Welcome! This is the first of a two-part interview with Jonathan Levin, the Philip H. Knight Professor and Dean of Stanford Graduate School of Business. Jonathan’s research has spanned a range of topics, including auction and marketplace design, the economics of organizations, consumer finance, and econometric methods. In addition to undertaking academic research, Dean Levin has consulted for a number of Fortune 500 companies, as well as the Federal Communications Commission and the U.S. Treasury.
This week Jonathan and I discuss the implications of subscription models on economic “intangibles,” strategy and competition, and data and regulation.
Welcome, Jon! I’m looking forward to this discussion. More and more, the physical world seems to be evaporating around us. By that I mean products are turning into cloud-based services. Ownership is declining, and a new model of consumption-based usership is ascendant. But most of our current economic tools are still predicated around product sales and asset transfers. So how should economists think about a world without products? To kick things off, would you mind expanding on the concept of “intangibles” for us non-economists?
Sure. So over the past 50 years there’s been a profound shift away from investments in things like land, buildings, and equipment, which were historically the foundation of a lot of the US economy, towards “intangibles” like software, data, brands, R&D, patents, and intellectual property. This shift is one of the major historical transitions in the economy, akin to the move from agriculture to manufacturing and now to services.
Right. Alan Murray of Fortune notes that among S&P 500 companies, intangible assets have gone from 17% of total assets in 1975 to 90% last year, which is really amazing. And that has all sorts of implications: you can scale faster because you’re building a user base instead of a fleet of equipment, you have less need for investment dollars for the same reason, disruption is easier because we’re essentially just talking about intellectual property, and pricing is way trickier because costs at the margin are close to zero.
And I think it’s important to note that the whole concept of “services” probably needs updating. In this discussion, for example, we’re not focusing on services in terms of hourly wage earners, tourism and hospitality, etc. We’re largely talking about cloud-based, digital services.
Exactly. So if you look at companies like Microsoft or Facebook, virtually all of their assets or a very large fraction are intangibles. There was a study done of Microsoft about 10 years ago, suggesting that only 1-2% of its market value could be accounted for by its physical plant and equipment.
We all see this happening. However, we don’t necessarily have great ways of measuring the change. The reason we call them “intangibles” is because they’re difficult to measure with our current tools. Our measures of intangible assets are nowhere near as developed as our measures of physical assets.
That has implications for the way we track national statistics and think about business cycles and so forth. For example, we’ve looked historically at investment in physical capital as an important economic indicator, but that investment has been falling in many sectors of the economy.
If we were still largely an agricultural economy, you’d be worried if we weren’t buying enough tractors. But we’re not. And that might not be a bad thing! But business investment in physical assets is still a really important component of a country’s GDP.
Right. So how do you measure investment when it comes to intangibles? If you’re building software it’s in the wages you’re paying to your software developers. If you’re assembling data, sure you’re investing in a data center, but the actual data is not embodied, it doesn’t look like an investment. Its value shows up in other ways. So we need better ways of quantifying these assets, because they’re now an important part of the economy.
2: STRATEGY & COMPETITION
We tend to frame this as the shift from products to customers. And then we try to question assumptions that we might take for granted, but are actually based on a product-centric view of the economy. Is there a way to rethink these things from a customer-centric view?
So in a product-based view of the economy, all your physical assets have value, and the rest is intangible, and difficult to measure. But I’ll tell you when we look at a subscription-based business, whether it’s a telecom company or SaaS company, what you find is that physical assets are in some ways are a liability, because what you’re actually building is a service.
For those types of businesses, where intangibles are important, the underlying economics also are different. They often involve significant fixed investments, and if they’re successful, large gross margins. Or put another way, in areas like data and software, there can be big investments and then very low marginal costs involved in scaling a business.
Right. A telecom company invests in switches, networks, wires, towers, satellites, and those are all the costs to provide a service. But guess what? The technology shifts. So if you over-invest in your assets you can get into trouble: you’re moving to 5G, you’re getting rid of your data center so you can move everything over to Amazon, you’ve got obsolete hardware, etc.
That’s why gross margins become so important. When Wall Street looks at this, they’re asking how well are you doing using these costs and extracting value? And that value is represented by how much your customers will pay. The value of a physical asset becomes more tied to business efficiency than a stand-alone investment.
Here’s one thought along those lines: in the digital economy, there are often significant scale economies, and in some cases considerable spillovers from investments. Data is an important example. The interesting thing about data is that it’s non-rivalrous. Which is to say that if I develop great data, I can use it, and other people can use it, without degrading it. Whereas if I build a factory, if I’m using it I can’t have someone else use it. That’s a very different structure because if I make a successful data investment, there’s a potential spillover or externality.
3: DATA & REGULATION
Data is kind of the ultimate intangible. We all know that it’s incredibly important. All the big tech companies built on server farms. But we don’t really know how to measure it from an economic point of view.
Right. If I make an investment that results in a research breakthrough, or in a great piece of software that many people could use, that’s obviously hugely important. So you might worry in some ways if you’d see under-investment in these kinds of assets. Or maybe you need different rules about how to protect them. For example, my colleagues Chad Jones and Chris Tonetti at the GSB have proposed regulatory structures for data based on exactly this line of thinking.
Sure. Because you need solid measurements in order to create smart regulations, right? To your point about data, lots of new services are about an exchange between the provider and the customer in the sense that the customers provide data about themselves. And that data follows the vendor to provide a better service offering, right? I have your viewership history, I have your purchase history, I know what features you like, and which ones you don’t.
Ultimately, at a macro level, the more data sharing that goes on, the more value is generated from the system. If that’s the case, then it’s really important that we have regulations that provide more efficient curating of this type of data in a way that promotes trust.
So I should be able to take my Netflix data and take it to Hulu, so because I want Hulu to use my purchase history now to provide me better service, and lower switching costs between vendors creates a more competitive environment. But the starting point is to reframe all this stuff as a world of service offerings versus a world based on an asset transfer model.
We clearly need different rules of the road for a digital economy than we had for our traditional physical economy. And you can see in so many places, people are trying to look at what should those rules of the road be? How should we regulate the digital economy compared to the regular physical economy? How should we think of antitrust or market power in the digital economy compared to the physical economy?
Next week we’ll be discussing he implications of recurring revenue models on traditional accounting and income statements, how Stanford GSB is thinking about subscription models in terms of subject matter, and the potential effects of subscription models on business schools in general.