Time for a revolution

I am rethinking the economics of technological change, especially in the context of cloud computing and its spectacular rise, as, essentially, a new and distinct segment of digital business. As I am teaching microeconomics, this semester, I am connecting mostly to that level of technological change. I want to dive a bit more into the business level of cloud computing, and thus I pass in review the annual reports of heavyweights in the IT industry: Alphabet, Microsoft and IBM.

First of all, a didactic reminder is due. When I want to study the business, which is publicly listed in a stock market, I am approaching that business from its investor-relations side, and more specifically the investor-relations site. Each company listed in the stock market runs such a site, dedicated to show, with some reluctance to full transparency, mind you, the way the business works. Thus, in my review, I call by, respectively: https://abc.xyz/investor/ for Alphabet (you know, the mothership of Google), https://www.microsoft.com/en-us/investor as regards Microsoft, and https://www.ibm.com/investor as for them Ibemians.

I start with the Mother of All Clouds, i.e. with Google and its mother company, namely Alphabet. Keep in mind: the GDP of Poland, my home country, is roughly $590 billions, and the gross margin which Alphabet generated in 2019 was $89 857 million, thus 15% of the Polish GDP. That’s the size of business we are talking about and I am talking about that business precisely for that reason. There is a school in economic sciences, called new institutionalism. Roughly speaking, those guys study the question why big corporate structures exist at all. The answer is that corporations are a social contrivance which allows internalizing a market inside an organization. You can understand the general drift of that scientific school if you study a foundational paper by O.D. Hart (Hart 1988[1]). Long story short, when a corporate structure grows as big as Alphabet, I can assume its internal structure is somehow representative for the digital industry as a whole. You could say: but them Google people, they don’t make hardware. No, they don’t, and yet they massively invest in hardware, mostly in servers. Their activity translates into a lot of IT hardware.

Anyway, I assume that the business structure of Alphabet is informative about the general structure and the drift of the digital business globally. In the two tables below, I show the structure of their revenues. For the non-economic people: revenue is the value of sales, or, in analytical terms, Price multiplied by Quantity.     

Semi-annual revenue of Alphabet Inc.

The next step is to understand specifically the meaning of categories defined as ‘Segments’, and the general business drift. The latter is strongly rooted in what the Google tribe cherishes as ‘Moonshots’, and which means technological change seen as revolution rather than evolution. Their business develops by technological leaps, smoothed by exogenous economic conditions. Those exogenous conditions translate into the Alphabet’s business mostly as advertising. In the subsection titled ‘How we make money’, you can read it explicitly. By the way, under the mysterious categories of ‘Google other’ and ‘Other Bets revenues’, Alphabet understands, respectively:

>> Google other: Google Play, including sales of apps and in-app purchases, as well as digital content sold in the Google Play store; hardware, including Google Nest home products, Pixelbooks, Pixel phones and other devices; YouTube non-advertising, including YouTube Premium and YouTube TV subscriptions and other services;

>> Other Bet revenues are, in the Google corporate jargon, young and risky businesses, slightly off the main Googly track; right now, they cover the sales of Access internet, TV services, Verily licensing, and R&D services.

Against that background, Google Cloud, which most of us are not really familiar with, as it is a business-to-business functionality, shows interesting growth. Still, it is to keep in mind that Google is cloud: ‘Google was a company built in the cloud. We continue to invest in infrastructure, security, data management, analytics and AI’ (page 7 of the 10K annual report for 2019). You Tube ads, which show a breath-taking ascent in the company’s revenue, base their efficiency and attractiveness on artificial intelligence operating in a huge cloud of data regarding the viewers’ activity on You Tube.

Now, I want to have a look at Alphabet from other financial angles. Their balance sheet, i.e. their capital account, comes next in line. In two tables below, I present that balance sheet one side at a time, and I start with the active side, i.e. with assets. I use the principle that if I know what kind of assets a company invests money in, I can guess a lot about the way their business works. When I look at Alphabet’s assets, the biggest single category is that of ‘Marketable securities’, closely followed by ‘Property and Equipment’. They are like a big factory with a big portfolio of financial securities, and the portfolio is noticeably bigger than the factory. This is a pattern which I recently observe in a lot of tech companies. They hold huge reserves of liquid financial assets, probably in order to max out on their flexibility. You never know when exactly you will face both the opportunity and the necessity to invest in the next technological moonshot. Accounts receivable and goodwill come in the second place, as regards the value in distinct groups of assets. A bit of explanation is due as for that latter category. Goodwill might suggest someone had good intentions. Weeell, sort of. When you are a big company and you buy a smaller company, and you obviously overpay for the control over that company, over the market price of that stock, the surplus you have overpaid you call ‘Goodwill’. It means that this really expensive purchase is, in the same time, very promising, and there is likely to be plenty of future profits. When? In the future, stands to reason.

Now, I call by the passive side of Alphabet’s balance sheet, i.e. by their liabilities and equity, which is shown schematically in the next table below. The biggest single category here, i.e. the biggest distinct stream of financial capital fuelling this specific corporate machine is made of ‘Retained Earnings’, and stock equity comes in the second place. Those two categories taken together made 73% of the Alphabet’s total capital base, by the end of 2019. Still, by the end of 2018, that share was of 77%. Whilst Alphabet retains a lot of its net profit, something like 50%, there is a subtle shift in their financing. They seem to be moving from an equity-based model of financing towards more liability-based one. It happens by baby steps, yet it happens. Some accrued compensations and benefits (i.e. money which Alphabet should pay to their employees, yet they don’t, because…), some accrued revenue share… all those little movements indicate a change in their way of accumulating and using capital.   

The next two tables below give a bird’s eye view of Alphabet in terms of trends in their financials. They have a steady profitability (i.e. capacity to make money out of current business), their capacity to bring return on equity and assets steadily grows, and they shift gently from equity-based finance towards more complex a capital base, with more long-term liabilities. My general conclusion is that Alphabet is up to something, like really. They claim they constantly do revolution, but my gut feeling is that they are poising themselves for a really big revolution, business-wise, coming shortly. Those reserves of liquid financial assets, that accumulation of liabilities… All that stuff is typical in businesses coiling for a big leap.  There is another thing, closely correlated with this one. In their annual report, Alphabet claims that they mostly make money on advertising. In a narrow, operational sense, it might be true. Yet, when I have a look at their cash-flow, it looks different. What they have cash from, first and most of all, are maturities and sales of financial securities, and this one comes as way a dominant, single source of cash, hands down. They make money on financial operations in the stock market, in somehow plainer a human lingo. Then, in the second place, come two operational inflows of cash: amortization of fixed assets, and tax benefits resulting from the payment of stock-based compensations. Alphabet makes real money on financial operations and tax benefits. They might be a cloud in their operations, but in their cash-flows they are a good, old-fashioned financial scheme.  

Now, I compare with Microsoft (https://www.microsoft.com/en-us/Investor/sec-filings.aspx). In a recent update, titled ‘#howcouldtheyhavedoneittome’, I discussed the emerging position of cloud computing in the overall business of Microsoft. Now, I focus on their general financials, with a special focus on their balance sheet and their cash-flow. I show a detailed view of both in the two tables that follow. Capital-wise, Microsoft follows slightly different a pattern as compared to Alphabet, although some common denominators appear. On the active side, i.e. as regards the ways of employing capital, Microsoft seems to be even more oriented on liquid financial than Alphabet. Cash, its equivalents, and short-term investments are, by far, the biggest single category of assets in Microsoft. The capital they have in property and equipment is far lower, and, interestingly, almost equal to goodwill. In other words, when Microsoft acquires productive assets, it seems to be like 50/50 their own ones, on the one hand, and those located in acquired companies, on the other hand. As for the sources of capital, Microsoft is clearly more debt-based, especially long-term debt, than Alphabet, whilst retaining comparatively lower a proportion of their net income. It looks as if Alphabet was only discovering, by now, the charms of a capital structure which Microsoft seems to have discovered quite a while ago. As for cash-flows, both giants are very similar. In Microsoft, as in Alphabet, the main single source of cash is the monetization of financial securities, through maturity or by sales, with operational tax write-offs coming in the second place. Both giants seem to be financially bored, so to say. Operations run their way, people are interested in the company’s stock, from time to time a smaller company gets swallowed, and it goes repeatedly, year by year. Boring. Time for a revolution.      

Edit: as I was ruminating my thoughts after having written this update, I recorded a quick video (https://youtu.be/ra2ztH3k0M0 ) on the economics of technological change, where I connect my observations about Alphabet and Microsoft with a classic, namely with the theory of innovation by Joseph Schumpeter.

[1] Hart, O. D. (1988). Incomplete Contracts and the Theory of the Firm. Journal of Law, Economics, & Organization, 4(1), 119-139.


[1] Hart, O. D. (1988). Incomplete Contracts and the Theory of the Firm. Journal of Law, Economics, & Organization, 4(1), 119-139.


I  am considering the idea of making my students – at least some of them – into an innovative task force in order to develop new technologies and/or new businesses. My essential logic is that I teach social sciences, under various possible angles, and the best way of learning is by trial and error. We learn the most when we experiment with many alternative versions of ourselves and select the version which seems the fittest, regarding the values and goals we pursue. Logically, when I want my students to learn social sciences, like really learn, the first step is to make them experiment with the social roles they currently have and make many alternative versions thereof. You are 100% student at the starting point, and now you try to figure out what is it like to be 80% student and 20% innovator, or 50% student and 50% innovator etc. What are your values? Well, as it comes to learning, I advise assuming that the best learning occurs when we get out of our comfort zone but keep the door open for returning there. I believe it can be qualified as a flow state. You should look for situations when you feel a bit awkward, and the whole thing sucks a bit because you feel you do not have all the skills you need for the situation, and still you see like a clear path of passage between your normal comfort zone and that specific state of constructive suck.   

Thus, when I experiment with many alternative versions of myself, without being afraid of losing my identity, thus when I behave like an intelligent structure, the most valuable versions of myself as learning comes are those which push me slightly out of my comfort zone. When you want to learn social sciences, you look for those alternative versions of yourself which are a bit uncomfortably involved in that whole social thing around you. That controlled uncomfortable involvement makes you learn faster and deeper.

The second important thing I know about learning is that I learn faster and deeper when I write and talk about what I am learning and how I am learning. I have just experienced that process of accelerated figuring my s**t out as regards investment in the stock market. I started by the end of January 2020 (see Back in the game or Fathom the outcomes ) and, with a bit of obsessive self-narration, I went from not really knowing what I am doing and barely controlling my emotions to a portfolio of some 20 investment positions, capable to bring me at least 10% a month in terms of return on capital (see Fire and ice. A real-life business case).

Thus, consistently getting out of your comfort zone just enough to feel a bit of suck, and then writing about your own experience in that place, that whole thing has the hell of a propulsive power. You can really burn the (existential) rubber, under just one condition: the ‘consistently’ part. Being relentless in making small everyday steps is the third ingredient of that concoction. We learn by forming habits. Daily repetition of experimenting in the zone of gentle suck makes you be used to that experimentation, and once you are used to that, well, man, you have the turbo boost on, in your existence.

This is precisely what I fathom to talk my students into: experimenting outside of their comfort zone, with a bit of uncomfortably stimulating social involvement into the development of an innovative business concept. The type of innovation I am thinking about is some kind of digital technology or digital product, and I want to start exploration with rummaging a little bit in the investor-relations sites of publicly listed companies, just to see what they are up to and to find some good benchmarks for business modelling. I start with one of the T-Rexes of the industry, namely with Microsoft (https://www.microsoft.com/en-us/investor ). As I like going straight for the kill, I dive into the section of SEC filings (https://www.microsoft.com/en-us/Investor/sec-filings.aspx ), and there, a pleasant surprise awaits: they end their fiscal year by the end of June, them people at Microsoft, and thus I have their annual report for the fiscal year 2020 ready and available even before the calendar year 2020 is over. You can download the report from their site or from my archives: https://discoversocialsciences.com/wp-content/uploads/2020/10/Microsoft-_FY20Q4_10K.docx .

As I grab my machete and my camera and I cut myself a path through that document, I develop a general impression that digital business goes more and more towards big data and big server power more than programming strictly speaking. I allow myself to source directly from that annual report the table from page 39, with segment results. You can see it here below:

Intelligent Cloud, i.e. Microsoft Azure (https://azure.microsoft.com/en-us/ ), seems to be the most dynamic segment in their business. In other words, a lot of data combined with a lot of server power, and with artificial neural networks to extract patterns and optimize. If I consider the case of Microsoft as representative for the technological race taking place in the IT industry, cloud computing seems to be the main track in that race.

Before I forget: IBM has just confirmed that intuition of mine. If you go and call by https://www.ibm.com/investor , you can pick up their half-year results (https://www.ibm.com/investor/att/pdf/IBM-2Q20-Earnings-Press-Release.pdf ) and their latest strategic update (https://www.ibm.com/investor/att/pdf/IBM-Strategic-Update-2020-Press-Release.pdf ). One fact comes out of it: cloud computing at IBM brings the most gross margin and the most growth in business. It goes to the point of IBM splitting their business in two, with cloud computing spinning out of all the rest, as a separate business.

I would suggest my students to think about digital innovations in the domain of cloud computing. Microsoft Azure (https://azure.microsoft.com/en-us/ ) and cloud computing provided by Okta (https://investor.okta.com/ ), seen a bit more in focus in their latest annual report (https://discoversocialsciences.com/wp-content/uploads/2020/10/Okta-10K-2019.pdf ), serve me as quick benchmarks. Well, as I think about benchmarks, there are others, more obvious or less obvious, depending on the point of view. You Tube, when you think about it, does cloud computing. It stores data – yes, videos are data – and it adapts the list of videos presented to each user according to the preferences of said used, guessed by algorithms of artificial intelligence. Netflix – same thing: a lot of data, in the form of movies, shows and documentaries, and a lot of server power to support the whole thing.     

My internal curious ape has grabbed this interesting object – innovations in the domain of cloud computing – and now my internal happy bulldog starts playing with it, sniffing around and digging holes, haphazardly, in the search for more stuff like that. My internal austere monk watches the ape and the bulldog, holding his razor ready, I mean the Ockham’s razor to cut bullshit out, should such need arise.

What’s cloud computing from the point of view of a team made of an ape and a bulldog? This is essentially a f**king big amount of data, permeated with artificial neural networks, run on and through f**king big servers, consuming a lot of computational power and a lot of energy. As cloud computing is becoming a separate IT business on its own right, I try to decompose it into key factors of value added. The technology of servers as such is one such factor. Energy efficiency, resilience to factors of operational risk, probably fiberoptics as regards connectivity, sheer computational power per 1 cubic meter of space, negotiably low price of electricity – all those things are sort of related to servers.

Access to big, useful datasets is another component of that business. I see two openings here. Acquiring now intellectual property rights to datasets which are cheap today, but likely to be expensive tomorrow is certainly important. People tend to say that data has become a commodity, and it is partly true. Still, I see that data is becoming an asset, too. As I look at the financials of Netflix (see, for example, The hopefully crazy semester), thus at cloud computing for entertainment, I realize that cloud-stored (clouded?) data can be both a fixed asset and a circulating one. It all depends on its lifecycle. There is data with relatively short shelf life, which works as a circulating asset, akin to inventories. It earns money when it flows: some parcels of data flow into my server, some flow out, and I need that flow to stay in the flow of business. There is other data, which holds value for a longer time, similarly to a fixed asset, and yet is subject to depreciation and amortization.

Here is that emerging skillset: data trader. Being a data trader means that you: a) know where to look for interesting datasets b) have business contacts with people who own it c) can intuitively gauge its market value and its shelf life d) can effectively negotiate its acquisition and e) can do the same on the selling side. I think one more specific skill is to add: intuitive ability to associate the data I am trading with proper algorithms of artificial intelligence, just to blow some life into the otherwise soulless databases. One more comes to my mind: the skill to write and enforce contracts which effectively protect the acquired data from infringement and theft.

Cool. There are the servers, and there is the data. Now, we need to market it somehow. The capacity to invent and market digital products based on cloud computing, i.e. on lots of server power combined with lots of data and with agile artificial neural networks, are another aspect of the business model. As I think of it, it comes to my mind that the whole fashion for Blockchain technology and its emergent products – cryptocurrencies and smart contracts – arose when the technology of servers passed a critical threshold, allowing to play with computational power as a fixed asset.

I am very much Schumpeterian, i.e. I am quite convinced that Joseph Schumpeter’s theory of business cycles was and still is a bloody deep vision, which states, among other things, that with the advent of new technologies and new assets, some incumbent technologies and assets will inevitably disappear. Before inevitability consumes itself, a transitory period happens, when old assets coexist with the new ones and choosing the right cocktail thereof is an art and a craft, requiring piles of cash on the bank account, just to keep the business agile and navigable.     

Another thing strikes me: the type of emergent programming languages. The Python, the R, the Pragma Solidity: all that stuff is primarily about managing data. Twenty years ago, programming was mostly about… well, about programming, i.e. about creating algorithms to make those electronics do what we want. Today, programming is more and more about data management. When we invent new languages for a new type of business, we really mean business, as a collective intelligence.

It had to come. I mean, in me. That mild obsession of mine about collective intelligence just had to poke its head from around the corner. Whatever. Let’s go down that rabbit hole. Collective intelligence consists in an intelligent structure experimenting with many alternative versions of itself whilst staying coherent. The whole business of cloud computing, as it is on the rise and before maturity, consists very largely in experimenting with many alternative versions of claims on data, claims on server power, as well as with many alternative digital products sourced therefrom. Some combinations are fitter than others. What are the criteria of fitness? At the business scale, it would be return on investment, I guess. Still, at the collective level of whole societies, it would be about the capacity to assure high employment and low average workload per person. Yes, Sir Keynes, it still holds.

As I indulge in obsessions, I go to another one of mine: the role of cities in our civilization. In my research, I have noticed strange regularities as for the density of urban population. When I compute a compound indicator which goes as density of urban population divided by the general density of population, or [DU/DG], that coefficient enters into strange correlations with other socio-economic variables. One of the most important observations I made about it is that the overall DU/DG for the whole planet is consistently growing. There is a growing difference in social density between cities and the countryside. See Demographic anomalies – the puzzle of urban density, from May 14th, 2020, in order to make yourself an idea. I think that we, humans, invented cities as complex technologies which consist in stacking a large number of homo sapiens (for some humans, it is just allegedly sapiens, let’s face it) on a relatively small surface, with a twofold purpose: that of preserving and developing agricultural land as a food base, and that of fabricating new social roles for new humans, through intense social interaction in cities. My question regarding the rise of technologies in cloud computing is whether it is concurrent with growing urban density, or, conversely, is it a countering force to that growth. In other words, are those big clouds of data on big servers a by-product of citification or is it rather something completely new, possibly able to supplant cities in their role of factories making new social roles?

When I think about cloud computing in terms of collective intelligence, I perceive it as a civilization-wide mechanism which helps making sense of growing information generated by growing mankind. It is a bit like an internal control system inside a growing company. Cloud computing is essentially a pattern of maintaining internal cohesion inside the civilization. Funny how it plays on words. Clouds form in the atmosphere when the density of water vapour passes a critical threshold. As the density of vaporized water per 1 cubic meter of air grows, other thresholds get passed. The joyful, creamy clouds morph into rain clouds, i.e. clouds able to re-condensate water from vapour back to liquid. I think that technologies of cloud computing do precisely that. They collect sparse, vaporized data and condensate it into effective action in and upon the social environment.

Now comes the funny part. Rain clouds turn into storm clouds when they get really thick, i.e. when wet and warm air – thus air with a lot of water vaporized in it and a lot of kinetic energy in its particles – collides with much colder and drier air. Rain clouds pile up and start polarizing their electric charges. The next thing we know, lightning starts hitting, winds become scary etc. Can a cloud of data pile up to the point of becoming a storm cloud of data, when it enters in contact with a piece of civilisation poor in data and low on energy? Well, this is something I observe with social media and their impact. Any social medium, I mean Twitter, Facebook, Instagram, whatever pleases, essentially, is a computed cloud of data. When it collides with population poor in data (i.e. poor in connection with real life and real world), and low on energy (not much of a job, not much of adversity confronted, not really a pile of business being done), data polarizes in the cloud. Some of it flows to the upper layers of the cloud, whilst another part, the heavier one, flows down to the bottom layer and starts attracting haphazard discharges of lighter data, more sophisticated data from the land underneath. The land underneath is the non-digital realm of social life. The so-polarized cloud of data becomes sort of aggressive and scary. It teaches humans to seek shelter and protection from it.           

Metaphors have various power. This one, namely equating a cloud of data to an atmospheric cloud, seems pretty kickass. It leads me to concluding that cloud computing arises as a new, big digital business because there are good reasons for it to do so. There is more and more of us, humans, on the planet. More and more of us live in cities, in a growing social density, i.e. with more and more social interactions. Those interactions inevitably produce data (e.g. #howcouldtheyhavedoneittome), whence growing information wealth of our civilisation, whence the computed clouds of data.

Metaphors have practical power, too, namely that of making me shoot educational videos. I made two of them, sort of in the stride of writing. Here they are, to your pleasure and leisure (in brackets, you have links to You Tube): International Economics #3 The rise of cloud computing [ https://youtu.be/FerCBcsGyq0], for one, and Managerial Economics and Economic Policy #4 The growth of cloud computing and what can governments do about it [ https://youtu.be/J-T4QQDEdlU], for two.