I think it was Mandelbrot that said something like the trend is killed by its discovery or something like that. Sure. It's amazing to watch that happen in real time in markets. Let's talk about scalability. One of my favorite posts in reviewing on the PlayNow yesterday, your big backlog was the 12 features of a scalable business. We don't need to go through all 12, necessarily, but we talked a little bit earlier about product market fit, a value hypothesis, which is you know it when you see it. The demand for the product or service is very high. The other side of that equation in Andy's way of thinking is the growth hypothesis and this is scalability. Maybe tell us a bit about what you think are the most important components of a business that has that scalability baked into it, presumably which makes that more attractive as a business person to run one, as an investor to buy an own one. I think this idea of scalability is fascinating in this kind of tech landscape that we're in today.

One of these things like product market fit and scalability is the kind of concept where there are many factors involved in it. Super dynamic. What you're looking for is a confluence of things and you may not have everything in a particular business, but some things are sort of critical and some things are nice to have. Usually, if you're looking for scalability, the VCs will tell you like Don Valentine was famous for this, which is he wants a big market and he didn't even want a market that might be big someday. He wanted it already to be big. The classic chart on the slide deck of a lot of startups is huge pie chart and they normally get a small part of it and everybody is going to have their own private island in the Caribbean. Huge markets really help because it's hard to scale a business for horse bridles or horse blankets or something like that. The big market and then the other thing is businesses that are highly people intensive like a law firm or an accounting firm, they are hard to scale. It's just so many people, people get jealous or they see a chance to make more money and they break off and whatever. If you have very people intensive services, it's hard to scale. You have size of the market that and then you have basically a tech, if you have a technology with cost or dropping as volume increases, if you have true economies of scale, business is more scalable. If you have demand side economies of scale, which are network effects and you have a lot of virality, a lot of new product getting more valuable, the more people use it, then that creates the business of scalable. Probably the key thing is you can't scale a business very well if you have inorganic approaches to acquiring customers. If you have to go out and buy radio ads, if you have to go out and do the stuff that these meal delivery firms are doing, it's a hard slog. If you have a product that is naturally viral like Facebook was or whatever, you're kind of running customers for almost nothing. The business is growing quickly. You don't need that many people. They still don't employ that many people given how they're more than that. That's how it is. Yeah. Go down the list of these things and there's sort of an additive effect. No business is perfect and they're not all quite the same. But those are the sorts of factors which determine the scalability of the business. It's sort of the last one. This is one of the trickiest ones is elasticity, which is if you take a business like SpaceX is in just the launch business. It's only a five and a half billion dollar business only, but there's a lot of competitors. There's Russians and Chinese and Indians and US government traditional launchers and the Europeans have an aerial on a spot. And so they're all fighting it out for launches. And if you get to a month, you have a big 30, 40% market share. So the market can only get so big. And so for launching payloads into space, also the payloads are getting smaller because of Moore's Law and a number of factors and they're coming lower or urban. And so the other way you can really grow that market is if humans start taking rides up there. So what's the market for taking rides into space? How many people have the kind of money they're going to go up in for four minutes can float around the space and come back down? And predicting that kind of thing is like what Craig did with the cellular business. He knew that these things were going to be popular someday when they got cheaper. So maybe there's a business doing that. It's certainly dicey. And so one of the things that SpaceX has done and said, okay, well, we want to have a backup. So we're going to put up our own broadband satellite communication system called Starlink. That's a completely separate business and they're creating a customer that's going to be within the same P&L presumably unless they spin it off because the business is only so big. So that's an example of a company that has, I believe, some

unanswered questions about the elasticity of demand. Maybe we could use a recent post you did on MoviePass as a way of talking about some of these other levers that we haven't gone deeply into. You mentioned Gurley who's probably my favorite writer on this concept of lifetime value to customer acquisition costs. And reading you, reading Gurley, reading others on this topic has really made me try to think about a lot of businesses in these really, really simple elemental terms. It makes life easier, which is great. But it's always easier to learn via example. It doesn't need to be MoviePass. You can pick your business. But I wonder if you'd be willing to pick a modern business one that people might be aware of and maybe use that as an excuse to talk about things like lifetime value, customer acquisition costs the money, how you think about these inputs into that equation. So

one of the key things in life is Roger Fisher's getting the S book as a concept called Best Alternative Negotiated Agreement. Basically what it says is if you're negotiating for something and you only have one choice, you're screwed. And that translates into another concept that another professor has across the river at Harvard Business School and Michael Porter said basically there's these five forces that you have to think about. And one of the five forces is supplier bargaining power. And so what I did in my own life with Craig was I got to have this conversation where I was listening to John Malone talk about wholesale transfer pricing. And basically it's supplier bargaining power all the way back to basically opportunity costs. And the MoviePass business model is fundamentally dependent on the price of those tickets and they're buying them at retail. And they only have one supplier because every movie isn't, you don't buy tickets in the movie industry. You buy tickets to Black Panther or whatever. You have to go into that movie, right? So they're buying at retail and they're selling at a discount in a subscription and they have a huge, is big a wholesale transfer pricing problem as I've ever seen anywhere

at scale. Can you describe that term in more detail? You talk about this a lot and it's such a great concept, wholesale transfer raising. Yeah, it's basically imagine you were at a bakery and

you were making bread. In the city they got a lot of bread bakeries and you go down there and you had to buy your flour from Joe. You couldn't buy it from Ben or anything. You had to buy your flour from Joe. You have a wholesale transfer pricing problem because he can price the flour, whatever price he wants. This is Spotify's fundamental problem in that they

have these- Their margin gets taken.

Yeah, it gets margin gets taken. And if they ever got some margin, they just up the price. And so that is the same principle as Roger Fittness-Battner, Charlie Munger's optionality, his Porter's Sustainable Conventive Advantages. You never want to have one supplier of anything. And that's something like, tell him to talk about Luigi or whatever, but some basic guy who's making pasta knows that he doesn't want to be completely dependent upon his brother-in-law for vegetables because he does, his brother-in-law is going to basically take a lot of his profit. And so that applies in SpaceX. Elon very smartly and brilliantly sort of doesn't rely on other people who can have- He's avoided wholesale transfer pricing. The inverse of that is movie pass, which is completely dependent on these movie studios, selling them- actually goes through the theaters, but they got to buy from the theaters like Regal at retail. And you can't make that up on volume. It just- negative gross margin is awful. It's sort of like another point we should sort of get into, which is I'm in the software business. I look at other businesses and I see the gross

margins and I go, my God. What a nightmare. How did they survive? They got

no money for anything. They got 2% net margin and the gross margin is tiny. And if it snows, the trucks don't roll, it's like, whoa! These guys, they better have some cash in the bank to go through these bumpy points. But I'll go back to my own business and say, God, we're lucky. This is great. These gross margins are

nice. You've written a ton about SaaS businesses, the SaaS business model, software business model. I'd be curious if there are other chunky categories that you think are worth explaining versus say the SaaS business

model in today's terms. Well, I think one of the things that's interesting and exploring on the SaaS business model is the idea that subscriptions are somehow magical. Subscriptions are a way of charging for a product and they're a way of getting a customer in a situation where you're not constantly fighting for renewal. And so there are some aspects of it that are good. But unless you have dog food that the dogs want to eat, subscription is

going to get you a buckus.

And so you still have to have a product that people really love and want to buy. Subscription can make the product better and it can make the model better because if somebody signs up for a year, you don't have to worry about them leaving for 12 months. That's the good news. Everything has a flip side. The bad news is you have to pay a little bit more to get them. Unless the product is truly viral, your caca is going to go up. And this is the point I talk about about all your variables in the LTV equation, which is they're all linked. They all have ropes to attach to each other. You tug on one and the other ones all move. It's like a spider's web or whatever. And so everything's dynamic, everything's related. And that's the fun of it.

It would be boring if it was just any year

and not dynamic. And that's why it's a game. That's why Buffett does a tap-dance on the work. It's a work every day because it's fun. To me, it's the biggest game of all.

Yeah, it's something I have to remind myself of all the time as someone that's prone to want to stuff everything into a formula that you can't always do that. And there is art to all this stuff. Even if you're building quant models, there's an art component to it. And I think that that's a fantastic reminder of it. Let's take a business like Apple as an example where I guess, are there any subscription areas of the Apple business model? Oh, sure.

Yeah, they have a lot of

them. And they're also a wholesaler of people who sell subscription

models. Right, sure.

It's an 80 model that Bill Gurley and I talked about. So during the .com era, I was sent down to Silicon Valley every week. I went down to Monday, come back on Friday, and I spent a lot of time with Benchmark. And Gurley and a guy named Bruce Dunleavy and Andy Ratcliffe and Bob Kegel, he got down the Steve Sperlock and all that. They were all mentors of mine, and they taught me a lot of things. But at the time, I was in the telecom business. And in the telecom business, we had a private equity firm. We went on Nextel, Nextel International, and XO. We had a bunch of different investments. I was just pining for software margins. And I was pining for pricing power because it was just tough business selling long haul against the level theory or whatever. And I got this idea for a thesis, I call it software in a box, which is basically, sometimes you have to sell a box to sell

software. It's like Peloton. Yes.

And that's the classic software in a box case, which is you sell this thing, which enables you to sell this service. And the margins are in the service. The stickiness is in the thing, and the enabling thing is the thing. It's like there are no Peloton clones you can just go out and buy. But the bad news is you have to create these and sell these things. The good news is that if you get it actually a feedback loop going, you can make them really more cheaply with everything else. And the thing that differentiates a GoPro from a Peloton is GoPro, I think, never my opinion is they never invested enough in software and creating the software that came with that box, because that would have given some staying power, because then you're not just competing against the latest thing to come

up. No, it's cost

producer. And so software, it becomes critical. But sometimes the hardware is an enabling thing. Even if you take the chips that Qualcomm or other folks make, a company like Intel employs more software engineers than hardware engineers. Software is key, and that's software in a box too. So software is truly eating the world, as Andreessen says. But sometimes hardware is enabling the distribution. Distribution is people who are able to get products distributed, they're magicians in their own right. And having somebody on your team who knows how to distribute and sell

products, highly underrated.