Subscription Capitalism: The Story of a Power Shift

The emergence of the internet changed the business landscape in fundamental ways. Computer-based services could be offered to anyone irrespective of geographic restrictions. This meant that internet companies could become globally significant with relatively little initial investment, as demonstrated by Facebook, Google, and several others. Their ability to rapidly concentrate wealth with relatively little overhead caught the attention of many entrepreneurs and investors and continues to evolve today.

Basic internet services began in the 1960s with time-sharing on computers. This gradually evolved into sharing entire applications online. It was a simple step from there for companies to offer complete online software services, some of the most prominent being full-service email such as Hotmail and Gmail in the 1990s.

Monetizing the internet, however, was not a straightforward proposition. When the internet first emerged, people saw it as a great democracy where anyone with a good idea could set up a business with little cost. This lowered barriers to entry and thus offered the promise of moving the economy closer to a pure “free market,” a system that better rewarded good ideas and merit over money or position.

Many free services appeared, not the least of which were search engines. When Google was first born, it was simply one of many search engines. In the process of becoming dominant, Google either bought out or made irrelevant all other search engines. Today, there are some people who think Google’s search engine is the internet instead of part of the internet.

Despite Google’s near monopoly, Google’s search engine remains free to use. At first glance, it seems puzzling that a large corporation would spend considerable resources to dominate a market and not use its place and power to force revenue from its users. This is because licensing software is not Google’s main money-making strategy. Google’s strategy requires large numbers of people to use its software, and the best way to generate traffic is to build high quality, free-to-use software packages that help everyone. On the surface, this aligns well with the widely held belief that the internet should remain free.

Although free products propagated across the internet, businesses never lost their desire to monetize the internet. At first, it looked impossible—how do you monetize free goods? To monetize such an environment, either everyone would have to agree to sell their software or no one could, with the exception of large, highly specialized programs such as Adobe Photoshop or (at the time) AutoCAD. There were efforts to create paywalls throughout the internet, but this largely failed, just as newspapers are rediscovering today. This forced companies to think of different ways to monetize their products. The process branched into two distinct paths depending on whether the customers were enterprises or consumers. Both these methods altered the basic business relationship from one of sellers offering property in exchange for money to one of sellers renting their property to customers. In other words, sellers are increasingly retaining the power and rights of property ownership while retaining its customer-derived income. This is changing the power dynamics between customer and seller.

The Growth of Cloud Computing

As soon as internet technology became sufficient to reliably handle high bandwidths, business-to-enterprise activity exhibited a marked increase. Successful marketing of online services to businesses largely succeeded because of demonstrated savings on the cost of doing business. These new enterprises have been classified into various categories, including Software as a Service (SaaS), Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Everything as a Service (XaaS), Enterprise Resource Planning (ERP), and Functions as a Service (FaaS). All these activities can be described as cloud computing, where a service (be it storage, software, or other service) is supplied via the internet.

Before cloud computing, every business had to set up their own independent onsite IT operations, often duplicating the IT functions of other companies. For example, large firms need software processes to manage, track, and report their sales activity. After the proliferation of fast internet connections and large shared servers, however, entrepreneurs saw that if they set up a secure online sales service package, they could offer the same or similar online solution to several companies. Development costs of IT functions are similar whether services are offered to one or many companies, so it is obviously more profitable to sell the service to many business customers.

The result, on the surface at least, is win-win—the seller makes greater profits while the businesses save money by not having to build their own IT functions. The customers receive leading edge IT services offered at a fraction of the cost. Further, should business IT needs suddenly increase, the cloud service is responsible for scaling the software project, not the business customer. The enterprise customer benefits from a lower cost of business and minimizes the responsibilities of managing IT resources as the business grows.

Another advantage of using existing services is certainty. For a business manager, purchasing a proven IT service removes considerable risk. An online journal, ZDNet, recently asked: “Would you rather host your data on an inflexible, costly and potentially unstable in-house resource, or would you rather work with a trusted external partner who is an expert in secure hosting?”¹ It is easy to understand how an experienced manager would be much more likely to reach for the proven software of a cloud service rather than ask his IT department to create something from scratch. This is borne out by the fact that the vast majority of cloud service contracts are initiated by business managers, not IT departments.

Ownership and Control

It is important to note, however, that once the IT department work has moved off premises, the business no longer owns that service or the infrastructure that delivers it. Instead, the company is investing in an intangible service from a centralized source. There are several consequences of this trend.

To an individual company, it might make no difference whether an IT outage is the result of an in-house IT problem or the service provider failing. IT issues were common before cloud computing. However, providers have different priorities than their customers. Consider Netflix’s experience with one of the largest cloud servers in existence, Amazon’s Web Services (AWS). Like other services, AWS has gone down several times since its inception. Whenever it does, it takes down thousands of other businesses with it, including big names such as Netflix, Instagram, Airbnb, and Reddit. On Christmas eve 2012, an AWS outage² caused Netflix to lose streaming, an event that some believe impacted Netflix sales for years subsequently. Worse, Netflix’s competitor, Amazon’s instant video service, remained online. The failure was caused by routine maintenance that had gone wrong.

Since then, Netflix has spent millions linking to an entire network of AWS servers and routinely tests the robustness of their network using their own custom software. The software, “Chaos Monkey,” randomly shuts down certain virtual machines, and “Chaos Kong” randomly shuts down entire regions to test streaming robustness. There is also “Armageddon Monkey” that simulates the loss of the entire AWS. In this scenario, even though it might take days, Netflix can recover from their backups. It took seven years of intensive effort, though Netflix declines to reveal how much they pay Amazon for their services.

While it is possible to create a robust cloud service connection, it requires a quantity of money and effort unavailable to most companies. In network terms, the bigger a cloud service node becomes, the larger the economic impact whenever it goes down. Further, whenever a cloud service outage occurs, there is always the potential for more widespread data damage. For example, the May 31, 2018 AWS outage³ ended with some customers having their backups erased from AWS. Many companies relying on AWS for their backups suddenly found themselves with either out-of-date data or none at all. Through no fault of their own, some businesses were forced to rebuild their web presence from scratch.

Replacing onsite capability with intangible services requires the customers to give up some control in the service design. Cloud services must link with the customer in some manner, meaning that all cloud services have two IT components—one from the customer and one from the service provider—that combine to create the full service. As cloud services must present a stable link to several businesses, the supplier is not going to change its core infrastructure to suit one customer. It is the customer’s responsibility to configure their IT to successfully link to the service. In the event of some failures, the customer can find themselves left to their own devices in trying to solve the problem, especially if the failure is specific to one company.

One way to overcome this problem is to contract the cloud service to build a custom service. However, the high failure rate of new IT projects does not change just because it is a cloud supplier. For example, consider the comprehensive cloud project “Connecting to Health,” designed for the United Kingdom’s National Health Service. Approximately £11 billion was spent before the project was eventually abandoned as unworkable.⁴

Property laws do not protect customers that have purchased intangible goods. Customers wanting similar protection add to their cost the time and effort necessary to ensure their rights are protected. For example, one way to mitigate against the downsides of purchasing intangible cloud services is to be diligent about negotiating a service level agreement (SLA). Whenever a cloud service fails to deliver, the SLA determines exactly where each party is responsible. A poor SLA can leave the two companies bickering over responsibility rather than getting to a solution, while a good SLA can head off problems before they appear. For the customer, the best SLA will likely be achieved through the customer’s IT department because that is where the IT knowledge lies. However, cloud service companies note that their services are usually commissioned by business managers looking for budget solutions rather than IT departments looking for technical solutions.⁵ Where the IT department is not involved, the integrity of the SLA can depend on luck and on the good will of the service provider, rather than the rigor of the customer. The result can be a poor or no SLA at all.

A poor SLA allows all the demons of uncertainty to appear in the event of a service failure. Did it fail because of customer specifications, poor implementation by the service, because of integration problems, etc. In the case of a temporary failure, the solution can be delayed as the service provider and the customer “negotiate” who is at fault. Poor SLA’s can be partly responsible for catastrophic failures as each side is making assumptions about the others perception of responsibility.

Another problem with poor SLA’s is they can make changing providers more difficult. For example, it is not uncommon for companies to make custom modifications to their cloud service. The service provider is the one writing the code and documenting (or not documenting) the changes. If for any reason the customer needs to modify the code or chooses to move to another cloud service provider, poor documentation might make any changes or moves difficult or impossible. This forms a natural barrier to exit, making it more likely customers will stick with their existing provider even in the event of declining customer service and/or the price increases. This, of course, benefits the provider. Though it cannot be said that they consciously do this, it is in the interest of cloud providers to “accidentally” set up their service in such a way that it becomes unexpectedly expensive for their customers to migrate away. As CIO puts it: “Cloud computing is evolving its own unique brand of vendor lock-in.”⁶

One of the main drivers for replacing operations with intangible services is cost savings. Nevertheless, apparent cost savings can be eroded by unintended productivity losses. For example, one common practice of businesses is to have their employees engage directly with a payroll service. For the administrative staff, this is simply one more thing to administer, but for the productive staff, it directly takes away from productive activity and replaces it with administrative activity. Further, the productive staff are unlikely to be familiar with administrative procedures.

Further, when a new operating system is implemented, old physical accessories can stop working. The result is equipment being made obsolete, not through lack of performance or functionality, but through lack of compatibility with the latest software. Sometimes, the physical equipment cannot be replaced at all because it is not made anymore. While the overall budget initially reflects cost-savings, inefficiencies are sometimes pushed down to the lower tier workers and later re-emerge in the income statement as an unexplained loss of productivity.

Regardless, as long as the adoption of cloud computing makes the company look more profitable in the short-term, most managers will be unconcerned about any long-term externalities. This is rational as their bonuses are dependent on today’s performance, not tomorrow’s costs.

From an economic point of view, the increasing size of cloud service nodes create an increasing stability risk. When all businesses had their own independent IT infrastructure, it was much less likely that large numbers of businesses would suspend operations because of IT issues at any one time. However, as cloud service becomes more concentrated, the likelihood of many businesses going down at any single time increases. If enough businesses go down in an interdependent economy, there is the possibility that much of the economy will stop whenever a large cloud service goes off line. A recent industry study suggests that a major cloud service failure lasting three to five days could cost the US economy $15 billion.⁷

Terms and Conditions

While businesses knowingly give up control in exchange for lower costs, this is not always the case with consumers. Making money from consumers through cloud services is frequently an entirely different proposition as consumers do not have running business costs. Google jumped ahead of their competitors by being one of the first to identify the extraordinary value of a new product made available by the internet.

Google’s genius was to position themselves as a broker rather than a seller. Brokers make money regardless of the performance of what is sold. What Google realized was that business would pay a fortune for a high number of high-quality customer profiles. All the time and money spent on developing their free search and email software was designed to outperform their competitors so that most people would use their free software. Once accomplished, Google would have one of the highest quantities of high-quality customer profile information available anywhere in the world.

The potential of high-quality customer profile information is far greater than most realize. Marketing departments use the same techniques that mentalist magicians use to appear to predict what people want or see, or imagine they see, to an extraordinary degree. The mentalist Darren Brown has repeatedly demonstrated doing this and how he does it. Once, he managed to get an advertising agency to come up a logo and catchphrase he previously put in a sealed envelope and left with them by using established marketing techniques. He described it as “turning the tables” on the marketing agency.⁸

Google and their business customers care about you as an individual no more than a fisherman cares about the individual life story of fish they intend to catch. All a marketing department needs to understand is your preferences in order to design the correct “lure” for you and people like you. Who you are as an individual is irrelevant. Facebook, Apple, and Microsoft are among those who also participate in this lucrative business model.⁹

Note that none of this is illegal or underhanded. Business have been changing preferences for profit since the beginning of trade. However, consumer awareness significantly lags behind the capability of opinion manipulation technology. Occasionally, the technology’s true capability reaches the public eye such as in 2016 when Cambridge Analytica’s use of Facebook profiles to manipulate voters’ choices accidentally became public knowledge.¹⁰ While these services may be intangible, they can have tangible impacts on our lives.

The profit in this intangible sector can also lead to incidental manipulation. For example, the BBC reported a group of teenagers living in Macedonia discovered if they paid Facebook to advertise their sites to people with specific preferences, their articles (usually plagiarized from US websites) became popular and the website owner subsequently made a lot of money.¹¹ Their activity aided the spread of misinformation in the United States to the benefit of Donald Trump’s campaign (this is not a comment on the Donald Trump as a person or president).

One wonders how willing consumers are to have their opinions manipulated if they were aware of the degree to which it is happening. Regardless, a degree of consumer sovereignty has been lost through the trend towards consumer intangibles being offered free online. The collection of high-quality customer profiles is so lucrative that Microsoft Windows 7, 8 and 10 is set up by default to send customer information back to Microsoft.¹² In this case, the customer willingly purchases and installs their own personal monitoring software on behalf of businesses. Microsoft’s official message is that monitoring consumers is in the best interests of the consumer as it results in the most relevant ads and the best tailored experience. Indeed.

This entire phenomenon has been recently labelled as surveillance capitalism,¹³ which is basically new economics of making profits from very detailed knowledge of you and your behavior. The purveyors of surveillance capitalism generally know how you behave and response to stimulus better than you do.

Another way intangibles have expanded into the consumer market is through subscription services. As customers have no built-in costs of doing business as enterprises do, companies have to convince customers that paying less money up front will save money in the long-run. This can only work if the provider first establishes repeat business. Amazon most successfully implemented this with Amazon Prime. Initially, Amazon Prime was the purchase of fixed cost shipping for a year (falsely advertised as “free” shipping). Overall, this was the same business model as Costco where customers pay for the privilege of spending money in Costco. Charging the customer up front to spend in a shop has two benefits: (1) the business receives a predictable revenue unrelated to merchandise sales, and (2) the customers, having a sunk cost, tend to go out of their way to shop where they have a sunk cost in order to maximize their benefits. In Amazon’s case, it also has the added benefit that Amazon no longer needs to be the cheapest on the internet.  As long as their cost remains below the cost-plus shipping of their rivals, they will appear to be the best buy on the internet even when they charge more.

This is usually seen as a win-win situation. The customer pays less for products and Amazon makes lots of money. But there is a shift in market power as Amazon drives out the competition. As in business-to-business software, once customers develop a dependency, prices can be raised on the intangible with relative impunity, especially if there are no equivalent alternatives. In 2018, Amazon boosted the cost of Amazon Prime in the United States by 20 percent, far above inflation.¹⁴ The rationale presented to the customer is that Amazon offers more services through Prime, such a video streaming, Kindle books, ad-free music, etc. Note that Amazon offered these services before the price rise and that Amazon dictates to their customers what is included in Amazon Prime. A dependent customer will end up paying for the extra intangibles whether they want them or not.

Another key trend in the development of the intangible economy is that the customers no longer fully own what they purchase. For example, both Samsung and Apple have been caught using firmware updates to significantly slow down their older phones.¹⁵

While what Apple and Samsung did was considered worthy of a fine, the next example was perfectly legal. Between 2006 and 2008, there was a struggle between Blu-Ray and HD as to which would dominate the market. Many consumers felt nervous about which format player to buy. During this time, some manufacturers sold players that could handle both formats, freeing the customer from having to be a prophet as to which DVD format would eventually dominate. Blu-Ray won. The people who purchased the dual format DVD players were pleased that their HD DVDs were secure. Then there was a firmware update. Newly purchased Blu-Ray disks insisted they would not play until the firmware update was downloaded and installed in these dual format DVD players. Wanting to watch new DVDs, the costumer did so—and then the previously dual format player stopped playing HD videos. The firmware had turned the DVD into a single-format Blu-Ray player. Further, there was no option to revert to the old firmware and the consumer could no longer purchase new HD format players. This meant the customer’s expensive collection of high definition DVD’s was turned into junk.

This is completely legal as set out in the terms and conditions. Note that the terms and conditions are written by lawyers solely on behalf of the business. Consumers have no say.

With software, the trend towards selling intangibles is more direct. Adobe, for example, no longer sells its high-end software at any price. Instead, you can license its high-end products through one of several payment plans. What this means is the end-user never owns the software. If the payment stops, use of the software immediately stops. For individuals and micro businesses (such as part-time professional photographers) this creates a number of problems.

In the past, if someone had an old computer running old software with peripheries, as long as parts could be replaced if they failed, it would continue to be viable. Sometimes things ran for decades without problems. However, if a licensed software goes through an update that requires a newer operating system, perhaps one that won’t run on the owner’s current computer, instead of a free update, the customer is faced with a high hardware upgrade bill or losing access to the software. This is not always possible. Further, as previously mentioned, upgrading the operating system can cause faithful peripheries to become inoperable and time spent on the core business must be diverted to learn the new software.

In the past, software was an owned product. The trend towards renewable licensing moves control from the purchaser to the seller. The more indispensable the software, such as Microsoft Office is to business, the more control a company gains over the licensing arrangement. Note that over time, companies make more money through licensing than they would get for selling it outright. This is disguised as an outright purchase is much more expensive than a year’s purchase of a subscription service. However, experts suggest that, on average, a subscription service will generate more revenue than an outright purchase that includes a service contract after the fourth year.¹⁶ Obviously, increased revenue will occur sooner for customers who do not normally purchase a service agreement. The reason subscription services are rapidly growing is that, over the long-run, businesses are making a lot more money for the same product. A trend where businesses make more money from the same basic product is a form of inflation and indicates a shift of power towards the business and away from the customer.

In an endless search for economic efficiency, two things can get overlooked. One is shifts in power. According to free-market economic theory, a meritocratic market necessitates all the power being in the consumers hands. This cannot be the case if businesses are controlling price through dependency and intangible benefits, or if businesses can repossess or redesign the purpose of a purchased product at their will. If the consumer is not the master, then the economy will not be acting as a meritocracy.

Intangible Goods and Property Rights

Computer-based services, or cloud services, have been defined as intangible goods.¹⁷ Intangible assets lack definition concerning ownership and responsibility. This unclear demarcation of ownership has allowed businesses to retain part-ownership even in goods that appear to be sold outright to their customers. This is contributing to a shift of power from the consumer towards businesses selling intangibles.

Property is wealth. An intangible is not. Converting sales from property to intangibles is a process that transfers wealth from the customer to the businesses. This impacts wealth distribution in society.

The total economy over one year creates a finite amount of wealth. That wealth then flows to either workers through wages or to wealth owners through profits. The workers and owners then spend a proportion of that wealth. When looked at in aggregate, practically speaking, workers spend nearly all their wages. Therefore, whenever owners end up with more wealth than they spend into the economy, wealth necessarily accumulates in their hands.

Physically, an economy is a set of stocks and flows. In terms of long-term flows, there are only three possible states in the economy: (1) more wealth flows to workers than owners; (2) more wealth flows to owners than workers; or (3) wealth flows equally to both workers and owners. Except when the economy is in the last condition, wealth increasingly concentrates on one side. If the workers end up with a greater flow, factories will no longer have the money to make new things. The markets will fill up with affluent buyers that the markets will be increasingly unable to serve. If the flows favor the owners, then the markets will increasingly exhibit productive wonders that fewer and fewer can afford. In either scenario, the increasing stock imbalance weakens the overall economic flows until they reach a tipping point. Then the economy crashes.

Since the 1970s, multiple economic studies have shown that the worker’s wage share has fallen. In other words, wealth has been flowing to owners at an increased rate compared to the workers. This has created the inevitable erosion of purchase power for the workers and a sharp rise in inequality. The rise of intangibles being sold to consumers increases the existing imbalance. For the owners who see their profits rising, this is nothing but good news. However, this euphoria is as temporary as a good narcotic fix. It feels fantastic in the moment, but the long-term economic consequences are very serious indeed.

For long-term health, the economy needs to have a healthy consumer sector. Without consumers, the economy dies. To consume what the economy can produce, consumers need wealth proportionate to the owners. The trends in several rich countries have been to decrease the wage share. This alarming trend prompted a group of U.S. millionaires to set up a movement that petitions the U.S. government to raise their taxes.¹⁸

Note that the golden age of American capitalism is frequently considered to be the decades after World War II to the 1970s. During this time, the wage share was constantly increasing. As with most things in life, the best result comes from balance. To get back to this balance, wealth needs to be placed increasingly in the hands of workers. Any policy or trend reducing worker and consumer wealth will create problems in the future.

This article is an American Affairs online exclusive, published May 20, 2019.

---

Notes