Since 2009 there has been a proliferation of VC-funded unicorns in a variety of tech sectors, and the bubble has been blown on the hopes of finding the quintessential dual-sided platform which will monopolize either the search for or distribution of a certain product (or both). As public investors are learning from the recent offering of Uber however, not all “tech” companies and platforms are created equal. The economics and potential of each differ with regard to capital and reinvestment needs, the minimum effective scale of each, the ease of entry, and the strength of either supply or demand side economies of scale.
Google and Facebook are the archetypes which many of these attempt to aspire to, but it should be very clear by now that they are the exception and not the norm. These two have a truly global market and have commoditized web-based search and social media, respectively. On the point of commoditization alone, you could say Google is the better business as it faces almost no competition within web-based search as it is nearly a pure commodity, while new social media offerings have been popping up to provide new features which cannot be easily provided by Facebook. As a result, it is likely that Google will retain a larger market share of search than Facebook will in social media, as social media is much more differentiable.
Both Google and Facebook also exhibit near-zero marginal costs when adding new users or customers, so their reinvestment needs have always been low- even in the early high-growth stage these two were highly profitable and gushed out FCF. Unlike some software or pharmaceutical firms, their initial fixed costs weren’t very high. With little initial investment, these two firms could monopolize their respective industries by virtue of direct network effects by which every additional user directly increased the value of the platform to every other user.
Nearly every other platform which has come along since is inferior with regard to one or all of these factors and so will experience far lower returns on capital in maturity.
Take Uber as a quick example. Far from having near-zero marginal costs, it must retain and pay drivers in each area it expands to and so costs bloat as it grows. There may be indirect network effects by which the number of drivers increases as its user-base increases and so users experience lower waiting times, but this is a far shot from the all-to-all direct network effects of FB or GOOG. Uber’s TAM is also not global, as it discovered with its experience in China. There are also many possible substitutes for hailing an Uber, such as driving, biking, public transport, etc. So you have a platform which will likely continue to share a duopoly in some national markets, but which only caters to a segment of the population which would prefer to pay to be driven around. It also has massive variable costs which will continue to squeeze margins whenever growth drops off. There is no clear operating leverage here, and I don’t think anyone knows for sure if Uber will ever be profitable long-term. To summarize, relative to FB/GOOG, its TAM is much smaller, there are many more substitutes, it has high reinvestment needs, and the network effects are iffy.
Netflix is a better business than Uber, and I believe provides much more value, but it still faces many of the same issues.
One of the main disadvantages that NFLX shares with Uber is that it has substantial variable costs and reinvestment needs, which grow as it grows. Unlike FB/GOOG, it must first pay for all content provided on its platform to end-users, and its suppliers have substantial bargaining power. Imagine for example, how much worse off Google would be if it had to provide itself all the websites and content found through its search, or if it had to pay other websites to do so.
Amazon also has a large advantage over Netflix in this regard- although it purchases/resells inventory and has fulfillment costs, it also has many sellers which list on its platform for free and AMZN profits from their transactions. Its sellers also have very low bargaining power given that there are so many of them.
Netflix’s variable costs just mentioned make a large dent in its potential to provide network effects and dominate globally. If it were simply a platform provider, and if content producers were willing to offer their product free of charge, NFLX could take a cut by providing a TV marketplace and have unit economics resembling FB. With this hypothetical setup, it would be able to grow much faster, provide the global inventory of TV and movies to all viewers worldwide at near-zero marginal cost per user, and experience extraordinary returns on capital.
In reality, the economics of NFLX are much closer to an “old-world” firm than they are to a pure platform provider. As mentioned, NFLX must pay for its inventory of content, whether it is licensed or created in-house. To grow, it must increase this base of inventory and therefore its marginal costs are actually quite high. To say that its marginal costs are low would assume that it could grow without expanding its library of content. Secondly, its suppliers have substantial pricing power, and NFLX is often forced to increase its license costs to keep the most popular shows or they will vanish. The recent appearance of Disney+ along with NBC’s threats to remove Friends/the Office is making it clear that NFLX cannot easily replicate all of what its suppliers provide, and that they have the power to inflict substantial pain on its results and offering. There is no seller on Amazon or website on Google which could make such threats, and which could reduce the total value provided to the platform’s users so dramatically.
Further, Netflix doesn’t experience the network effects of a pure platform. The first of two arguments in favor of network effects for Netflix is that as its membership increases, its cost per user will decrease and its ability to pay for content will increase. This however, is just supply-side economies of scale that have been known of since the industrial revolution. For supply-side EOS to hold, a fixed cost is depreciated among a large number of units (members) and so unit costs are lower than they would be for a new-entrant. This seems to make sense for NFLX, although I would say that most of its base of content is not fixed, it is probably closer to variable. It outlays cash for a show license, which is amortized over 4 years and in year 4 NFLX has to decide to renew or attempt to replace it. It would be like if Nike had a few factories owned by other firms and every 4 years it was asked whether they would like to pay higher fees or lose one of its factories, with no other firms to provide the same quality factory. The licenses are short-term assets and rapidly depreciate. As a result, much of its content investment is needed to replace what will be going out the door in the future.
NFLX has of course realized that its suppliers have substantial power, and that it is beholden to licensing costs and so has tried to in-house production, which also comes with many problems. Firstly, it has to bid on ideas, producers, and actors with other networks and content producers. Secondly, reinvestment and total costs have risen by in-housing a broad TV studio which attempts to cater to the entire world, limiting its organic growth rate and ability to rapidly expand its base of content without drowning in debt. You could say that its total profit potential in maturity has grown as they are decreasing the bargaining power of their suppliers by vertically integrating upstream, but this comes at a substantial cost and this cost must come with additional profits. As a result, its potential returns on capital are unlikely to have increased much from this decision but their capital needs have exploded.
The second point thrown out for NFLX’s network effects is that as its content base increases, it becomes more valuable to users as they have more to watch and so its TV-buffet expands in variety. On its face I agree with this, but as mentioned, this comes with substantial cost. Unlike a software platform, this network cannot increase in value free of charge, it requires massive reinvestment and ongoing costs. These costs dictate that it will take longer to grow, it will experience lower profitability on capital than software firms, and it would never be able to dominate the global TV audience without paying for/creating all popular TV and movie content globally. Unlike web-based search, TV and movies are highly differentiable and other companies will own valuable content that is in high demand- either NFLX will have to pay for it or it will face competition elsewhere.
You could say, well although this is true, content suppliers will be harmed if they don’t sell to Netflix, because their platforms have few users and there aren’t many other distributors willing/able to pay the same price for content. That may be a good point but it is less true by the day. Given that network effects are very costly thanks to content costs, I think it is likely that the SVOD industry will shake-out to be oligopolistic with at least a few major players that have different content and pricing. I also believe that Hulu and others can benefit dramatically from advertising revenues. NFLX shuns ads, but they provide a subsidy which many consumers will suffer as long as they have access to a wide variety of cheap and high-quality content.
One final point is usually mentioned with regard to NFLX’s network effects which is the term ‘data network effects’, which basically purports that as a result of servicing so many members over the years, NFLX has a better understanding of what they want and can recommend and provide content which they enjoy. There is probably some benefit from this, but its much iffier than the other advantages. I think other SVOD providers will be able to find out very quickly what their members enjoy and I doubt that NFLX’s recommendation tab will detract from upstarts (they can often know which content is popular before buying it). To say this is a durable advantage would really be going out on a limb and should be viewed with much skepticism.
From all the points mentioned above, I find it highly likely that Netflix will experience the typical U shaped cost curve resulting in unit costs eventually rising as they push expansion (while for a pure platform it would remain near-zero permanently). I think Netflix is very good at distributing content that consumers globally are attracted to- blockbusters and popular mainstream shows like Friends, for example. As they diversify their offering and attempt to add users who are currently not attracted to NFLX, I believe they will increasingly be forced to add content that is outside of their current scope and which another firm or offering is currently supplying.
This, along with content suppliers either attempting to provide their own SVOD service or use a new service as a bargaining chip leads me to think that the operating leverage inherent in NFLX’s model is lower than expected by most investors. I do not see a point at which they will keep content spending flat, increase prices and continue to rapidly grow users worldwide (or have spending increase but revenues increasing much faster), which is what is currently priced into the stock. I also do not see an outcome whereby more and more consumers worldwide are attracted to NFLX without it rapidly expanding its content library. This means that they will eventually have to raise prices without increasing churn, which I also find unlikely.
It is becoming clearer that neither the optimist nor pessimist view is on the mark with NFLX. It seems highly unlikely that it will monopolize the SVOD market globally for wide profit margins, but it also has a large and growing membership which appreciates the variety and breadth of what they offer. It is likely to remain the largest player in SVOD but will experience rising content costs, substantially more competition, particularly from live-streaming and ad-funded offerings, and pressure from content suppliers. If it executes well, in maturity I could see it having margins and returns on capital better than cable providers at their peak given that NFLX benefits from virtual distribution and a larger TAM. This is still however, a far shot from the returns on capital of a pure software or platform provider and likely would be disappointing to most investors.