How Spotify, Zillow and more Could Dominate Their Markets
Organizations that excel in data science are going to see their productivity compound exponentially. The current tech sell-off presents a meaningful long term opportunity.
Power stems from capitalizing on the invisible, long term trends
Over the last few days I have been reading Peter Drucker´s “Post-Capitalism Society”. It´s a very thought provoking book, but the most relevant aspect (in my brain, today) is the overview it gives of the evolution of productivity over the last couple of centuries. Productivity is the sort of thing we all hear about, but that you have to zoom out a couple of centuries to really understand, because it evolves in the background and quite slowly compared to our rather short term impulses and thoughts. “The Clock of the Long Now” by Stewart Brand talks about the importance of thinking long term and how power stems from capitalizing on the changes that happen slowly, for a long period of time. Power in the form of our ability to preserve and enhance the environment, to expand human lifespan or to make great investments and create meaningful organizations that foster prosperity and peace, for instance. Productivity is such a factor in our world. I believe that our ability to harness data is going to produce a world changing acceleration in productivity and that there are a number of stocks out there, that currently sell at a discount, that will play a key role in this.
The history of productivity and knowledge
According to Drucker, the predominance of the middle class in the world, together with rising standards of living is the result of having transitioned into a world in which we apply knowledge to knowledge. Initially, knowledge was confined to narrow pockets, in the form of guilds, for instance. Guilds essentially owned the knowledge relevant to its function and they would be rather secretive about it. Such a dynamic as it pertained to knowledge essentially inhibited productivity, because it stopped people from knowing how to do more with less. In the second half of the 19th century, we arguably transitioned fully towards empiricism and ultimately, we learned to apply knowledge to work. The first industrial revolution was powered by a series of advances in knowledge (like interchangeable parts and motion powered by steam) that permeated verticals and enabled a productivity revolution. This enabled the proletariat to become the middle class or near-upper class, because it enabled people to create more with less and make a higher income.
Allow me to take a step back and explain what I mean by knowledge. Drucker argues that the stirrup enabled feudalism, by enabling knights and their associates to fight from a horse and in this way, be technologically far superior to their corresponding peasants. This is a stirrup:
If you see a stirrup from a materialist point of view, you will see a piece of metal shaped in a certain way — an object. But in fact, you can see the stirrup at a deeper level, by realizing that it is actually just information and energy. It is the result of the acquisition of knowledge by humans (steps to perform) together with the input of a certain amount of energy. In other articles, I have explored how advances in information and energy are precursors to human wealth. The stirrup is one of the many seemingly petty and ridiculous but very real examples that history grants of this.
After the First Industrial Revolution, Drucker argues that we gradually learned to apply knowledge to knowledge. Today, it is evident that we live in a world in which knowledge crosses domains and evolves at impossible speeds. As automation increases, humans are increasingly focused on delivering value in the knowledge side of things, versus in the mechanical work side of things. This is evidenced by the % of the workforce that works in manufacturing in developed economies (relatively low, getting lower). Knowledge now drives the world.
Knowledge on knowledge on data
We humans have infinite potential in many ways, but not when it comes to processing data. This is a dark corner to explore and I will do so only briefly, but it turns out that reality, both within and outside of us is far more predictable that we believe it to be, if enough data can be gathered on the area of reality in question. This has been brought about by advances in neural networks and big data and it presents a fundamental advancement in regards to our productivity, as the stirrup did for the knights (unfortunately for the peasants).
Why? Because anyone who owns a dataset rich enough on a particular matter is now able to perform all sorts of magic with it, that our biological brains could not. Consider Spotify, for instance. What it is doing is gathering data on what we like to listen to. It is gathering data from hundreds of millions of humans, in different locations and with different cultures. The dataset it is generating is lethal for music labels and radio channels, because eventually, Spotify will know exactly what song or podcast to produce or feed to you, so that you stick with them. It will know what combination of notes is best to produce/feed during a warmer than usual Christmas to people who prefer pop music over hip hop, for example. Spotify´s productivity (and therefore margins) will be much higher than that of incumbents. This will manifest as an ad network similar to that of Google´s, but in audio format. Already, Spotify´s recommendations are making my life way more fun.
Zillow is another one. People buy, sell and rent homes and will do probably forever. Agents and firms participate in the market to deliver and capture value. Meanwhile, Zillow is doing to real estate what Spotify is doing to audio. Eventually, it will know more than all the firms and agents in the US combined. It will know when to buy what property at what price, in order to sell it for a profit, for instance. It will do so far quicker than any agent operating in the analog space. In essence, it will have the best knowledge on how to perform any function in the market optimally.
In essence, Spotify and Zillow follow the same blueprint. They are both platforms that facilitate the interaction between participants in a market, whilst carefully taking notes of what they do and do not do, when and how. As such, they gather vast amounts of data, which together with AI, turns into a large, nimble and powerful moat of knowledge. Now, both these companies sell at meaningful discounts with respect to their recent ATH and do not sell at vast multiples. SPOT P/S = 4.4, Z P/S = 7.58.
In the next few weeks, I will be diving into companies that follow this blueprint and that trade at reasonable multiples. Another potential one could be Teladoc, but I am yet to take a good look at it. In this post, I explain how Palantir enables companies to create digital twins of themselves and therefore generate large datasets and then perform work on them. PLTR P/S = 20 ish, still a bit high.