing set-top boxes, modems and other network equipment…

So over time, it’s very possible that the actual synergies end up being double or even triple management’s original $800mm estimate – which would still be a MSD % of pro-forma revenues – and roughly in-line with prior cable merger deal synergies.

Why is there so much upside on the table? It’s obvious there was management sandbagging from the get go. This was done to avoid overbearing regulatory scrutiny over job cuts and announcing an attractive profitability outlook for rationalizing a huge merger. Also, Rutledge’s compensation package (including options and RSUs with exercise prices up to the $500’s per share) for the pro-forma business was struck shortly after the merger closed, incentivizing him to downplay the potential upside. If these ambitious share price targets are met over the next 5 years, he will become a billionaire on paper. The only strong pushback I can think of is that Rutledge is not known as a cost-cutter, despite being a great cable operator. However, now with the Trump administration in place along with a less onerous regulatory environment, I see potentially more leeway for aggressive cost cuts. On the revenue front, the new industry-friendly FCC could mean the approval of data caps and usage-based pricing which would likely mean additional upside for broadband pricing.

As for the HSD penetration runway, If we define “high-speed broadband” as a service providing minimum downstream speeds of up to 25 – 50mbps, then “real” US high-speed broadband penetration is understated relative to published industry figures (which typically define a broadband household as one with any fixed-line internet connection including DSL). If high-speed broadband penetration defined under these minimum speeds eventually reaches 80 – 90% in the US over the coming 10 – 20 years (not unreasonable given the robust demand outlook), then it’s possible that Cable will take the vast majority of incremental share over the next couple of decades. We’ll likely move to 10GB eventually, but this will probably be closer to 15 – 20 years from now. Cable’s technology roadmap looks promising and it appears DOCSIS 3.1 alone (which is slowly getting rolled out by Comcast and Cox in select markets over the next few years) can get Cable to those speeds without upgrading the last-mile to fiber-optic.

Currently Charter’s HSD penetration is ~45% within its footprint. If the industry moves towards 90% penetration eventually, and assuming a 50/50 share split between Cable and FiOS (16% overlap), a slightly more favorable share split between Cable and U-Verse/upgraded DSL (25% overlap), and of course a 100% share split vs DSL in the rest of Charter’s network, Charter can potentially grow its top-line at at least MSD rates for the next couple of decades, driven by new net adds, modest base inflationary pricing expectations, upselling faster data speeds, and increasing commercial sub penetration. In any event, I think with the synergies outlined and the fantastic incremental margins from broadband subscription growth, Charter’s EBITDA margins should be closer to at least 40 – 45% 7 – 10 years from now, which would still be below Altice US’s targets for CVC and Suddenlink at 50% and European peers.

As for 5G, I’m too lazy to discuss that today, so let’s leave it for another time, or on a good day.

On valuation, assuming a continuously levered capital structure at 4.0 – 4.5x with aggressive capital redeployment into share buybacks, I see a pathway to $40 – 50 of fully-taxed levered FCF per share by 2020 or $20.5bn+ of EBITDA at a slightly less than 40% margin. At an implied 9 – 10x EBITDA exit multiple / or ~18x levered FCF per share of $40 – 50 (potentially very conservative assumed exit multiples given the aforementioned multi-decade attractive growth runway), Charter stock is worth $750 – 900 per share 4 years from now.

In conclusion, I believe Charter Communications / Liberty Broadband remains the most attractive cable asset to own globally and one of the most attractive long-term investments in the US large-cap TMT universe today. Did I mention John Malone? 🙂

“All of humanity’s problems stem from man’s inability to sit quietly in a room alone.” – Blaise Pascal

Happy Holidays,


[1]Outside of getting acquired which I think is increasingly unlikely other than a carve-out of the studio business, I still think Viacom is a short and the equity could potentially be worth close to zero 4 – 5 years from now. The cable network segment comprises of the bulk of the value of the company and earns juicy EBITDA margins north of 35%.

[2] We also know for a fact that their last-mile unit cost per order is lower or at least on par with any of the prices third-party last mile franchises charge to merchants (assuming we take into account all the costs across the entire last-mile delivery chain including inbound delivery, sortation centers, line-haul, delivery stations I estimate JD’s average fulfillment and last-mile delivery cost per order was ~$1.70/$0.85 in FY2015 (vs 8 – 10 RMB for the industry ASPs), at the same time JD’s service quality is superior (based on customer service ratings/surveys, cash on delivery service available), and faster, more reliable, with a friendlier return policy. So better service than the industry with a marginal cost structure equal to or lower than industry ASPs.

[3] In fact, e-commerce search engines will likely become the most valuable advertising properties long-term over horizontal search engines as incremental traffic growth increasingly shifts directly to e-commerce properties; these properties are superior in creating purchasing intent right at the bottom of the funnel.

[4] There’s a big difference between how asymmetric the risk-reward profile of an investment (or how wide the distribution of future outcomes) is vs. the probability of a large permanent loss of capital.

[5] The barriers to entry in new or adjacent markets are lowered due to the valuable customer data and existing scale built in an adjacent market.

[6] I’ve observed that successful leaders may have many different management styles and philosophies. If I had to guess the most important thing for a CEO on a daily basis is to prioritize tasks. But on a more strategic level prioritizing company long-term initiatives is as equally if not more important. How Bezos allocates his time within an increasingly complex business such as Amazon with many different projects and moonshots is a mystery to me. At Baidu you can bet Robin is spending most of his time on AI as he knows the search business will soon be in secular decline. Like a lot of great tech CEOs the recurring pattern appears to be that they are extremely passionate about their business, have a solid technical background, quite introverted and typically not very charismatic. The best tech CEOs have all these traits at minimum and are great allocators as well!

A Response to the JD Short Thesis on SumZero

Some of you may have seen the 50-page JD short write-up that was recently posted on SumZero. The author essentially compares JD to Valeant and slaps on a $12 price target for his short thesis.

Coming up with such a sensasonal comparison, I was expecting to come across some high-quality analysis to back-up the attention-grabbing headline and hopefully learn some new insights from someone with an opposing view. Instead, what I got was a very biased and inaccurate report. A condensed version of the short seller’s paper can be found a here.

Below I will address all of the short seller’s major arguments.


  1. Returns/cancellations may be as high as 20-25%. Net GMV removed after IPO.

Anyone who has put in some effort to study JD and the e-commerce industry knows that gross/reported GMV is before returns/cancellations (which the company clearly defines in its filings). That is practically how any discerning analyst models the business including the sell-side. I don’t know anyone who models gross GMV as being completely monetizable, or bases their valuation off of gross GMV alone. All Chinese e-commerce companies disclose only gross GMV which is common knowledge for those that have studied the industry, but somehow for the short seller this is a red flag for JD…

JD stopped disclosing net GMV after the IPO because all of its competitors including BABA only report gross GMV and reporting agencies/media were using JD’s net GMV figure to calculate their respective market share vs. Tmall which basically penalized them vs. the competition. But if you call up IR which doesn’t seem like the short-seller did, they would tell you what the net GMV and implied return/cancellation rates are. To the extent that gross GMV is useful as a financial metric, growth in gross GMV could be used as a rough proxy for measuring growth in net GMV or “net monetizable revenue” once one has come up with a reasonable return/cancellation rate.[1] Certainly this metric when properly adjusted is a key business driver for a company scaling in high-growth mode and by nature is very different from the questionable metrics used by internet companies during the dot-com boom as the author tries to draw a historical comparison to…

Historically returns/cancellations have been about 18-21% of GMV for the 1P and 3P business (IR will confirm this with you). Another way to confirm this is through public filings as JD reports both the gross GMV and net revenue figures (under the “online direct-sales” line) for the 1P business (the only difference between the company’s definition of net GMV and net revenue is the VAT which is 17%).

For example, in FY2015 JD reported gross GMV of RMB 255.6 billion and net revenue of 167.7 billion for the 1P business. The way to get the return/cancellation rate is to arrive at net GMV by adding back the VAT (167.7/0.83 = 202 billion). The difference between gross GMV and net is 53.6 billion (255.6 – 202) which implies a 20.9% return/cancellation rate. Where the short-seller miscalculated is instead of dividing the net sales by 0.83 (if the tax is 17% then what remains is 83%) to properly account for the effect of the VAT, he mistakenly grossed it up by 17% which understates net GMV and overstates the return/cancellation rate.

point 1 jd

In any case, we can see that the return/cancellation rate has trended up over time. There are two primary reasons for this as management has clearly communicated: 1. The mix-shift away from electronics into more general merchandise, especially apparel, has increased the return rate (the return rate is still at industry-low levels around mid-to-high single digits whereas cancelled orders make up around 80% of the gross-to-net GMV gap); 2. More purchases are now done on mobile (mobile accounted for 30% of total orders in Q3 2014 vs. 72% by Q1 2016) where consumers tend to leave more orders uncompleted and where there is likely more impulse-type buying. Management is working to bring the completion rate on mobile down to a level more similar to desktop over time.

Accusing the company of a lack of disclosure where JD clearly defines gross GMV and net GMV in its filings and would readily supply information on net GMV over a simple phone call seems pretty ridiculous.


  1. Brushing on 3P marketplace reported by media and recent JD comments: 10% GMV discount

Brushing is a common phenomenon on all marketplace e-commerce platforms including Taobao/Tmall, eBay and Amazon (I am reasonably knowledgeable about this as one of my best friends was a Powerseller on eBay, where brushing is extremely common). The idea is that to get better reviews and more importantly, a higher review count for their products, some 3P merchants would “brush” by buying their own products and write favorable reviews (and never delivering the product, for obvious reasons), in turn driving up the reputation of their stores and their rankings over keyword searches.

This problem in particular is prevalent over Taobao which does not charge a sales commission, and hence “brushing” is practically free for the merchant. On Taobao and to some extent on Tmall, this introduces an adverse selection problem on the platform as the bad actors crowd out the good actors who refuse to brush and artificially boost their ratings.

Brushing is a much smaller issue on JD for a number of reasons. Firstly, 56% of JD’s LTM GMV is 1P (self-operated) which by definition has zero brushing. Secondly, a very substantial % of GMV on the 3P platform is generated by flagship stores owned by the brands themselves who rarely engage in brushing. Thirdly, JD’s commission is much higher than Taobao/Tmall which makes brushing very costly for the sellers and lowers their ROIC on brushing relative to more legitimate marketing channels such as keyword bids and display ads. Finally, all of JD’s merchants are large brands or retailers that are over a certain size limit; JD is much more selective about its 3P merchants (far more initial due diligence and ongoing monitoring on marketplace merchants) and has a soft cap on the number of merchants allowed on its platform (JD only has around 100,000 3P stores vs. over 10 million on Taobao).

When a 3P merchant is caught brushing, JD generally kicks the seller off the website permanently whereas Alibaba follows a much more lenient “3-strike” rule (only removing the seller after catching him brushing/selling counterfeits three times). Unlike Alibaba which refuses to acknowledge brushing as a problem, JD has been heavily investing in anti-brushing technologies and has specifically called out their recent anti-brushing campaign as having a modest impact on GMV growth.

Furthermore, JD’s reported GMV has always excluded GMV generated by merchants with abnormally high return and cancellation rates, as well as any cancelled orders above 2,000 RMB to partially account for the effect of brushing. No other Chinese e-commerce company is as conservative with GMV reporting. Brushing is estimated to account for as much as 15 – 25% of Taobao’s GMV whereas most estimates put it at low to mid-single digits of 3P GMV for JD. Assuming 5% of JD’s 3P GMV is “brushed”, the impact on JD’s total GMV is less than 2.5%. Moreover, since 3P merchants mainly use brushing as a marketing tool, as JD’s technologies make it harder to brush, these sellers will increasingly have a stronger incentive to shift their marketing spending to keyword search advertising which would ultimately help offset any lost commission revenues related to brushed transactions.

As for the newspaper article which alleged that a JD employee helped sellers with brushing activities, JD has specifically responded that the employee was new to the company and was immediately terminated after the incident. The reality is no matter how much effort is exerted, you cannot completely prevent corrupt behavior at a company with over 110,000 employees.


  1. Improper counting by orders placed and limited disclosure: 5% GMV discount

This argument is just plain bizarre. The author claims that GMV is inflated because 1. It includes shipping fees (well, nearly all of JD’s orders include free shipping to the customer and the shipping service is a key component of the online shopping value proposition; Alibaba also includes shipping revenue in its GMV and Amazon includes shipping in net revenue); 2. It includes GMV from equity investees – by that logic if Amazon owned 20% of one of its suppliers it should not count revenue derived from that supplier as actual revenue!?); 3. It includes internet finance revenue (will be discussed in point 9 below).


  1. China-based auditor is unable to audit 100s of millions of orders: 5% GMV discount (We note that it is a gigantic task for JD’s auditor, PricewaterhouseCoopers Zhong Tian LLP to audit 100s of millions of orders placed which count towards GMV. As such given even the auditor cannot audit GMV, we would place an additional 5% discount/hair cut to the reported GMV figure.)

So the author has decided to give JD an arbitrary 5% GMV discount simply because it is impossible for the auditor to independently verify every single order on JD. I’m out of words here. Seriously, the same can be said about any large company engaged in large transaction volumes – Amazon, P&G, Wal-mart, JC Penney – you name it… Perhaps the author is trying to imply that management are disingenuous when disclosing key business drivers? In no part of the author’s paper did he touch upon any aspect of the founder’s background, character, or past dealings with major stakeholders – perhaps this not so surprising in such a biased paper as Liu could potentially be one of the most ethical corporate executives alive today.


  1. JD is Enamored with GMV While Losses Grow Larger

JD’s core business is actually profitable, despite operating in a growth phase. The current reported losses are due to one-time impairments and current investments running through the PnL that are tied to fast-growing businesses such as JD Finance and JD Home/Daojia that are at present subscale. JD’s reported FCF profile also turned negative in 2015 in part due to the heavy investment in fulfillment center construction – this is mostly growth-related CapEx. But if you analyze JD’s FCF generation a few years back, you can clearly see that the core business is solidly profitable and this is in large part due to the Finance unit being started only in 2013, coupled with the company’s prior strategy of leasing warehouses to meet order capacity. If the author provided a sound argument that JD is investing in low-return projects, then criticism of the current heavy spending is warranted.

Perhaps the author didn’t catch this, but JD added two new internal KPIs for 2016 (that will affect employee compensation) – and both are focused on increasing profitability – Cash Flow Days and Operating Profit.  The author has also omitted the steady gross margin expansion JD has shown year after year driven by rapidly growing 3P contribution profits.

I believe the right way to think about the GMV metric is to use it as a relevant benchmark for comparing the profitability and margin structure of the target business to other offline and online retailers at similar scale. GMV is a relevant metric for hybrid or pure-play marketplace retailers since a marketplace business (commission-based revenue model) has a very different margin structure to a direct-sales/1P business. Depending on the average product/gross margin of the 1P business, and the average commission/take-rate of the 3P business, comparing the profitability of a hybrid online retailer to a pure-play direct-sales retailer by assigning a steady-state or long-term margin on GMV could be a useful tool for analysis. Conventional analysis such as assigning a margin as a % of revenue for both 1P and 3P units is also possible, but less useful for straight comparisons between different retail models. Remember that GMV should always be adjusted for returns/cancellations, and any VAT taxes. No competent analyst I know of uses gross GMV for straight comparison/modelling purposes.

What the author seems to be missing is the massive operating leverage that will become evident from owning your own fulfillment assets as order volumes continue to grow as well as the sizable gross margin expansion from 3P GMV growth and increasing scale in the 1P business. The author seems to be skeptical that JD’s 1P gross margins will eventually expand despite his own admission that GMV will likely grow at least in the 30%+ range (faster than any direct competitor). I view this risk as very low given JD’s track record in execution and the fact that they are already the largest direct-sales retailer and are growing all categories at a volume rate multiples of that of their largest rivals Suning and Gome. The author also does not take into account the considerable likely incremental cost-savings in the future from more automated warehouses and the use of delivery drones.

Ultimately, it is the margins (gross and operating) and FCF, in particular, that matters most to JD’s long-term intrinsic value (something the author has yet to have an opinion on) and JD management understand this as well (apparently the author doesn’t think so) and have their own opinion of what type of margins JD could achieve long-term if you speak with them. Overall, it appears that it is the author who is myopically focused on GMV and short-term reported losses over having an informed view of what the economics of JD’s business could look like longer-term.


  1. Paipai acquisition boosted JD’s GMV by RMB 33.6bn ($5.3bn). GMV was counted, but RMB 2.8bn impairment created for shareholders. Questionable capital allocation for GMV?

The $5.3 billion the author quoted is cumulative GMV over two years (seems like he made the number artificially larger to exaggerate his claim). Ever since the Tencent – JD strategic partnership transaction closed, JD has always disclosed core GMV on an ex-Paipai basis (Paipai is a C2C platform originally owned by Tencent). JD shuttered Paipai last year because as a C2C platform similar to Taobao, Paipai is fraught with counterfeit products and given the sheer number of small merchants on the platform, it is practically impossible for JD to enforce a counterfeit-free platform. In sum, JD’s decision to kill Paipai was motivated more by reputation management over short-term economic considerations.

More importantly, the author omitted the biggest component of the Tencent deal which gave JD an exclusive position on Tencent’s flagship mobile apps WeChat and QQ (35% of all time spent on the internet in China is spent on Wechat). WeChat and QQ now account for 20 – 25% of JD’s customer traffic and have helped JD significantly accelerate its growth profile. To completely disregard this contribution while isolating a minor part of the deal to discredit management’s capital allocation track record is one example that highlights the author’s overly biased position.


  1. JD counts in its GMV orders placed in JDs websites but operated by equity investees like Bitauto and Tuniu where traffic generated through JD platform…The question is why should JD be counting in its GMV when this is an equity investee?

It doesn’t appear that the author has ever used JD’s website. GMV generated through both Bitauto and Tuniu actually take place on JD’s platform under its travel and auto verticals so Bitauto and Tuniu are effectively 3P sellers on JD just like any other 3P merchant. Both account for a negligible amount of JD’s GMV currently.


  1. These investments (Tuniu, Bitauto and Yonghui) have been a dud and burned real cash…

This argument is analogous to saying that because a stock in your portfolio has fallen below your cost basis that it must be a permanent impairment of capital… please spare me.

Both Tuniu and Bitauto are publicly traded on the NASDAQ and JD’s write-down of value in both was driven by the fact that their market capitalizations have fallen substantially over the past year (JD acquired a stake in both during 2015). For context, Tuniu is the third largest Online Travel Agency in China (think Priceline) and has recently established an alliance with industry leader Ctrip (forming effectively a monopoly); Bitauto is one of the top two advertising platforms in the internet auto vertical. Both are rapidly growing players in industries with massive TAMs. To say that the investments have been a waste of money simply because of short-term volatility appears quite unfair to say the least.

JD has not realized any capital losses on these investments and it is also obvious that Chinese-related securities as a whole have not been performing well over the past year. Both stocks have gotten crushed over the past year (each down by more than 50%+) and I am not saying that over time JD’s investment in either Tuniu or Bitauto might not turn out to be a disaster but I definitely think it is premature to jump to a conclusion that they are without properly assessing management’s decision. One has to consider that management is very focused on long-term value creation and that is also reflected in their capital allocation decisions. Now if the author provided some sound fundamental analysis to support his claim that Tuniu and Bitauto were a waste of shareholder cash instead of relying on Mr. Market as his informer of intrinsic value then I’m all ears. In any case whether these equity investees become very valuable or not, they are not key value drivers to either the long or short thesis.


  1. JD’s consumer finance business GMV as % of GMV has been rising (with limited disclosure). JD has no track record in growing, operating and managing risks in internet finance.

First of all, what JD refers to as consumer finance GMV is not the revenue of the JD Finance business but rather the portion of GMV that’s tied to JD’s consumer financing offerings (for example, it’s the GMV recorded when you buy, a fridge on a consumer loan from JD Finance). JD has separated JD Finance into a standalone unit that will be self-financed in 2016. This unit was valued at $7.2bn post-money in January (which the author conveniently omitted in his analysis) by top-tier investors and will likely be IPO’d sometime within the next few years. It is true that JD’s experience managing internet finance businesses is rather short but the same can be said to just about any internet finance company on the planet. The whole point is they will try to get better at it and create value by leveraging JD’s massive distribution platforms. To reiterate in my original JD paper posted several weeks ago, JD Finance’s management team is prioritizing risk management over speculative lending growth.


  1. As GMV grows, are investors underestimating the risks in JD Finance business? Internet Finance is a space with intense competition with larger, more profitable players!

Internet finance is a multi-trillion dollar opportunity in China. It appears the author believes this is a liability (note that the business is now siloed away from the core business in the worst case scenario that it never achieves minimum efficient scale) solely on the rationale that 1. This is a new area for JD and 2. This area is competitive and yet extremely well-informed people at Sequoia think this business is worth at least $7bn.

Admittedly there is competition in the Chinese internet finance space, especially on the online payments side given the larger size of AliPay for third-party payments. But as I mentioned in my paper, the market opportunity is large enough that at this early stage multiple players can create a considerable amount of value given their online distribution advantages and richer data sets over the state-owned banks.

Has the author provided any rationale that this business is not worth at least $7bn today? No, in fact he did not touch upon the nature of the fundraising round in January. Bigger picture, JD’s core business alone could be worth at least $200bn over the next 5 – 7 years, which frankly would make the internet finance division’s value look like peanuts today.


  1. JD’s Management Has Limited Tenure For a 12 Year Old Company Which is Rarely a Promising Sign For a Company Yet to Make Money

The author really just answered his own questions here. JD is a very young company and as a result has short-tenured executives. It is also important to keep in mind that when JD was founded it was very much a one man show where Richard pretty much took on most of the executive responsibilities as the company did not have the money nor reputation to hire professional managers. JD upgraded its C-suite in the early 2010s after receiving investments from Hillhouse and Tiger Global and that allowed Richard to delegate more responsibilities to professional managers.


  1. Who is Jiaming Sun Owning 55% of the VIE, more than Richard Liu, JD’s Founder? A Mystery

It wouldn’t be such a mystery if the author had done any research in actual Chinese. Sun Jiaming is one of JD’s earliest employees and as recently as a year ago was the SVP of General Merchandise at JD. He recently left the company for personal reasons but like Liu he had to enter into a series of contractual agreements to ensure that JD would retain effective control over the VIEs should he stop being a JD employee.


  1. JD is a Hedge Fund Darling and Has Nearly Unanimous Buy Ratings by Analysts. Is It A Crowded Trade Like Valeant?

Yes there are some very smart people who own JD in size who hold a better opinion of the company than the author. By the author’s logic is any name heavily owned by hedge funds doomed to fail as an investment?

It’s simply not enough to be “contrarian” to do well with an investment – you also have to be right. Simply calling JD a hedge fund hotel and comparing it to Valeant without providing any sound fundamental analysis to back-up the claim basically kills any credibility the author had in my mind.

Looking at JD’s shareholder base and the extensive sell-side coverage, I would view the story as pretty well-known by major institutional investors (so it is definitely not an “under-the-radar” type of mis-pricing): JD is typically pitched as a vehicle to gain exposure to the growing Chinese internet and consumption theme. However, I am not convinced that the long thesis is well understood by the majority of potential shareholders who sit in an office in New York and lack a variant perception. As for my numbers they are materially higher than Street consensus over the next several years and I am highly skeptical that the market is accurately pricing in the long-term earnings power of the business. If the current stock price represents the market consensus then I think a sound case can be made that consensus is way too conservative on a longer-term time frame.

I would also note that there are those that actually conduct their own independent due diligence in order to develop a well-informed view of their holdings (and pencil out the numbers accordingly) vs. those that conduct limited due diligence and simply piggyback off of their buddy’s ideas at other funds. Price-agnostic buying of “outsider” companies based solely on getting exposure to a robust narrative such as TDG, CSU, or any Liberty-related story, for example, is almost a guaranteed path to generating mediocre returns.

It is also important to consider the quality of the funds holding the stock vs. simply labelling the stock as a hedge fund hotel that is ready to crash. Different funds have different time horizons, investor bases (quality matters) and average cost bases that could affect the short-term liquidity of shares. In JD for example, despite the numerous Tiger Cubs in the stock I basically treat their exposure as one de facto position since they like to share the same ideas with each other (hedge fund circle jerk is real). Hillhouse of course is well-known for taking concentrated positions with a very long-term time horizon.

Frankly when it comes down to it whoever owns the stock is irrelevant for the long-term fundamental investor in my view – except in the rare case that a large shareholder is actually adding value to management by leveraging its exceptional proprietary research to create additional value. Over the short-term it could affect the stock’s liquidity and cause higher than average volatility since shorter-term funds can crowd in and out of the stock – I view these liquidity events as more of a potential opportunity for the longer-term investor – but it could very well be a real risk for the shorter-term investor/trader.

Finally the stock is hitting fresh 52-week lows and is near its IPO price which took place more than 2 years ago; sentiment at the moment appears quite poor – calling it a crowded trade like Valeant at this point seems like a stretch at best… I am not even going to get into how JD is different from VRX since the author obviously made no worthwhile effort in his own case despite his claim.


  1. “Services and other” revenue has deferred revenue included and limited disclosure

The author seems to take issue with the fact that 4% of JD’s deferred revenue account (a whopping $80 million) is related to Bitauto and Tuniu. “At APS, we note that there are accounting ways counted here like deferred revenue from 2015 resulting from “resource” arrangements with Bitauto and Tuniu, equity stakes where JD is deeply underwater. Burning cash but booking accounting revenue?”

Really not sure what the author here is suggesting really. There is a clear business relationship between JD and Bitauto and Tuniu both of which are NASDAQ-listed industry leaders. To not book revenue would be illogical when a service is provided or expected to be provided.


  1. JD’s Non-GAAP gross margins need to be adjusted to reality
  1. The author is correct in stating that JD’s 1P gross margin has been flat in the last four years. However, this has been the result of an intentionally low-pricing policy as part of management’s plan to quickly gain scale (management has stated quite explicitly at meetings that they have been deliberately holding gross margin flat and reinvesting scale benefits back into pricing), similar to what Amazon has been doing for the past twenty years. One interesting data point is that despite being the largest retailer in China, JD’s gross margin is over 10 points below its offline peers such as Suning and Gome which has helped them to win substantial share in the price war. 2016 is the first year JD is letting 1P gross margin expand (already increased 80 bps in Q1 this year) and management’s plan calls for 1P gross margin to increase from around 6.5-7% currently to 13-14% in the next few years. In my own estimates I am assuming 1P gross margins to expand closer to 12% by 2020.
  1. Outbound/last-mile shipping costs are not captured in COGS – this is also correct, but the author is completely wrong on everything else when it comes to JD’s cost structure compared to Amazon. First of all, JD lumps its shipping expenses into the fulfillment costs line under Opex because JD’s fulfillment operations are integrated with its proprietary last-mile delivery. Amazon’s fulfillment cost is also recorded as an Opex line item but outbound shipping is lumped into COGS because Amazon outsources the majority of its last-mile delivery to UPS/USPS/FedEx. The short-seller’s analysis which adds all of JD’s fulfillment costs back to COGS is misleading and not comparable to Amazon. To make JD’s results apples-to-apples to Amazon we need to add back only the last-mile delivery portion of fulfillment. More importantly, Amazon also has a substantial 3P operation (50% of Amazon’s orders are 3P compared to 44% for JD) which similar to JD is 100% gross margin and overstates Amazon’s 1P gross margin in the consolidated revenue line. Finally, prime membership fees and AWS revenue at Amazon are both very high gross margin revenue streams (offset by much higher Opex ratios than retail) that skew the numbers.
  1. JD’s use of weighted average inventory – the author’s assertion that electronics prices are falling 10 – 15% a year is purely fictional. According to the National Stats Bureau of China deflation in prices of electronics and appliances has only averaged about 1.3% for the past three years. Also the author seems to be forgetting that JD is a direct retailer with substantial scale, not a supplier that may have to eat deflationary pricing. More importantly, JD’s inventory turnover is very fast at about once every 35 days so COGS for the first 11 months of the year would be the same regardless of whether JD uses weighted average or FIFO. The impact of the inventory accounting choice is very small.


  1. JD’s inventory has grown faster than 1P direct sales in 4 out of the last 6 years and in the last 2 years. Rarely a promising sign given huge $3bn of inventory as of Dec 2015

Even according to the author’s table inventory turnover has been very much flat for the last five years (2009 is irrelevant given the company was super tiny at the time):

jd point 16

Also it’s important to note how fast the inventory turnover is – if you calculate inventory turnover correctly (by using average inventory for the denominator rather than ending inventory like the author did), JD’s inventory turnover comes out to be over 10x or once every 35 days.


  1. JD relies on Non-GAAP accounting metrics but several of the expenses should be viewed as real expenses and given real cash was burned on investments

The author is correct in stating that stock-based comp should not be added back as it is a real expense. However, practically ALL best-in-class high-tech companies do it so JD’s practice is not out of line with industry standards. Furthermore, stock-based comp at JD makes up a much smaller % of its market cap than most Silicon Valley tech companies, suggesting that JD dilutes shareholders less on an annual basis. Note that on my own numbers I fully expense stock-based comp as a real expense when assessing profitability. In reality whether you adjust for SBC or not, the equity here is demonstrably undervalued.


  1. JD consolidated operating cash before internet finance working capital is scant and GMV conversion to operating cash is limited even before internet finance business started

JD actually has generated positive maintenance free cash flow every year in the last four years when CFO is adjusted for the internet finance business which is a negative drag to JD’s working capital position. But more importantly, JD is currently in a high-growth phase and accordingly has been heavily investing in growth-related capex over the past couple of years (tied to expanding fulfillment capacity via fulfillment center buildouts, equipment, etc) like any fast-growing online retailer well-positioned in the most attractive e-commerce market globally would be.

It is true that JD has reported little accounting profits despite raising a substantial amount of capital over its rather short 12-year history as a company. But I think it is first-level thinking to take reported financials as a proper measure of the economics of the business or assume that the business will be in a perpetual state of cash burn – without being thoughtful about how the economics of the business will look like in the future. My own view is that JD has steadily increased its earnings power over time, management has consistently focused on maximizing value over the long-term (and this means raising enough capital historically to pursue this goal), and the time and capital it took to build a moat around the business today illustrates the substantial barriers to entry in becoming a tier-1 e-commerce platform.

Allow me to demonstrate in greater detail why I believe the author’s thinking is flawed here. Aside from wrongly adjusting the Gross GMV figure, the author failed to properly adjust JD’s reported EBIT and FCF to normalized levels in order to account for the substantial growth-related investments. A more sophisticated type of analysis would include a deep dive into the unit economics of JD’s business on a customer and/or order-level basis in order to assess whether any incremental invested spending/capital is generating a sufficient incremental rate of return. Now one can have a legitimate debate about the magnitude of the rate of change over time for these metrics, yet the author has failed to even mention them. At this point in time, all the key drivers are solidly moving in the right direction (and have been improving historically) and JD’s unit economics have already hit a positive inflection point (customer-level IRRs are likely already in the triple-digit range driven mainly by the growth of 3P contribution profits, a steadily declining fulfilment cost per order/customer, and a larger customer base). For the long-term investor, it is a thing of beauty to be invested in a company that has a long runway to invest substantial amounts of capital at high incremental returns.


  1. JD has a number of related party transactions that are large and merit closer attention. What is Staging Finance entity given it is responsible for RMB 1bn of related party transactions?

All of the so-called “related party transactions” took place with companies that are equity investors/ investees of JD (Tuniu, Bitauto, Tencent, Staging Finance). Nothing strange here.


  1. Some Insiders Have Been Net Sellers In Last 12 Months. Share Buyback Not Yet Conducted. Instead, JD Has Raised $2bn In the Last 5 Months

The claim that insiders have been selling stock is simply bogus. Fortune Rising is an employee Incentive Pool where shares are periodically sold for employee awards. Richard Liu has not sold a share since the IPO and owns close to 20% of the company. The buyback hasn’t been conducted but I expect them to be buying back shares very soon, especially with the stock trading at these levels.

It also appears that the author taken the recent financings out of context by suggesting that JD has conducted $2bn of straight dilutive equity raises when in reality $1bn was a series A funding round raised for the JD Finance unit at a $7.2bn post-money valuation (worth a whopping quarter of the current JD market-cap today) and another $1bn was a recent bond issuance at very attractive terms – likely done to increase the company’s firepower to repurchase undervalued shares. The author seems to believe that JD is doomed in a perpetual state of capital raising when in reality the balance sheet is very strong, the core business is already FCF positive and its margins are on the verge of a major inflection point. Perhaps most importantly, JD Finance will now also be independently financed. Overall, I believe the risk is to the upside with respect to a greater chance that JD will be buying back their own stock at these levels vs. diluting the equity; count this as another potential near-term catalyst to my own thesis.


  1. Other additional points

The author’s subsequent analysis regarding 1. JD’s slowing GMV (which is natural given the law of large numbers and it’s impossible to maintain 80% growth for a $70 billion retailer??), 2. Intense competition from Alibaba/Suning (JD is growing at 3-4x the speed of Suning’s overall business which is struggling to clip a double-digit growth rate and is taking market share from both Taobao and Tmall), 3) However, some investors had valued the stock once at peak market cap of $50bn due to its GMV and revenue growth. It’s nice to know that the author is competent at reading stock price histories. Seriously… is it not a surprise that a Chinese “high-growth internet stock” could have a volatile share price print over any period of time?


  1. vs.

So I think I have a right to a damn opinion here since I’ve studied both companies quite extensively. Overall, I agree with the author that a comparison between JD and Amazon is irrelevant given where each retailer is in its respective growth cycle and their different business mixes. Amazon is different from JD in many ways, particularly in its business mix (AWS + online video streaming + Prime + other moonshoots vs. JD’s end-to-end fulfilment network + JD Finance + O2O), but similar in many ways also, namely the corporate culture, extreme emphasis on the customer experience and long-term focus on creating shareholder value.

The author seems to be suggesting that the market consensus is long JD to get exposure to the “Amazon 2.0” theme in China which admittedly could reflect the thinking of some of the longs in the stock today. Where I disagree is that 1) JD is an extremely well-managed business, (like many of the large Chinese internet companies are) – likely better managed than Amazon – look at what they did to Amazon in China for example 2) JD is a much more attractive investment compared to Amazon today.

Consider that Amazon is currently trading around 1.5x 2016 GMV (or slightly less than 1 times if you assume AWS is worth $150 billion), and that JD is trading at less than 0.4x 2016 net GMV (less than 0.3x when JD Finance is excluded), it’s obvious to me that JD is being valued at a massive discount relative to Amazon despite growing GMV at more than twice the rate that Amazon is and operating an extremely attractive market for e-tailing.

Also the author’s negative view of JD’s massive employee growth and his analysis of the cost structure is simply misplaced. Other than Baidu’s growing self-owned O2O unit Baidu and Tencent operate a much more asset-light business model than JD and hence do not have a large growing blue-collar labor force and are not comparable benchmarks. So JD’s employee base has grown a lot over the years – cool – has it ever come across to the author that JD’s fulfilled order volume has grown from ~194mm in 2012 to ~1,263mm units in 2015 – a more than 6x increase and much faster than total Chinese parcel volume growth? I guess the author assumes that parcels deliver themselves magically in China. Has the author mentioned at all in any part of his write-up about JD’s growing competitive advantage in logistics by operating its self-owned fulfilment network? The author seems to be missing the whole essence of JD’s business model which is a combination of managing an online platform and end-to-end fulfillment network which is quite a powerful model in the right market and if managed by the right people. It is also interesting to note that Amazon has started to in-source parts of the logistics process as their order volumes continue to grow. (For the record I have great respect for the leadership in both companies and believe that over time both companies will become increasingly more valuable).

So the author seems to jump to a simplistic conclusion that operating an asset-heavy business is an inferior model to an asset-light one such as Alibaba and appears to be basing his conclusion on the current financial profiles of both companies. This is almost like saying that Walmart is an inferior business to eBay simply because it is a more asset-heavy business and thus has a higher cost structure, yet we all know how this big-box retailer came to dominate retailing in America. I have detailed extensively in my own paper why I think JD will be a long-term winner in China and will not go over it here again.


Final Thoughts:

What I always try to do is to have a superior understanding of the opposing argument better than the other side in order to deserve the credibility to voice an opinion. Although it is obvious that the author has spent a considerable amount of time on this short thesis (50 pages…), when it comes down to it he provides zero justification for his arbitrary $12 target price (FWIW at that price the stock trades at less than 2.5x and 3.5x EBIT and FCF on my 2018 #’s and less than 1x for both metrics by 2020).

If I were to construct a robust bear thesis on JD I would try to prove that it is a massive fraud since it’s the only scenario in my mind where the equity is worth near 0. However, the author has provided no “smoking gun”, nor has he approached with his analysis from this angle – the author claims that he has conducted channel checks, but where is the evidence of empty warehouses or fake suppliers or merchants? Frauds are also typically highly promotional companies that are overvalued on fake reported numbers alone which I argue that JD is neither.

Instead, the author primarily relies on speculative near-term catalysts (that are unlikely to materialize in my view) for his thesis over sound fundamental analysis (doesn’t seem like there is any margin of safety at all in his trade). With nearly $3bn of net cash sitting on the balance sheet, an improving margin profile over the next few years, and the JD finance unit being independently siloed – I see a future equity dilution as a low probability. Over time, the company’s earning power will become increasingly evident to the market – this is a long-term thesis that requires time to play out after all (if the stock was obviously undervalued to the casual market observer, I would be less confident that it was wildly mis-priced).

The only thing in the report that the author brought up as a legitimate concern to me is the CFO Sidney’s background at Longtop Financial Technologies – Longtop turned out to be a major accounting scandal that was exposed in 2011. Sidney worked briefly at Longtop when the company was still private as the CFO from July 2005 until March 2006 (around 8 – 9 months). From my understanding Longtop was engaged in fraudulent accounting since 2004 until it went bust. I believe Sidney was unknowingly a victim of fraud at Longtop (keep in mind he was there years before the company went public and he was not named in the lawsuits that ensued after Longtop’s collapse in 2011). The subsequent company he joined thereafter was acquired by Blackstone. I will definitely circle back to this issue at a later date and appreciate the author for highlighting this. For now, Sidneys’ position as JD’s CFO implies that Liu has basically endorsed him. My own impression of Sidney is that he is very straight-forward in his communication, is very thoughtful about the business and does not overly promote the business like what the author has implied throughout his report. Richard Liu himself is not promotional at all but it seems like the author has a different opinion. If he author can provide a sound argument that Liu is unethical and thus the equity is uninvestable then that would be a much more robust short thesis.

Other than that the author’s work here was extremely biased (the analysis was even misleading in my view), omitted major facts pertinent to his argument and failed to explore any element of the bull case (there was no acknowledgement at all of the real possibility that JD could be a multi-bagger investment over the next several years). The paper overall was very short on any real fundamental analysis, with little thought about the future economics of the business, and lacked any rigorous analysis on JD’s key business drivers. I’m also sure that Mr. Munger would be absolutely delighted if he found out that his picture was included in this paper…

In sum shorting JD stock presents an atrocious risk/reward profile since the equity can double or triple overnight and remain demonstrably undervalued (that’s what I would call a large margin of safety).


[1] This is a bit of a side note but to clarify in my paper I defined 1P net GMV as 1P Gross GMV net of returns/cancellations and the VAT, and 3P GMV as 3P Gross GMV net of only returns/cancellations (since JD takes a commission out of the entire 3P-related transaction where the VAT is not stripped out of that calculation). Originally I took a more conservative route and stripped out the VAT from the 3P-related transaction when calculating 3P net GMV.


Idea Tracker Update, Part 4: Interactive Brokers

I’ve slowed down the updates since it’s my favorite time of year in Toronto – when it gets warmer.

Interactive Brokers reported their q1 results today and I believe the brokerage business remains on track to generate just under $1B in pre-tax earnings power by 2016.

ibkr valuation

They say bulls make money, bears make money, and pigs get slaughtered.

I guess I’m a bull on Interactive Brokers…


SoftbankA great capital allocator and sum of the parts

The long thesis here is basically that Softbank is run by an excellent capital allocator and is trading at a deep discount to its sum-of-the-parts or break-up value. I largely agree, however I want to have greater clarity on how Mr. Son is going to unlock shareholder value over the near-term. Any eminent catalysts over the next 2-3 years would be great for closing the holdco discount gap and consequently increasing the IRR on this investment. But I also wouldn’t worry too much if the stock stays flat in a trading range for a little while. Given Son’s exceptional track record, (I believe Softbank has been able to compound intrinsic value per share at 20%+ since it went public) I think it’s a very high probability that Softbank will continue to compound intrinsic value per share at superior rate of return for a very long time. If we analyze all of Softbank’s major investments since inception, the compounded rate of return has been roughly in the mid 40’s%, driven largely by the investment in Alibaba; Son invested $20MM in Alibaba when the business will still being run out of Jack Ma’s apartment. Talk about a visionary. Even if we exclude Alibaba from the return figures, I believe his track record is still in the low 20’s%. I think how Son runs Softbank is actually quite similar to how Buffet grew Berkshire – intelligently investing all the cash flow from existing cash cow businesses into opportunistic, very long-term investments. I guess for Buffett it was more like investing the low-cost float received upfront from his insurance businesses. For Son his Japanese telecom business is basically the cash cow. I really like how Son is building Softbank’s ecosystem to prosper in the “mobile revolution”. He’s investing in the infrastructure (Sprint), the content via mobile games (Gungho, Supercell), and in the services (Alibaba).

Aside from the typical complex holdco discount, I believe that investors’ worries over the outlook for Sprint and the negative short-term sentiment for Alibaba is why Softbank’s shares are still trading at such a large discount to intrinsic value. But for the enterprising, opportunistic value investor that’s willing to dig through all the moving parts, I believe Softbank is quite cheap.

Domestic Telecom Business: The Japanese wireless industry is basically a 3-player market and although Softbank continues gain incremental market share, it remains third behind its competitors DOCOMO and KDDI. Apparently Son was able to convince Steve Jobs to give him exclusivity in bundling the original iPhone models with wireless plans when they were first released. Along with his foresight in betting on the iPhone, Softbank was able to undercut industry pricing by a large margin which led to huge market share gains. Some analysts were expecting Apple’s iPhone’s handset margins to come down as competitors would eventually catch up and pressure pricing in an increasingly commoditized product; if this were to occur then the telcos would be able to pass on some of these device cost savings to consumers. Despite the increasing competition, the iPhone product lifecycle doesn’t seem to be ending anytime soon and pricing hasn’t been pressured at all. I’m not sure if Apple’s ecosystem will allow them to earn above-average profits for the long-term but what I do know is that Softbank’s customers continue to buy lots of expensive iPhones among all the alternatives. So overall the business remains healthy today. They’re currently #1 in terms of networks speeds and they’ve recently completed a major CapEx upgrade cycle to 4G, so I expect their free cash flow to materially ramp-up over the next several years. Moreover, the industry players have each recently shown signs of ending aggressive promotional pricing, which should further support healthy long-term ARPU growth, or at the very least support stable pricing. I value their wireless business in-line with comps at 6x EBITDA.

Sprint: Sprint definitely has its problems. The US wireless industry remains an extremely competitive 4-player market as T-Mobile and Sprint continue to invest large amounts of capital on upgrading their networks and hurting the overall industry profitability with their deep pricing cuts. I’m not an expert on the US wireless industry so I don’t think I have a huge edge here. Some analysts even believe that Sprint’s equity could be worth $0. Fair enough. Let’s assume Sprint’s equity is worth 0. Worst case scenario, Softbank would inject a little bit more equity capital in a cash-hemorrhaging Sprint. Even under this scenario, my break-up analysis suggests that Softbank’s shares are still worth nearly 15,000 YEN 2 years from now. So whether or not Sprint is successful in reaching minimum efficient scale doesn’t make or break the thesis.

Alibaba: I believe Alibaba’s value is worth much more than the current market cap of Softbank today. I value this business at 25x 2017E EBITDA of $12B, or $300B. As they continue to scale their marketplaces, their operating margins should continue to ramp-up, along with an increasing incremental return on capital. With such a large fixed-cost base, I don’t see how their incremental margins on transaction growth won’t hold up despite the monetization concerns. I also think the fake product concern is short-term in nature. On top, they have a huge payments processing arm in AliPay, an infrastructure-as-a-service (IAAS) business that sells excess cloud capacity, and a logistics joint-venture; this is a monster of a business and likely one of the highest-quality in the world. These guys were basically a start-up that drove eBay’s business out of China. Their end market growth runway is huge, they’re in a net-cash position, and for a monopolistic, asset-light, toll-booth business on China’s growing e-commerce market, I think $120 per share is reasonable. Since Mr. Son is also the largest shareholder of Softbank, I doubt that he would want to incur a large tax hit should he decide to monetize the Alibaba stake with the huge tax basis. I assume, like Yahoo!, that their stake will eventually be spun-off in a tax-efficient transaction, so I apply no after-tax value. I also think the domestic regulatory or currency risk is quite limited. Longer-term I think as the market better appreciates Alibaba’s value it will be a driving force in lifting Softbank’s shares.

softbank nav

If Softbank shares can close 90% of the holdco discount in 2 years, the IRR on this investment will be ~38%. Like most other major non-US currencies, the Yen has been getting annihilated. In terms of structuring this position, I think I’m going to fully hedge out the Yen, and I might even short the proportional Sprint stub. Based on the proportional values of each piece, you can create your own stub if you like.

Idea Tracker Update, Part 1: EBAY

I’ve been looking for some garbage to short in this market. I think it’s been increasingly difficult to find attractive long opportunities in larger companies that meet my hurdle rate. The only real opportunities I see right now are special situations where the underlying company is undergoing some type of corporate event and coincidentally these are in industries I think I understand best. I don’t think the market is overly expensive nor cheap by the way. Over the next little while I’m going to post a mini marathon set of posts that will update some of the ideas recently mentioned on this blog.

First up is eBay.


eBay – Revisiting the spin-off thesis

The stock has done quite well over the short-term since Steven recommended it. I original liked the idea as well, but now that I have a much higher hurdle rate, it simply doesn’t meet my return requirements unless I structure a long position in combination with more leveraged instruments such as derivatives which I’m unwilling to do at these trading levels.

The original thesis was focussed on the pending spin-off of the PayPal unit from the Marketplaces unit which would unlock substantial shareholder value mainly through improved capital allocation, better corporate governance, operational focus, cost rationalization and in general just getting rid of a rather idiotic CEO. I think the CFO needs to go as well.

I think the opportunity still exists partly because of everyone’s frustrations with the mediocre operating performance and the subpar capital allocation.  But I also think investors are increasingly fixated on faster-growth opportunities in the general tech space. Ebay is sort of becoming a “dinosaur” compared to some of the more newly emerging, sexy, rapidly growing technology-based industries such as enterprise SaaS and social media. Moreover, from an event-driven angle there was also little guidance given from management (at the time of the spin-off announcement) for how each business would get capitalized, the complexity of where to allocate the corporate overhead, D&A, etc and the typical failure of most market participants in properly valuing 2 divergent earnings streams.

The thesis for the marketplaces segment is based primarily on financial engineering and returning capital to shareholders via leveraged share buybacks. I had modelled ~$4.1B in EBITDA for marketplaces by 2017, which is actually on the back of low single-digits top-line growth and ~300 bps in EBIT margin compression. I’m actually quite bearish relative to consensus on the long-term prospects of the marketplaces business given what I think is a few emerging structural challenges that this business will face in a changing e-commerce competitive landscape. I think the two large concerns I have is the slowing organic growth and the additional marketing and R&D spend that I think will need to be incurred in order to even sustain positive long-term revenue growth rates. The recent security failure might have also turned off a lot of users and damaged brand equity for the long-term; I know I would be pissed if my account and password got hacked. 1) In the recent full conference call, management projected 0%-5% fx neutral top line growth for marketplaces which I think is pretty ridiculous in light of the fact that US and global e-commerce is still growing at around high-single to low double-digits. This tells me that management have mismanaged the business to such an extent that they’re now losing massive market share. Given the international exposure, I think there’s even a possibility for reported revenue growth to go negative, especially if the US dollar continues to surge against major international markets. Given the operating leverage and rather concentrated US cost base of this segment, the downside won’t look pretty. For a typical dominant network, winner-take-all marketplaces business model that should be taking incremental market share this is just a very worrisome possibility. But honestly, after going over some of management’s bullish forecasts in previous analyst days (that have badly missed or have been revised down), 0%-5% could potentially be very realistic. On the margin side I think larger incremental spending will be required since this business increasingly looks outdated (doesn’t seem to be attracting millennials) and irrelevant.

So I feel like there’s quite a bit of execution risk going forward to revitalize this business and better compete against Amazon. I think going forward the only real value proposition for this platform will be for existing “power sellers” that have a long merchant track record and have less of an incentive to move their business. The major business lesson here: Despite having a business with superior economics on the surface: an asset-light, cash-generative toll-booth business (with ~40% operating margins) vs. Amazon’s fulfilment model (with ~1% operating margins), a dominant network effect doesn’t automatically off-set an inferior service. In Tony Hsieh’s book, Delivering Happiness, he talks about the great lengths Zappos went to continually improve their customer service even if it sacrificed near-term profitability; this included things like free shipping for customers, free returns, and an exceptionally trained, passionate customer service team. Zappos was eventually acquired by Amazon and I feel like even in this new information age economy, the level of customer service remains a critical factor to the success of these e-commerce service platforms.

I assume the new marketplaces management will increase the pace of buybacks and leverage up to 2.5x net debt to EBITDA by 2017. Working from EBITDA to FCF per share, I modeled out nearly $2B in fully-capex’d FCF by 2017. If marketplace can raise debt over the next 3 years at an average after-tax cost of 3.5%-4%, leverage up to 2.5x by 2017, and assuming they pay full repatriation taxes on existing foreign cash, I think they can repurchase ~22% of the total current float over the next 3 years. They should be over a comfortable interest coverage ratio of over 5x as well. Originally I expected closer to 30% of the float to be repurchased but since they’ve announced that they’re capitalizing PayPal with $5B in cash, I think this is less likely. Given the number of large actavist funds in this name (Jana, Third Point, Glenview, Icahn), and the leveraged share buybacks that have already taken place over the past year (~$4.6B in share buybacks) I’m rather confident that they will shrink the equity accretively in this low rate environment – I just don’t know to what degree (2.5x is what I’m assuming or 4x that some have suggested). For a still rather high-margin, cash generative business with a network effect moat, (although structurally challenged I believe) I value marketplaces at a range of multiples of 14x-16x (or a ~6%-7% fcf yield) my estimate for $2 per share by 2017. On these numbers, marketplaces is basically worth $28-$32 per share 3 years from now. Unlike most of the Street that values marketplaces on an EBITDA multiple, I believe we have to take into account a more efficient capital structure that involves leveraged share buybacks.

However, if top-line growth rates dip to negative territory and FCF growth decelerates even faster, the entire leveraged buyback thesis falls apart. On the upside case, it’s very possible that Alibaba will eventually acquire eBay and fix its problems. I think Jack Ma has real ambitions to expand outside of China. I would just assign a 0 value on the GSI segment (which is now getting sold) since it’s immaterial and was a horrible acquisition. On a side note, Donahoe has repeatedly said that it was a great acquisition – I wonder why the sudden change of heart?


For the PayPal unit, here’s an excerpt from my last letter to my investors:

“… Although we acknowledge that Apple Pay can potentially become a major contender in the payments industry, we remain skeptical of a wide spread merchant and consumer adoption of Near Field Communication (NFC) technologies, along with the issues of convenience and security being adequately addressed. Apple Pay is currently only in ~220,000 merchants in the US, which is ~2% of the total addressable market. Furthermore, we believe that their competitive focus will remain in the offline, mobile Point-of-Sales (PoS) space which is currently a miniscule part of PayPal’s existing business. In fact many market players, including PayPal, have failed to gain much market traction in the mobile offline payments industry over the past several years. This is partly due to the high barriers of entry including scale and security requirements, but more challengingly the need to get the different parties with differing economic incentives in the payments ecosystem including customers, merchants, financial regulators, banks, and the payment networks to facilitate wider spread adoption of PoS mobile payments.  If Apple Pay can indeed generate considerable traction and wider spread adoption of PoS mobile payments[1], we actually believe that this will ironically be a net positive for an independent PayPal; it can it open up a large offline payments opportunity as a free option, currently not being priced into shares, in our view. As per the unit economics, Apple Pay currently takes $0.15 for an average $100 PoS transaction versus PayPal’s ~$1 – a stark differential. Although PayPal may have to sacrifice some existing unit economics in order to create substantial value in PoS mobile payments, we believe given PayPal’s strong brand equity and franchise value, PayPal can eventually generate sufficient scale to participate in this opportunity.


PayPal’s ~$216 billion in Total Payments Volume (TPV) is currently ~20% of the $1 trillion online commerce total volume – second to only AliPay – and only ~2% of $10 trillion in the total commerce opportunity. The online eCommerce market is a long-term secular growth market and should continue to grow faster than brick-and-mortar TPV. We believe that post-spin, independent PayPal will have ample strategic options to aggressively pursue the total eCommerce opportunity – both online and offline. PayPal can position itself as either platform agnostic like it has for the most part historically, or it could potentially form strategic partnerships with either Apple, Google, Facebook, Samsung or even Amazon, among other large market players once it is separate from core eBay. We believe there was likely internal conflict and misaligned incentives between past PayPal executives and the C-suite, leading to poor strategic decisions for PayPal’s positioning in the marketplace. An independent PayPal will likely become a more nimble, focussed, disruptive “Silicon Valley-like” entity, with future business performance properly aligned between management and employees. The bottom line is that we believe spinning-off PayPal is the right strategic decision and will better position PayPal to tackle a rapidly changing global payments competitive landscape.”


In sum, I think an independent PayPal makes a lot of sense. In order to attract some of the brightest talent needed in Silicon Valley today, I think an independent stock currency is required for stock-based compensation. Base Case: PayPal earns nearly $3B in operating income by 2017. I assume a bit of margin compression (~200 bps) if they go after the offline payments opportunity and continue to deploy large amounts of R&D spending. So net-net the lower incremental margins from the offline opportunity should outweigh their existing scale benefits. I also think there’s a bit of risk from PayPal’s projected ~20% top-line growth rate – if we assume eBay’s top-line is slowing down to LSD-MSD, then by default PayPal’s related transactions on the eBay platform (around one third of PayPal’s business) should be slowing down dramatically as well.

On a 25x-30x net operating profit after-tax (NOPAT) multiple, PayPal is basically worth $50-$60 per share 3 years from now, or nearly double eBay’s value. I’m not quite sure what management plans to do with the $6.3B in incremental fcf that they’ll generate over the interim years so I very conservatively assume that it just builds up on their balance sheet. There’s definitely a bit of upside if the actavists push new PayPal management for better capital allocation as well. PayPal will be capitalized with no debt and $5B in cash to help fund the credit part of their business but I think realistically they’ll just need half of that amount. PayPal’s $9B in excess cash by 2017 will be basically worth ~$7.5 per share.

So let’s add all the pieces up to arrive at intrinsic value: 1) eBay marketplaces: $28-$32, 2) PayPal: $50-$60, 3) PayPal’s excess cash: $7.5, and 4) a 9x multiple on corporate overhead’s negative after-tax income: ($13.5). If we add it all up the stock is worth $72 – $86 by 2017. If we take the mid-point of the intrinsic value range and if the post-spin catalysts materialize as expected, the stock should generate a near 18% 2-year IRR under my base case from today’s $57 price tag. Unfortunately this is way below my hurdle rate, so I’m officially closing this position for the blog. How I’m going to play this situation out is to observe the stock up to the spin-off event and see if I can get more clarity on post-spin capital allocation and if there’s any post-spin forced selling. I’m not long now, but at least I’ve done the necessary work should Mr. Market present an opportunity. I’m also much more interested in the PayPal piece. For bigger funds that have a lower hurdle rate, the opportunity might be worth it if you’re confident that capital will be adequately allocated in both pieces and if you think marketplaces isn’t in a long-term structural decline.

One last note on management that I want to get out of my system. These guys (management) are honestly amateurs when it comes to capital allocation and corporate governance. If I were to grade them alongside the management teams that run my portfolio companies on a relative scale from 1-10, these guys would be sitting on a negative digit. I have absolutely zero confidence in existing management to execute on what is necessary to build long-term shareholder value. I recall reading a transcript with the CFO basically saying in like a proud way that the ~$15B of cash that they’re holding on their balance sheet is generating a lot of interest income. Well, no shit. If you give me $15B I can sit on my ass and buy long-term US treasuries too! These guys seriously need to go. The high executive turnover, constant flip-flopping on major strategic issues and operational missteps is suffice evidence. Next time don’t hire a management consultant to run a once leading, global technology company.

[1] NFC only works on the newest iPhone 6 models, which is ~70 million users as of the end of 2014; Apple has large existing installed base of ~500 million iPhones

DirecTV Update & The Future of the Pay-TV Industry

It has been a while since I have updated my thesis on DirecTV. The stock has had a good run since my write-up as the market has re-rated the shares to a more reasonable multiple. The Brooklyn Investor, a high-quality investing blog I follow also recently wrote-up DTV here for those of you who are interested. What caught my eye was the Brooklyn Investor’s point surrounding the cost of distributing content in the current Pay-TV ecosystem.
In all honesty, I had little to no clue how the Pay-TV ecosystem would look like 10 years into the future when I first bought my shares in DirecTV at $50. But did I really need to have a crystal ball to justify buying into a good business that earned returns on invested capital greater than 20%, had high incremental returns on capital and growth prospects in Latin America, was cannibalizing its own shares at good valuations, and was run by a fantastic management team? Oh yeah, it was also selling at around a 10x multiple! Maybe I didn’t know it back then, but perhaps the real reason I should have used to justify buying DirecTV was because of its long-term cost advantage.


Satellite is an attractive business because it’s actually relatively asset-light compared to Cable and Telco fiber. A mini-dish and set-top box for every home, several broadcast stations, call centers and a few strategically positioned satellites is all that is needed to run a ubiquitous satellite TV company. This, vs. a ton coaxial fiber or fiber optic needed for every home or business, that needs constant upgrading every several years in order to remain competitive, and you can see why cable and Telco fiber are much more capital-intensive businesses and why media moguls Rupert Murdoch, John Malone and Charlie Ergen were all bidding for Satellite orbital slots when they first came to auction. They all wanted to own a piece of the new death star which was Direct Broadcast Satellite. As an aside, one relative advantage cable has over telco fiber is that although DOCSIS technology continues to get upgraded for faster bandwidth speeds, the vast majority of the cable industry’s CapEx is front-loaded, so only the “last mile” to the home has to be upgraded, vs. Telcos who have to replace their entire copper-wire infrastructure. This is also why Liberty Global holds a competitive advantage over Telcos in Europe.


So back to DirecTV. Having a cost advantage is enormous in a scale-driven business. Direct Broadcast Satellite technology provides the means to efficiently broadcast several hundred HD channels, bargain for exclusive content such as the NFL Sunday Night Ticket and generate superior unit economics (IRR) relative to Cable/Telco; DirecTV’s subscriber churn is the lowest amongst US peers, its ARPU is the highest, and despite paying higher Subscriber Acquisition Costs (SAC) and programming costs per sub than its Cable and Telco peers, DTV ends up ahead because it has a less capital-intensive business model that delivers superior content to a stickier customer base. Looking back, this should have been the original rationale for buying DirecTV. And I believe Satellite TV’s cost advantage will remain even 10 years from now. Let me explain.


Nearly everyone is worried about internet-TV disrupting the Pay-TV ecosystem and everybody and their mother “cutting the cord”. I suspect when some of these major networks such as ABC, FOX and CBC start selling highly-rated channels directly to consumers, the economics would be questionable. Let’s not forget that the Cable networks will most likely have to charge more on a per-channel basis online in order to make up for their lost advertising revenues that they would otherwise receive in the current ecosystem. So I believe either online advertising rates will have to go up in order to justify a large-scale move into a direct-to-consumer model, or even a wholesale model to broadband providers, but either way, the Cable networks remain incentivized to maintain the current ecosystem and will not want to see their affiliate fees evaporate overnight.


Another phenomenon that I believe most industry pundits and analysts have failed to recognize is that the cable firms have basically been indirectly subsidizing the bandwidth costs of OTC players such as Netflix, Hulu, etc. If you take the view that we are going to fully transition towards an Internet-TV based world, then that would mean that currently Cable TV service revenues are subsidizing their broadband service, and broadband service in turn, I believe, is currently subsidizing online streaming services. Quite simply, broadband providers have yet to fully monetize their broadband service. A lot of people think that US broadband services are a rip off because of the speeds you get relative to the price you pay. But what about the other important component of a broadband connection – bandwidth capacity? As more and more video is consumed online, bandwidth demand is going to explode upward and I don’t think Comcast and company are going to idly sit back and let online streaming providers “free ride” the system. After all, Netflix’s largest operating costs after content are probably their server expenses. I would not be surprised at all if the Cable/Telcos eventually implement usage-based pricing for their broadband service in order to make up for the eventual lost cable TV revenues. Let’s not forget that Americans on average still watch a shit ton of television per day, around 4-5 hours on average I believe. Given the explosion in growth in smartphones and tablets, I think its safe to say that online video consumption will continue to grow rapidly, and someone, probably the OTC guys are going to have to eventually pay that price.


One last point I want to make regarding the Internet-TV threat: I believe the internet-TV industry will remain quite fragmented in terms of content type for some time, and it would be rather difficult for any one player to become dominant across all content categories, yes, even you, Netflix. This is partly due to the nature of the industry, but also because online rights will remain competitive. Online streaming services don’t have the luxury of scale to buy up every single content category like Satellite/Cable/Telcos do. This is not only due to a lack of scale but a tradition of content producers and distributors securing long-term, content deals. Netflix may be wise to strategically pursue exclusive original content series such as House of Cards to help differentiate their platform, but will they be able to expand into Live News, Sports, or even killer nature shows? I think not. Exclusive content comes at a high price, especially if they are online rights. To me Netflix looks like a cheaper version of HBO, and even Reed Hastings admits that HBO is Netflix’s largest long-term competitor. In fact the only reason not everyone can watch HBO online without a Pay-TV subscription is because Time Warner doesn’t want to piss off the the DBS/Cable/Telcos. Platforms such like HBO, Starz and Netflix tend to focus on theatricals and original series, so they won’t appeal to every type of customer, and especially those that want a very broad range of content.


Eventually everybody will have an online video streaming service, including DTV. In fact DirecTV is already starting to complement its existing satellite TV service with an online-video product. As more customers demand “TV-everywhere” and place-shifting television, TV distributors will focus on building their online product. That’s why I think DirecTV tried to buy Hulu earlier last year. And with DTV’s large existing customer base and strong relationships with content providers, they should have no problem scaling a solid online video service. Who knows how it would actually look like, but I wouldn’t be surprised if it was sports focused.


So to sum it all up:

1) Despite all these fears of the death of the Pay-TV industry, the market has basically misunderstood the DirecTV story for a while now, and using some second-level thinking, it’s safe to conclude that DTV’s business will not get fucked overnight.

2) Online streaming providers such as Netflix are basically selling their services at an artificially low price, but they can’t have their cake and eat it too. Eventually the economics of online streaming will not be as attractive.

3) The US Pay-TV industry is very slow to change, but even if the bundle breaks DTV can successfully transition to an Internet-TV world. Some of the best investment opportunities are created by uncertainty and taking a contrarian view. As long as one is able to price in the risks and come up with a range of possible future outcomes, one can make money.

4) DirecTV is led by CEO Michael White, who was hand picked by John Malone. This guy knows what he’s doing, and is a very savvy and disciplined capital allocator. Eventually they may spin-off the Latin American business when the time is right which would help unlock shareholder value.

5) If all else fails DTV’s backdoor play is to merge with DISH, and maybe even AT&T will buy them both eventually :). Although I’m not betting on it, it does provide some downside protection.

6) By the way, John Malone actually holds a lot DirecTV shares, and someone may want to confirm this, but last time I checked he owned ~3% of DirecTV’s total shares outstanding. Malone was former chairman of DirecTV, and he was basically forced to step down by the FCC as he had competing cable assets in Puerto Rico. Let’s also not forget that the Berkshire guys own the largest stake in the company, and oh yeah, they are pretty damn good investment managers.

7) I am an idiot for selling a bunch of shares when DTV was in the low 60’s. I didn’t maintain conviction when I should have. Sometimes that happens when you maintain a very overweight position (it was more than 30% of my portfolio at one point). Mistakes come with learning to become a better investor, and I have a long way to go. At least I put the money back into Liberty Interactive :).

For those of you who read sell-side research, I would recommend following Jason Bazinet @ Citigroup, Matthew Harrigan @ Wunderlich Securities and Craig Moffet. They seem to know what they’re talking about in the Cable and Satellite industry.


For now I am holding on to my shares, and have modelled out my IRR assumptions for DTV below. Click on the DirecTV icon below and you will see my income statement and key driver assumptions. A 12% IRR at a modest 14x multiple is nothing crazy, but will beat 99% of institutional investors out there, including hedge funds, mutual funds, pensions, and endowments.

My DirecTV Income Statement and Operating Model – You may have to zoom in:
       DirecTV Model

If you like reading our blog then please follow us on Twitter  @valueventure101 to get updates on our latest investing ideas!


The author is this post owns shares in the company mentioned (DTV) and may purchase or sell shares without notice. This post represents only the author’s personal opinions and is not a recommendation to buy or sell a security. No information presented in the post is designed to be timely and accurate and should be used only for informational purposes. Readers of the post should perform their own due diligence before making investment decisions.


LINTA finally uploaded their Investor Day slides:

Other than the proposed “spin-off”, there are a few other things I find interesting:

1 The e-commerce group, which is a very small asset compared to QVC and generally unnoticed, will actually be one of the largest focused e-commerce companies in the world (slide 8) in revenue. This means as a stand-alone this thing will get a reasonable amount of sell-side attention.

2. Management refused to give a specific FCF guidance for QVC but made some comments on the variables (slide 18): 1) They expect maintenance Capex to be 200 – 230 million (I used 240 mil in my model); 2) working capital is volatile but management pointed out that WC has been on average 10% of revenue in the last five years (I used a 9.3% last three year average); 3) they factored in dividends to the Japanese minority interest as a cash outflow, which I accounted for separately as a reduction in overall enterprise value. So overall if you use these numbers that management provided and model FCF my way (accounting for the minority interest on an enterprise value basis), you end up with roughly the same number that I have, or perhaps slightly higher – about a little over $1 billion TTM.

3. QVC US is now 42% e-commerce. Management’s goal is to bring it to 50% by 2014. Mobile as of Q3 was 30% of total e-commerce sales. QVC was ranked no.3 this year in sales among all mobile services vs no.5 last year; remains no.2 among retail apps after Amazon. Management believes mobile sales will top $1 billion in 2013. 67% of new customers are from digital devices which is helping the average age of the core customer group trend lower.

4. Customer retention has been pretty consistent – a little under 90% per year for the last 20 years. Management highlighted the scalability of the QVC retail model by pointing out that purchase behavior (average no. of purchases, average annual spending) across all of QVC’s markets is nearly identical.

5. Some very interesting comments on the Chinese asset – Basically before QVC made the JV investment CNRS was for two years solely run by China National Radio who did not have a lot of retail experience. QVC’s plan with the JV is to upscale CNRS by bringing in more QVC brands. Recent product launches have already helped drive a 40% sales growth. Under the current JV model the goal is to bring QVC practices and discipline to the business to make it run more like one of QVC’s international businesses rather than a Chinese state-run business.

Top 2 players in the business (far older and more established) did $1 bil and $750 mil in revenue vs. $100 mil + for CNRS on basically the same level of home carriage as CNRS. CNRS is currently ramping sales at the same rate the Big Two were in their growth stage.The big difference is that CNRS is younger.  Imagine CNRS at $500 mil in sales maybe 4 years from now which I think they are perfectly capable of managing. At that stage it will probably do half of QVC Japan’s EBITDA and will have solid double digit growth potential vs. a market that has suffered from chronic deflation. In my model I valued CNRS at the same value as QVC Japan which is doing a little over $1 billion in sales. I think this is still appropriate given how fast the industry is growing and the potential for taking market share. QVC’s CEO seems to believe that their practices are superior to the Chinese incumbents.

6. Management would like to do one new market every 18 – 24 months; hired a new VP to lead the efforts.

7. This is interesting. QVC has been growing on average 4% in its mature markets over the last three years. The CEO’s view is that “there is no reason (they) can’t continue to grow at rates (that) at least look like ’10 to ’12 historic average”. They also expect margins to slightly improve every year as more customers switch to digital. So if you add in growth from their newer markets, and potential international expansions, assuming constant currency, this is basically implying something like a 6 – 7% annual growth rate in EBITDA. This is a little higher than I expected.

8. Malone’s comment on the new split was “you have to look at the trackers as a half-step. I think the right way to think about it. It’s got its own identity but it isn’t totally separated from the mother ship. And there are certainly tax synergies, still, in the combination. So at an appropriate time, it could be separated. But I think it has some more growing up to do as a collection of businesses.”

9. Greg Maffei seemed amused about QVC’s transition from a perceived “Internet roadkill” to today one of the biggest e-commerce businesses.

Disclosure: I am buying more LINTB.


Funny how quickly things work out sometimes. LINTA management announced today that they plan to separate QVC from LINTA’s e-commerce assets in a new tracker deal, which I guess I sorted of predicted in the write-up I posted two days ago. LINK:

The way the deal will be structured – QVC and LINTA’s 38% stake in HSN will be one part (New tickers QVCA & QVCB) ; the e-commerce portfolio will be another part (new tickers LDCA & LDCB). My understanding is that they are effectively turning two tracking stocks (LINTA & LVNTA) into three (QVCA, LDCA & LVNTA). Sounds complicated?

Basically this deal will turn QVC into a pure-play with its own SEC filings that people will actually bother to locate & read, which I think is huge news because now people can just look at QVCA and say this is basically the same thing as HSN (which is publicly traded and richly valued) but only better and should probably deserve to trade at a premium. The e-commerce assets will also be able to highlight its high growth profile and hopefully re-rate to where their peers trade at (mid to high teens x EBITDA). Also, currently LINTA disclosure lumps the financial results of all of the e-commerce sites into one line and does not do a very good job explaining these businesses; once the spin prospectus is filed there will be a dramatic improvement in transparency and quality of disclosure of these smaller assets so investors can gain a better appreciation for their business models.

Other than that, LVNTA will be spinning off its stake in Tripadvisors and will be making portfolio arrangements by making some minor asset transfers with LINTA’s e-commerce group (LDCA) to give the new vehicles more visibility. The effects of this are fairly immaterial to LINTA shareholders.

I think management is probably quite frustrated with where the market is valuing QVC and their plan is to give QVC its own stock so the multiples can be raised to where HSN is, which will make it easier for them to do a merger. LINTA also updated on its share repurchase program. They bought a little over $200 mil in the last two months and I expect them to accelerate that throughout the end of the year given that they still have $500 mil outstanding for the year but only 3 months to complete the program.

The stock was up 7% today but given the simplification of the capital structure this is gradually turning more and more into a no-brainer for a long-term investment. I will be buying more and will follow up with a spin-off model some time next week.