Scale and Loyalty are more important online than offline, which drives much of the “winner take most” reality of the internet.
Scale and loyalty are generally the most important metrics to me as an investor as they are the most sustainable — and measurable — competitive advantages. They are *much* more important online than offline, which is why the internet economy has such pronounced “winner take most” dynamics relative to the offline economy. Barriers to entry on the internet are low, but barriers to scale are high.
Scale and loyalty are more important online than offline primarily for four reasons. 1) The fact that “CAC is the new rent,” 2) the mechanics of how the Google and Facebook auctions work, 3) the reality that customer loyalty effectively lets online companies avoid paying this “rent,”and 4) the fact that AI/ML are so important on the internet and scale drives better AI/ML.
The statement “CAC is the new rent” comes from a Fast Company article. It refers to the fact that advertising spend, i.e. Customer Acquisition Cost, is generally the largest expense for most consumer oriented internet companies and that while in the offline world, it is impossible to be a retailer without paying rent, it is impossible to be a retailer in the online world without paying CAC. The most profound implication of this is that the TAM for online advertising spend is much larger than the TAM for offline advertising spend as online advertising is effectively replacing costs other than advertising when comparing online vs. offline consumer companies. This suggests that advertising should take share of global GDP as a greater % of GDP shifts online. And given that Google and Facebook have dominant share of online advertising, understanding how their auctions work is critical to understanding the internet economy.
The Google and Facebook auctions privilege scaled bidders via the ad quality score, which usually means bigger companies can pay less and still win the auction. Google and Facebook generally only get paid by an advertiser when a consumer clicks on an ad, so they assign a “quality score” that attempts to predict the likelihood that an ad will be clicked on. Hence the fact that a company with a higher quality score can win auctions with a much lower bid vs. a company with a lower quality score. The quality score generally works to the advantage of larger companies with brands and superior customer experiences that increase the likelihood that a consumer will click and Google/Facebook will get paid by the advertiser. Given that “CAC is the new rent” this is a bigger economy of scale than any comparable offline metric I can think of — it’s not as if a retailer or restaurant’s rent per location decreases significantly every time they double in size. This is one of the factors driving much higher ROICs for dominant online companies vs. comparably dominant offline companies.
Scale is almost always correlated with customer loyalty and customer experience; both online and offline. However, customer loyalty is even more important online given the importance of the ad quality score for costs and the fact that a loyal customer base lets an internet company avoid paying “rent” to Google and Facebook. Essentially it is almost impossible to be a large internet company if you don’t have a loyal customer base that raises your ad quality score and thereby lowers your CAC. Customer loyalty *generally* means more repeat customers and more organic traffic, which means the larger companies effectively are not paying “rent” — i.e. advertising — for a larger %’s of their sales/customers relative to smaller companies. Sidenote: if this isn’t happening, run away. Repeat customers and organic traffic are effectively “rent avoidance,” which magnifies the importance of frequency and retention. The cream really rises to the top on the internet, but once it is at the top it is hard to displace. Path dependence is real within each vertical as there was an initial period where market share was up for grabs, but if you didn’t grab it during this initial state it is really hard and expensive to get it later. The most fascinating takeaway for me from reading “Netflixed” was how close Blockbuster came to displacing Netflix as the #1 player only to give up at the 1 yard line and snatch defeat from the jaws of victory.
Beyond this, scale matters because so much of internet economics revolve around making accurate predictions and the more data one has, the more accurate one’s predictions are in an AI/ML environment. Internet companies need to be able to make accurate predictions around customer LTV, payback periods and recommendations (both for building baskets and increasing engagement). These are all critically important and become much more accurate with scale. One of the most fundamental principles to understanding a world dominated by AI/ML is that scale matters more than almost any other variable for AI quality. Multiple research papers from both Google and Microsoft have consistently found that AI quality, defined as variance from zero errors, doubles with every order of magnitude increase in the amount of data used to train the model. These scale based advantages combine to create positive feedback loops that often lead to share gains accelerating as companies grow larger rather than decelerating which is generally the case offline.
Important to note that relative market share is always the best way to measure scale online. 50% share when the #2 player has 3% share is *very* different than 50% share when the #2 player has 40% share. And even more so on the internet. This is one reason that I distrust the HHI index for online companies — a market with a 60/40 share split has an HHI of 5200 but it is much better to be a 60% share player in an online market where the #2 player has a 10% share (HHIs would generally be below 4000).
Perhaps the most elegant way to say all of this is that *variable* costs per unit generally decline online along with fixed costs per unit with increasing scale which is very different vs. the offline world where generally only fixed costs per unit decline with increasing scale. And this makes it very hard to regulate the internet — in fact the only effective “regulators” of the internet are actually Google and Facebook themselves as we shall see.
The increased importance of scale and loyalty for online co’s vs. offline co’s is why the internet economy is much more “winner take most” in almost every vertical relative to the physical economy despite the frictionless, low barrier to entry nature of the internet. .i.e. Amazon has higher share of e-commerce than Wal-Mart has of physical retail and the same goes for Netflix, Booking and every other major internet company. Even goes for The NY Times and WSJ’s share online vs. offline. And not only do the #1 online players generally have higher share of revenue than the comparable offline #1 player, they have a higher share of profits. Booking had $3 billion more revenue than Expedia in 2019, but $4 billion more EBITDA as a result of primarily operating one brand vs. Expedia’s six brands and focusing more on hotels relative to Expedia which results in much higher effective scale than explained by revenue differential alone.
All of this helps to explain why it is “only on the rarest of occasions” to quote Michael Scott opining on the right time for him to use a bailer 😀 that one sees an internet company with dominant share be displaced. And these displacements generally only happen when there is a fundamentally superior customer experience that results in higher customer loyalty that enables a competitor to avoid paying rent to Google and Facebook. Mobile can be understood as a one time discontinuity that let smaller competitors with superior app based experiences take share given that search — and all of the scale based advantages that go with it — was initially less important on IOS given apps and mobile happened at the same time that Facebook ads were priced inefficiently relative to search. Given the prevalence of marketplaces online, also critical to realize that these dynamics are even more important with marketplaces than other internet businesses given the obvious network effects which make scale, relative market share and loyalty even more important than in other internet businesses.
These winner take most dynamics also lead to three other features of the internet economy. 1) Google and Facebook will be inevitably pushed into adding more and more marketplace features over time. 2) There is a much higher sensitivity to changes in competitive intensity relative to the offline economy. 3) Wherever relevant all online players will eventually have an offline presence.
The“winner take most” dynamics for their customers are not good for Google and Facebook in the long run as they lead to significant leverage to the winning customers advertising spend on Google and Facebook. This is inevitably pushing Google and Facebook towards adding more marketplace features to their own platforms over time. Booking’s dominance in travel led to Google pushing further down the funnel to somewhat equalize the competitive playing field between Booking, Expedia and the hotel brands themselves as having Booking consistently winning auctions at a lower CPC because of their higher ad quality score was not good for Google. Amazon’s dominance is pushing Google towards adding more “shopping” functionality for retail customers and Facebook into effectively becoming a marketplace over time. Basically, Google and Facebook periodically need to “reset” the competitive environment by adding marketplace features that help the long tail of advertisers with lower quality scores that pay more per click. This is actually a very modest long term negative for Facebook and Google as advertising is a better business online than being a marketplace due to the simple fact that most advertisers structurally overestimate their LTVs by overestimating the likelihood of repeat business which leads to inflated CAC to LTV ratios. Also why it is critical to *always* ignore LTV calculations and focus on gross profit $ payback periods as an investor. Ironically, all of the EU’s regulatory efforts hurt the long-tail of small advertisers who are helped by these resets and helps the dominant advertisers that can afford to pay for lobbyists and white papers. EU regulation effectively prevents Google and Facebook (to a lesser extent) from “regulating” the internet themselves to both their own advantage and the advantage of SMBs.
Online businesses are also more sensitive to changes in competitive intensity than offline businesses given that the Facebook and Google auctions are second price auctions, which are inherently chaotic and nonlinear in terms of sensitivity to changing auction pressure. i.e. It is almost impossible to significantly change the “CAC” of a scaled offline retailer, but it is possible to significantly alter the CAC of a scaled online retailer if the challenger spends enough money. This possibility is why #2 online players — especially those with a good story around customer experience and loyalty — can get funding and why their increased spending can be so painful to the #1 player in a short period of time. This strategy assumes that the #1 player will not respond and the #2 player will be able to displace them, but almost all internet companies understand that they have to respond so it rarely works but does lead to Pyrrhic victories especially in VC land where the spending drives so much dilution (have to raise money to spend) that it really hurts founders and early investors.
Finally, the fact that “CAC is the new rent” for the internet economy also advantages omnichannel players. One of the best ways to lower online CAC, especially if you don’t have scale, is to have a physical retail presence and build your brand via traditional CPM advertising, especially television advertising which is likely undervalued in todays world. Said another way, on a long enough time horizon all scaled online retailers will have a physical retail presence albeit much smaller than the retail presence of pre-internet retail incumbents. A corollary to the principle that on a long enough time horizon, all internet companies become fin tech companies.
Understanding all of this is critical to being an investor in internet companies. Started to write this up as a Tweetstorm and then realized it was just too long. Hope it is helpful. TL, DR: Scale and loyalty matter more online than offline, creating the “winner take most dynamics” seen on the internet.