Thursday 7 March 2019

Is another tech bust coming?

The technology sector (referring to online services, and to a lesser extent, SaaS, which is somewhat different in the B2B space) has been one of the strongest stock market performers over the past decade. Growth rates in users and top-line have been incredible, but with a relatively small handful of exceptions (Facebook, Google, Tencent, Alibaba, Expedia & Priceline, and a few others), profitability has been vanishingly rare. However, managers with heavy technology exposure - either in the listed or unlisted space - have nevertheless made a tonne of mark-to-market/capital-raising-valuation profits, as valuations have escalated alongside galloping top line and user growth.

The received wisdom at present is that the lack of profitability reflects heightened investment in growth and customer acquisition, and that as these businesses scale, profitability will eventually mushroom upwards as they did at the likes of Facebook, an in some cases that may well prove to be the case (see my post on Spotify). However, that this will happen in general is still an article of faith, and as I hope to argue in this article, I believe there is a good chance those cases will prove to be much rarer than is currently believed, and that it is at least as likely that losses actually significantly increase in coming years, and significant valuation (and economic) carnage subsequently ensues. Indeed, I'm convinced it's possible we could see a tech bust on par with the 2000 dot.com fallout, as the two booms, and potentially the coming bust, have a lot of dynamics in common.

The dot.com bust revealed to investors that while technology can (and has) change(d) society, it doesn't change the laws of economics and business profitability, or how companies are ultimately valued. From an aggregate industry standpoint, the level of profitability will - like any industry - be ultimately determined by the balance of demand and supply, with some necessary adjustments made for the existence of genuine network effects and other sources of customer lock-in/pricing power, which I believe are likely to prove much rarer than is currently believed.

The past decade has seen a mushrooming in demand for online services, driven by the advent of the affordable smartphone (note that in this post, I'm focusing primarily on B2C, including social media, video streaming, online gaming, online news like Buzzfeed, e-commerce, and other B2C services like travel booking, ride hailing, food delivery, etc, rather than B2B SaaS, where the dynamics are somewhat different). Unlike the 1995-2007 period, where consumption of online services/media was constrained by the amount of screen-time people were able to spend in front of their PCs, the advent of the smartphone - coupled with increasingly fast and affordable wireless data networks/services - has now put a small computer screen inside the pocket of the majority of human beings on the planet.

This has been transformational on the demand side. The total demand equation is determined by the aggregate number of 'eyeball hours', coupled with people's incomes/purchasing power. While consumer purchasing power has grown relatively slowly, smartphones and wireless networks have exploded the amount of available eyeball-hours, by expanding not only the pool of connected customers to the majority of the world, but also the number of available hours people can spend interacting with online media/services.

Unlike in the pre-smartphone era, people can now consume online media/services while commuting, waiting in line at the bus stop or in a store, walking around at the mall, or sitting at their desks at work. All they have to do is pull their mini computer screens out of their pockets/bags and get online. Furthermore, the addictive nature of many online platforms (notably social media) has also kept people glued to their screens to the maximum extent possible. In combination, these factors have propelled the eyeball-hours demand-side up the steep section of the 'S' curve.

However, the whole online/mobile services eco-system is about to run into a serious problem: there is a finite global population base, and there are only 24 hours in a day, so the amount of humanity's collective eyeball-hours available to tap is also equally finite, and is rapidly becoming saturated. When this saturation point it hit, industry growth in online services/media is going to become much more zero-sum/cannibalistic, and morph into a market share fight for our finite attention.

Netflix CEO Reed Hastings recently noted that one of his biggest competitors is the hit online game Fortnite, and he is absolutely right. Hastings recognises that there are a finite number of eyeball-hours up for grabs, so any source of online media/entertainment that absorbs a bunch of eyeball-hours is necessarily a competitive threat to all other sources of online media. When industry eyeball-hour saturation is reached, the success of one platform/service is going to have to come at the expense of other platforms/services, rather than via expanding into what was previously a void of underutilised attention (e.g. instead of sitting around the airport with nothing to do waiting for a flight, or queuing at the bank, people can now interact with their phones, etc).

The past decade has been the golden era of the demand-side of the equation, which is rapidly drawing to a close, and yet, most online services companies are still not making much money (if any). Why is that? Because supply has grown as fast as demand. The exciting growth narrative has drawn a flood of capital into the industry, from the VC/privately funded tech start-up space to Softbank's Vision Fund, to now also the reinvested cashflows of the small pool of giant, highly profitable tech platforms such as Tencent, Baidu, Alibaba, etc. Baidu for instance has funded numerous loss-making businesses over the past five years that have consumed most of the cash flow its highly-profitable search engine business has been generating. iQiyi is an example, which while growing fast, has negative operating margins of nearly -50%, as it fights for market share with Tencent's loss-making streaming service.

Valuations have nevertheless mushroomed because, much like in the late-1990s dot.com bubble, investors are focusing on growth in active users and revenues, rather than profitability. Much like in the dot.com bubble, multiples of revenue or eyeballs (today, rebranded as 'active users') are being used, based on faith that as scale is acquired, profitability will inevitably follow, alongside rising economies of scale. But the laws of economics and demand & supply - not to mention the dot.com bubble precedent - would warn against making such an automatic assumption of future prosperity.

What is just as likely to occur - if not more so - is that rather than profits inflecting upwards, losses actually deepen as demand-side attention saturation beckons, and tech companies start fighting with each over a finite pool of eyeball-hours for market share, triggering a massive industry fallout and (eventual) consolidation cycle. As the demand-side moves towards the shallow, right-hand section of the S curve, while supply growth continues at its former exponential pace, the fight for users and eyeballs is only going to intensify, and aggregate industry operating losses are apt to mount.

It is telling indeed that after a decade of hyper-growth in demand, profits are still as rare as they are. It stands as a testament to how effective free markets are at stimulating supply to meet growing demand (especially in loose capital market conditions) and crushing profits, where sufficient switching costs, network effects, and other sources of customer lock-in capable of thwarting such an outcome are absent. And it raises the question - if these companies haven't been able to make money during the past golden-era decade of demand growth, what hope is there for them in an environment of demand saturation which sees aggregate industry demand radically decelerate?

One thing investors have forgotten of late - as they did in the late 1990s - is that the switching costs for consumers of (most) online services is relatively low. E-commerce has removed many of the frictions associated with time and space that exist in the real world, and this is what has enabled the platforms to grow so rapidly, and capture share off old-world industries. However, this lack of friction cuts both ways. With some important exceptions, the lack of these frictions means that it is also very easy for customers to shift to alternative platforms/apps/services. It takes merely a few minutes, for instance, to download and set up a new food delivery app.

In the 'real world', frictions exist. In traditional bricks and mortar retail, for instance, there is a real time cost for customers associated with shopping around. If you see an item you like in the store, you're aware you might be able to find it cheaper elsewhere, but you don't necessarily know where to find it, nor have the time or inclination to shop around to save $5-10. These frictions have been a source of profit opportunity for incumbents, and even today, B&M retail is vastly more profitable than e-commerce. However, such frictions don't exist in the online world. It is easy for customers to compare prices online and simply choose the cheapest option. One might use Uber for years, but change to Lyft or some other service as soon as a friend tells you, "use XYZ, it's cheaper".

In addition, customer acquisition is more commoditised in the online space. Anyone can pay Google for traffic, or pay Facebook for ads on a click-through basis. There is less benefit to incumbency, because customer acquisition/marketing costs become variable, on a per-customer basis. This means a lot of the potential profit pool is competed away in high customer acquisition costs - if you can make $5 per customer on a mobile game, it makes sense to pay not only $1 to Facebook per app download, but also $2, $3, $4, and even $4.99. The profit pool can therefore be easily competed away in the form of escalating CACs as the supply of available apps/services continues to grow.

These are the real reasons why profits are so rare, despite very rapid growth: it's not a lack of scale; it's because these are lousy businesses, with limited differentiation or pricing power - what would traditionally be described as 'commodity businesses'. I don't buy the argument, for instance, that JD.com, at US$56bn in revenues, still has not achieved economic scale (JD still loses money). Nonsense. US$56bn already makes it one of the biggest businesses in the world. The truth is, it doesn't make any money because it's a lousy business - if it was a good business it would already be making a lot of money. There is a lot of competition, and limited differentiation, with little to no frictions with respect to price comparisons. When purchasing goods online, customers show limited brand loyalty and just choose the cheapest option, just like they do with airline tickets, which is why airlines have been lousy investments over the years. This is why it has only been possible to grow and acquire 'scale'/customers/volumes by offering prices so low as to give away all your profits. As soon as you try to raise prices, customers go elsewhere.

As bad as the status quo currently is for many tech companies, however, it could well be about to get much worse, because the level of competitive intensity is likely to radically intensify from here, driven by three factors: (1) increasing eyeball-hour saturation on the demand side, which results in demand transitioning from the steep section of the S-curve, towards the far right plateau; (2) continuing exponential growth in supply as VC funds continue to poor into tech start-ups; Softbank's Vision Fund (attempts to) raise as much as US$500bn in seed capital; and a record level of tech IPOs are expected to occur this year; and (3) old economy companies/industries also race to get online.

The latter is also a very significant and under-appreciated development. As the tech industry has rapidly grown, taken market share off traditional businesses, and garnered an increasing amount of media attention, old world companies have started to rush to digitise their businesses and transition their own sales channels online as well. Traditional retail, for instance, is investing heavily into 'omni-channel' capability; traditional media companies are investing in their own streaming services; and a plethora of other businesses are also setting up online storefronts and/or payment systems.

This trend is now fairly widespread. I was floored to see last week, for instance, that Gazprom Neft - the oil subsidiary of the Russian gas giant Gazprom - which is regarded by investors (somewhat unfairly) to be one of the most inefficient and corrupt state-backed oil companies on the planet, has launched an app that allows its customers - when refueling at one of Gazprom Neft's stations - to pay for their gas with their smartphones, without leaving their car.

Now let's think about that for a second: Even Gazprom is adopting digital payments for its customers. Ask yourself this - how hard is it, really, to launch all these online services/platforms/payment systems, if even a state-owned oil company in a developing country can do so? And if everyone can do it (and is doing it), the replacement cost is obviously relatively low (just get a few smart young coders in to write the software), and why do investors think such easily-replicable services/software are so immensely valuable?

Now, granted, there are some very good tech businesses out there, which enjoy genuine network effects and other sources of customer/supplier lock in that will render them relatively immune to these mundane competitive forces of demand and supply. But their number might be fewer than investors realise. Take the online travel booking platforms, for instance (e.g. Priceline, Expedia, Webjet), which have hitherto been very profitable - partly due to the frictions associated with getting hotel inventory (which has to be done one hotel at a time). Ctrip in China (which has about 50% market share in China) was making 30% operating margins as recently as 2011, but profits have since collapsed to less than a third of that amount, as a market share shit fight broke out, which resulted in heavy discounting and industry losses. Supply grew faster than demand and profitability collapsed.

The market has consolidated down some in the past few years, but Ctrip's profitability still remains depressed. It is possible it recovers (and the stock is certainly priced that way), but online giants such as Alibaba are competing very aggressively in this space, and have a structural advantage in terms of customer acquisition costs. China's airlines are also launching their own online booking services and are attempting to disintermediate the OTAs. So there are still limited constraints on supply, and the reality is, customers will choose the platform offering the lowest prices, and suppliers of hotel or airline seat inventory will use the platforms that offer them the best prices (and lowest take rate). It is therefore entirely possible the industry remains highly competitive and relatively unprofitable indefinitely, reflecting the fact that many aspects of these platforms are little more than computer code, that can be easily replicated by armies of smart coders.

As demand growth slows down, tech companies of all stripes - accustomed to rapid growth - are likely going to start competing with each other aggressively for market share - not least because growth has always been their marching beat, and is what their investors (and stock prices) demand. And they are going to be doing so in an environment where aggregate supply is starting to far outgrow aggregate demand. Traditional economics would suggest that it is hard to see how mushrooming losses are not an inevitable outcome, and this is exactly what happened in the dot.com boom and bust.

Markets are not likely to take these developments well. Tech valuations are extremely elevated, and bake in an expectation of both continuing rapid top line and user growth, and sharply rising profitability as scale is built. Instead, they might get slowing user growth/revenues (or even falling revenues in some instances), and mushrooming losses. The de-ratings are apt to be severe, and at some point the narrative will 'flip', and investors will start to look at these businesses in a very different light. They won't see sexy growth stories, but instead businesses going ex-growth in commodified industries, with extremely uncertain revenue and earnings outlooks.

The fact that this outcome is on the horizon is probably one reason why VC funds are currently rushing to IPO their privately-funded, loss-making investees at present (2019 is expected to be a record year for tech IPOs). There is a window to monetise these loss-making businesses on the cusp of a rapid slowdown in user/revenue growth at nosebleed valuations at present, but it is unlikely to remain open for very long, because at some point investors are going to wake up and realise the emperor has no clothes. Investors are soon going to realise companies like Dropbox (IPOed last year) are likely to remain unprofitable, and probably won't survive long term.

A tech bust could also have a negative impact on the economy - particularly the US economy. The tech boom has been positive for the economy as it has acted as a means to redistribute excess savings held by sovereign wealth funds, high net worth individuals and endowments, etc, into millennial salaries, as the capital raisings have funded mushrooming Unicorn losses which have been channeled primarily into the high incomes of 20-something coders. The largess has also found its way into high real estate prices in the Valley (amongst other things), and this stimulus could come to an abrupt end.

A tech bust could also harm the advertising giant Facebook (and also Google, to a lesser extent). How much of Facebook's mushrooming ad revenue comes from loss making tech businesses desperate to acquire new customers, who have been willing to pay through the nose for click-throughs? If ad budgets get curtailed, it might not only slow Facebook's ad growth, but even (gulp) cause it to decline, amidst a backdrop of rapidly rising costs (associated with growing platform monitoring/compliance issues). And as the eyeball-hour attention market is saturated, all future innovations in online media/games/services will have to come at the expense of the eyeball-hours currently being captured by existing incumbents. Increasingly, if people are spending more time (for e.g.) playing Fortnite, they will be spending less time on Facebook, as the attention market is set to imminently become a zero-sum game.

I could be wrong. Perhaps I'm old hat. I've sat almost the entire tech boom out, and still have zero exposure. I've certainly been wrong to date. But it looks to me like a demand/supply imbalance is brewing, and investor optimism has been carried away. Time will tell.

Comments welcome.


LT3000