One person whose opinions I value is Ben Thompson of Stratechery. He produces a daily newsletter from his lair in Taiwan but once a week it is free for anyone to read (I don't have a subscription so I only see the free ones). Also, roughly weekly, he creates a podcast called The Exponent along with James Allworth. Normally they discuss mostly what is in the free edition of the daily newsletter, but they range wide and sometimes Ben will come to a conclusion under James's challenging that is different from the one he came to in his newsletter.

Nothing directly to do with this post, but a month ago, in a podcast called It's Been a Week, the two of them discussed the passing of Clayton Christensen that I talked about in my post RIP Clayton Christensen. Not only was James Allworth a student in Clay's MBA class at Harvard, but he was also invited to stay on and do research with him. But soon after he started, before he'd really got going, Clay had a stroke. Maybe due to this intimation of mortality, his next book was How Will You Measure Your Life? but having had a stroke he needed a lot of help to get it written from a co-author. That co-author was, yes, James Allworth. So his insight into Christensen's ideas, personality, and motivations are at another level from almost anyone outside his immediate family, and I recommend listening to the podcast.

This week's podcast, Scale Scale Scale, is about the direct-to-consumer market (DTC). This is not directly a semiconductor issue, we don't sell chips or EDA direct to the consumer, but it is an aspect of the global cloud scale of the internet. Also, all of the trillion-dollar market cap companies build some of their own silicon, and they are all making major investments in artificial intelligence and deep learning. It's important to keep an eye on the dynamics out in the marketplace, especially when the government starts to get involved. The big reason for that is how poorly the government seems to understand tech.

As Bruce Schneier recounted at this week's RSA conference, a sitting US senator asked Mark Zuckerberg how Facebook made money.

I can't believe two things about this. One is how ignorant the question is. And second, that nobody on his staff pointed out to him that it was a stupid question.

Scale Scale Scale

The DTC market is composed of companies that sell product directly to consumers without going through normal brick and mortar stores, Amazon, or anyone else. They typically operate by acquiring new customers through advertising on Facebook, Google, Twitter, Snapchat, and similar places (I'm just going to say Facebook from now on, like Ben and James do in the podcast). They then fulfill your orders, and, if you like them then they can keep you as a customer for some period of time.

One company that did this quite successfully was Harry's Razors. These are actually quite a good product for DTC since it is naturally recurring as blades get blunt (I should point out that I've had a beard for 40 years so this is a fight in which I have no dog). When I first heard about selling mattresses over the internet (Casper, etc), I had the opposite thought, since that is almost a one-time purchase. It seems most people only buy a new mattress when they move house, or perhaps every decade or two. The non-DTC market for razor blades, sold in drugstores, supermarkets, and places like Target is dominated by two companies, P&G (Gilette) and Edgewell (Schick). Even as Harry's was growing and being successful, these two companies were sufficiently insulated from the DTC market that they could increase prices regularly in their comfortable duopoly. They operated at scale in research, manufacturing, advertising, and shelf-space.

Harry's decided to move into the offline market, exclusively at Target, with prices lower than P&G and Schick. Probably their manufacturing costs were higher, despite setting a lower price. Harry's immediately started winning share at Target. Gilette and Schlick both quickly reacted by lowering prices to compete. Edgewell closed a deal to acquire Harry's for $1.37B. The FTC has moved to block it and Edgewell has abandoned the deal (amazingly, Harry's had not negotiated a breakup fee, it seems).

As the FTC said in their document challenging the acquisition:

Harry’s significant entry into brick-and-mortar retail transformed the wet shave razor market from a comfortable duopoly to a competitive battleground...The Proposed Acquisition is likely to result in significant harm by eliminating competition between important head-to-head competitors. The Proposed Acquisition also will harm competition by removing a particularly disruptive competitor from the marketplace at a time when that competitor is currently expanding into additional retailers.

Three interesting things discussed in the podcast are:

• The FTC is right, and clearly Harry's has put downward pressure on pricing
• If Harry's had stayed as a DTC brand, it would not have affected the brick-and-mortar market much, and probably the acquisition would have been allowed.
• So if DTC brands want an exit through acquisition, they should not try and compete directly with offline brands in brick-and-mortar stores. Namely, by blocking the acquisition, the FTC may be making an improvement in the market for shaving products but decreasing competition in all other brick-and-mortar products where there are DTC competitors. Yet this is about the only way a new competitor in the brick-and-mortar market has a chance of emerging. A non-DTC startup has no chance to even get shelf-space in drugstores, supermarkets, and Target. So the FTC is likely to achieve more market concentration, not its goal at all, but an unintended consequence.

Ben's view is that Harry's was pretty much forced into the offline market since it couldn't really make money in the online market (he admits that he has not seen their financials). The reason is that in the online market, having access to the customer base is the important thing and the place to where all the profit flows. That means that Facebook takes all the profit since they know who to show the ads to. An important point is that Facebook doesn't set its ad rates like an old-fashioned newspaper, it is set by auctions between all the companies trying to reach the appropriate demographic.

One lesson from Casper mattresses is that DTC brands have very low barriers to entry—If you start a brand and pay more for the ads then you reach the customers and the previous DTC champion does not. As I quoted in my post Benedict Evans 2020: Standing on the Shoulders of Giants about Casper:

A vacuum-packed mattress was a brilliant idea...until everyone else did it. There are now 175 online mattress companies.

Those 175 companies all have the same business model. Advertise on Facebook and try to get to bigger scale than the other 174 fast enough that they are not shipping money with every mattress. They also have another problem, that the razor blade people do not have, that people have noticed that a free 90-day mattress return policy means you get a free mattress for three months, and then you can get another new one from another supplier. Repeat.

As I said, Facebook doesn't set its ad rates. What it does set is how many ads each person gets to see. In some sense, they want to show each person as many ads as they can get away with, without driving away disengagement. They absolutely do not want to do what AT&T has done, and allow so much "advertising" (aka spam) that people disconnect completely (yes, mobile is a bigger trend, but it seems to be junk calls rather than non-use that drives people to finally cut the cord). Of course, AT&T has very little say in the matter since they are required to connect calls (imagine if each phone was only allowed to receive one junk call a day and they could auction it to the highest bidder).

Credit Karma

Another acquisition that seems similar but is very different is Intuit acquiring Credit Karma (CK) for $7.1B. CK has a different business model. They provide free financial services such as credit scores, financial analysis, and tax filing. Users don't pay them. As Ben says "they don't even have a Stripe account". They have about 90M customers in the US alone (customer meaning users who have consented to receive emails about their credit scores and finances). Half of these are millennials. Google tells me there are 70M millennials in the US, so they are reaching about 60% of all millennials. They make money by analyzing people's financials and recommending credit cards that match their profiles (you do a lot of business travel, or you buy a lot of gas, etc). As you can imagine, the lifetime value of acquiring a millennial as a credit card customer is high and so the card companies will pay a lot. The deal is expected to be approved, perhaps with CK's free tax filing sold off.

Monopolizing Supply or Customers

A key insight is that antitrust law is worried about monopolizing (or having big share) of the means of supply. Rockefeller's Standard Oil, IBM in mainframes (and earlier in typewriters!), and so on. At internet scale, it is monopolizing (or having a big share of) the sources of demand that should be the concern. Ben and James stated this but passed over it. I think it is a deep observation, and the main reason I wrote this post.

It is also the the key to understanding why we are concerned with just how big and powerful Facebook, Google, and the rest have become. It is not traditional antitrust, we are not worried the prices are too high or are going to go up—they are zero. It is their concentration in the advertising market that makes them powerful. But they can't even be ordered to lower their prices by the FTC (or something similar) since they are auctions. They could, in principle, be ordered to show more ads (meaning the auction prices might go down) but we users would be annoyed if every other item in our feeds was an ad, and it is poor politics to annoy your voters. I don't see any good solution. Breaking Instagram out of Facebook and YouTube out of Google would make advertising a little more competitive, but I don't think it is not going to have the effect of creating a sudden flurry of new photo and video sharing sites. This is why traditional anti-trust is inappropriate and doesn't really match the perceived problem. To make a change would require deciding what outcome would be appropriate and Congress doing something about it.

That is what Facebook and its competitors do so well. Since we all get Facebook for free, we are not worried about the price going up. Facebook is not monopolizing advertising, it is just that they really are better at finding customers for DTC customers like Harry's than anyone else (and mobile games, and other services). Since Facebook is not setting their ad rates, their advertising customers are only complaining about those prices going up since they are bidding against their competitors for a limited resource. As Casper discovered in mattresses, their competitive edge was not that their mattresses were better, but they paid more for Facebook and podcast ads than anyone else—not much when they had the DTC mattress business to themselves, but all their profit (and maybe more) now that they have 174 competitors. I notice Casper is now starting to have its own brick-and-mortar stores and even selling through some Target stores. Sounds familiar!

The thing that makes Credit Karma so valuable to Intuit is the parallel of this. They provide a source of new customers for Intuit services without having to pay the Facebook advertising tax, and without having to compete with any startups who will have to pay the tax to find its customers. Intuit will already have a relationship with 60% of millennials (and some older people, too). It seems unlikely that 174 competitors are going to try and compete head-to-head with CK anytime soon.

Summary

Modern internet companies with zero marginal cost upend who in the supply chain has the leverage to make most of the money. You can make a similar comparison (as Ben has) about PCs and mobile phones. In PCs, the leverage was with Wintel (Intel and Microsoft) and the companies that build the PCs got comparative crumbs. In mobile phones, Apple makes all the profit, the chip/processor vendors make a little, and Apple creates its own operating system software.

The big thing in an internet company is owning the customer list, and charging a lot for access to it. The suppliers buying ads on Amazon, Google, Facebook, etc are getting the crumbs because they have to compete with all their peers to get their hands on the customers.

More Information

How Will You Measure Your Life? (Amazon link) by Clayton Christensen, James Allworth, and Karen Dillon

Stratechery and the post Email Addresses and Razor Blades that discusses the two acquisitions I discussed in this post.

The Exponent Podcast (series). That's the Apple link but is available in all the usual places. Here's the Google Podcast link.

It's Been a Week (podcast on Clayton Christensen)

Scale Scale Scale (podcast on the DTC market)

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.