CHAPTER 10

The Man-Less Business Model

Why did the 19th century economist David Ricardo change his mind about money and machinery?1 Economists still speculate about the reasons. David Ricardo is one of the most influential classical economists along with Thomas Malthus, Adam Smith, and James Mill. He was a self-made millionaire who made his fortunes on financial speculations about the outcome of the battle of Waterloo. He lived in times when inventions like the steam engine and the textile mills were transforming capitalism and developed a strong interest in the effect of machines on labor. He formulated the machine/labor substitution effect, that is, the ratio at which machines can replace human labor in the production process, which is at the core of the above-mentioned speculation.

Prior to 1821 Ricardo saw the introduction of machines as beneficial. He viewed the displacement of workers by machines as a temporary inconvenience and the benefit of machine production to lower prices as a permanent gain to society. Since lower prices made many goods affordable to lower income people, machines increased the overall standards of living. Ricardo believed that the demand for labor is permanent over time and, thus, the temporarily displaced workers will eventually find new jobs. Furthermore, since machines replace low skilled jobs, workers will be retrained for higher level and higher paying jobs. Who would want to be a taxi driver today, if they can get a better paying job and ride to work in an autonomous vehicle? Ricardo’s belief was based on his observations of how machines worked at the time. He noted that: “machining cannot be worked without the assistance of men, it cannot be made without the contribution of their labor.”2

But in 1821 he added another chapter to the revised edition of his book “The Principles of Political Economy and Taxation” in which he changed his mind:

… I am convinced, that the substitution of machinery for
human labour, is often very injurious to the interests of the class of labourers.

My mistake arose from the supposition, that whenever the net income of a society increased, its gross income would also increase; I now, however, see reason to be satisfied that the one fund, from which landlords and capitalists derive their revenue, may increase, while the other, that upon which the labouring class mainly depend, may diminish, and therefore it follows, if I am right, that the same cause which may increase the net revenue of the country, may at the same time render the population redundant, and deteriorate the condition of the labourer.3

Ricardo’s new belief is that the substitution effect does not necessarily benefit society and that the displacement may not be just temporarily. The substitution effect is a rate. The faster the increase in the rate, the larger the profit. Today, this is known as the marginal rate of technical substitution, or the rate at which an investment in automation has to increase in order to keep the same level of production taking into account the displaced labor. Once a rate is defined, humanity will do anything to accelerate it and bring it to its logical conclusions—complete substitution. Perhaps Ricardo changed his mind because he could not avoid the logical conclusions of his own argument. But, as one economist discovered, Ricardo might have changed his initial observations on how machines operate and how technology will evolve:

In a letter to J. R. McCulloch written in June 1821, Ricardo wrote: “If machinery could do all the work that labour now does, there would be no demand for labour. Nobody would be entitled to consume anything who was not a capitalist, and who could not buy or hire a machine.” (Ricardo 1951–1973 8: pp. 399–400)4

Ricardo foresaw that a day will come when machines will neither be operated nor be created by man. The invention of autonomous machines will lead to complete automation and to new men-less business models.

The tracking of rates typically indicates that the future is closer than we think as all effort and investment will go toward accelerating the rate. In 2016, I was invited to Lockheed Martin to brainstorm together with 90 other executives about a Condition-Based Maintenance (CBM) system for military ships. Equipment maintenance is typically done based on a predefined schedule. Every 12 or 18 months, ships are brought back to the base and many parts are serviced or replaced according to the predefined schedule. Today, most cars alert the drivers to go to the dealership for certain types of scheduled maintenance. Each type of maintenance requires the service or replacement of certain parts even if it is not needed. Once the alert goes on the dashboard, the dealership mechanic does what is required in order to turn off the alert.

But what if the ship can sail for another year without maintenance? What if you could drive another 10,000 miles? It will certainly save you money. The length of operation without defects depends on many factors and conditions, while the scheduled services and parts replacements are based on the average expected time in service without considering the variations in operating conditions. The problem with such averages is that some machines break before the scheduled maintenance and others work much longer. In both cases unnecessary costs are incurred. If a ship breaks at sea, either a crew of experts has to fly to repair it or the ship has to be brought to the shore for the repair. On the other hand, if the scheduled maintenance is not really needed, the cost of early service and parts replacement is unnecessary. Condition-based maintenance is a data-driven alternative to scheduled maintenance as it delivers repairs at the point of need, that is, right before the equipment breaks. Similar to the just-in-time inventory management practice, CBM leverages data and advanced analytics to predict and deliver just-in-time services and repairs and eliminate unnecessary costs and waste.

The moderator asked us to imagine what would a military ship look like 25 years from that day. Once we had this idea, we could think backwards and design a CBM system that could be implemented with today’s technologies, but also be extensible over time to meet the needs of the future ship without a complete redesign. We all agreed that the future ship would be men-less. Why put crews at sea for many months away from their families? Why risk human lives in military operations? Any rational commander would like to minimize risks and inconveniences, and best way to do that is to design a men-less ship.

One obvious suggestion was to build a remote-controlled drone ship. But a ship is a large piece of equipment sailing many miles away. Many unforeseen things can happen and interfere with the remote-controlled operations. Hence, the future ship had to have autonomy and make decisions similarly to how a crew of specialists would decide and act in certain circumstances. But what if some system on the autonomous ship breaks down while it is 10,000 miles away. Coming to shore for repairs is costly and delaying the repair exposes the ship to environmental and military risks. The future ship must be able to self-diagnose and self-heal itself. Self-diagnosis requires a robust CBM system to prevent possible equipment failures just-in-time, while technologies like 3D printing, robotic automation, and virtual reality enable autonomous or remotely assisted repair. All people in the group acknowledged the validity of the concept—a men-less, autonomous, self-diagnosing, and self-healing ship. We all thought how powerful the vision was. Why couldn’t we have it sooner?

The rate of technological evolution is not a straight line. At some periods it grows exponentially and what seems a faraway future comes faster than we expect and often catches us unprepared for the social and other consequences. On April 25, 2018, the first drone ship joined the U.S. navy.5 The ship is crewless, 140 ton, 132-foot-long autonomous sailing robot. The concept of the men-less ship was formulated in 2010 and just in six years it became a reality. It was deployed for testing in 2016 and released for service in 2018. At the time when we were discussing the distant possibilities for a men-less, autonomous ship and flattered ourselves with our thought leadership, the concept was already near its completion.

It is interesting why the army may prefer a completely crew-less ship instead of one with a few crew members who can react quickly to any emergency with the remote support of highly qualified experts. This is a general question about why we strive to develop completely autonomous machines versus hybrid technologies where humans and machines can work together to achieve the desired outcome. The rate of substation is one directional.

On July 2019, The Atlantic published an article titled “At Work, Expertise Is Falling Out of Favor.”6 It reviews an army experiment to operate a highly sophisticated multipurpose warship with significantly reduced crew. The ship could hunt submarines, sweep mines, enter combat operations, and much more. A ship of even less complexity typically required a crew of 200 highly specialized professionals, but the new ship had only 40 “hybrid” crew members. The shortage of personnel was supplemented by advanced intelligent technologies that made it possible to do more with less resources. As the author of the article pointed out the ship was a complete departure from the 240 years of management principles and operating traditions. The “hybrid” crew was essentially a band of jack-of-all-trades that replaced the traditional masters-of-one-trade. With the assistance of the smart technologies crew members could perform multiple highly specialized tasks. As a result, idle time could be eliminated, and the crew could be reduced by 80 percent. This new management paradigm became known across industries as “minimal manning.”

Why didn’t it work? There were many well-known trade-offs such as the ship’s low survivability, which required the crew to abandon it in some cases of emergency, exposing them to significant risks. But it boiled down to unforeseen human factors. A smaller crew had more tasks to do. Hence, they had time just to react to problems without investigating and giving them due consideration. As the researchers discovered the more a crew member spent time on investigation and consideration, the lower the productivity of the crew members. Thinking can be expensive on the assembly line! To the surprise of everyone, the “hybrid” teams were not only unable to solve big problems, but they also failed to properly route tasks. One crew failed to oil the main engine causing the ship to return to base for repairs that cost the military $23 million.

The lack of expertise led to a lack of accountability. As everyone was able and tasked to do everything, no one was directly responsible and accountable for anything. Many small and big mistakes at peace time showed the dangers of combining minimal manning with smart technologies, and thus, the strategy was abandoned. Creating optimal coordination between less skilled men and smart machines is much harder than the development of autonomous machines. To achieve beneficial coordination men and machines have to be “equal” in expertise and sophistication as is the case with robotic surgery, which we will discuss later in this chapter. Augmentation of lower level skills with smart technologies can lead to more negligence on part of the employees as they begin to delegate more of the “thinking” and decision making to the machines.

As we make advances to full automation, two intermediate business models have emerged: Product-as-a-Service (PaaS) and Service-as-a-Product (SaaP).

Product-as-a-Service

Why would a traditional equipment manufacturer like GE want to become a software company? GE manufactures and sells large-scale expensive industrial equipment. But recently they changed their strategy and made a significant investment to develop the Predix software platform in an effort to become one of the largest software companies.

All industrial equipments have an expected life span. These expectations are used to forecast and plan maintenance and new sales revenues. Some machines have a longer- and others have a shorter life span than the average expected life span. All machines are built in the same way and have identical parts. Thus, the differences in their life span are due to variations in the operating conditions.

Equipment manufacturers and their customers have exactly the opposite preferences with respect to machine life span. Manufacturers want shorter life spans as this leads to more sales, and the customers want longer life spans as it reduces their capital investments. Manufacturers gain when the actual life span is less than the average expected life span, and customers gain if it is higher. Since the interests of manufacturers and customers are not aligned, there are business opportunities to monetize the gap.

Numerous innovative companies have developed specialized equipment monitoring and predictive maintenance solutions to optimize the operating conditions and beat the average machine life expectancy. Such companies charge a fraction of the equipment replacement cost for each year of extended service. Over time the manufacturers noticed that their sales cycles had become longer. The new analytics and monitoring services were in effect eroding their margins. The only way to combat this situation is by selling not the equipment but the outcomes of their products. When the customer buys the outcomes of a machine and not the machine itself, the interests of manufacturers and customers become 100 percent aligned. Since the outcomes are not dependent on the life span of the machine, the new PaaS business model eliminates the life span arbitrage opportunities that existed between manufacturers and owners.

The concept of selling the service instead of the product itself is not new. In 1962 Rolls Royce started a marketing campaign for its aircraft engines under the tagline “power by the hour.”7 This was essentially a maintenance service at fixed price per flying hour. However, in the absence of detailed operational data the problem of moral hazard arises. This is a situation when one party engages in a riskier than usual behavior knowing that another party will carry the costs if something goes wrong. Once the cost is fixed per flying hour, some operators may undertake riskier flights that increase the wear and tear of the equipment or lead to sudden breaks. The collection of detailed data solves this problem as it can demonstrate any violations of the operating conditions. Today, fleet management companies build driving profiles of the individual drivers that can pinpoint how sudden acceleration or sudden breaking increases the wear and tear of tires and other parts of the trucks and increases the cost of maintenance.

In 2017, Rolls Royce and Nor Line, a Norwegian shipping company, announced a completely new service level agreement based on big data and analytics.8 It was announced as a new kind of a “power by the hour” agreement, as Rolls Royce had developed a complete sensor-based remote monitoring of all cargo vessels which gives them full transparency into each ship’s operations and conditions. Each ship is being treated like a Check Point Cardio’s remote monitored patient that we discussed in Chapter 3. The complete transparency into the ships’ operations makes the fixed price fair as the manufacturer knows best how its equipment should work and be serviced.

Service-as-a-Product

“There are many more people who want to be healthy and fit but who do not go to gyms than those who go to gyms”—Min Kim, founder and CEO of Wise Wellness (wisewellness.co.kr) told me. Kim lives in Seoul and shows me from his office window how many people do physical exercise on the sidewalks in front of their office buildings. He explains to me that these are people who find a few free minutes during their busy days, walk outside, and do a few simple exercises for 5 to 10 minutes. Some of them do it several times per day. This is so normal in Korea that pedestrians do not even pay attention to the people doing stretches, jumps, and other exercises, which we may find weird in New York.

Kim explains that people need 2.5 hours of exercise per week according to a U.S. army research paper. The physical exercises do not have to be done at once or in 30 minutes increments as we do when we go to the gym. All people need is a few minutes of moderate intensity anaerobic activity per day. There are many reasons why people do not follow this simple recommendation. Lack of motivation is the key factor for not exercising. It is also the reason why people hire personal coaches or join gym classes.

Min Kim’s idea was to make these short exercises as easy and convenient to do as it seemed to be for the people who do them on the streets and in the parks of Seoul. No need for wearable devices, no trips to the gym, and no live coaches. What people got out of the various exercise devices and facilities had to be packaged as an app in the one device that no one in Seoul could live without—the smartphone.

To replace the fitness instructor the smartphone had to be able to monitor accurately what a person does, assess the intensity of the activity, track progress over time, and provide advice. Kim packaged the services provided by gyms and fitness instructors in a SaaP mobile app. To make it simple Kim focused on the five basic exercises—squats, push-ups, sit-ups, short climbs, and walking. All exercises, except walking, require visual monitoring to prevent cheating.

You may wonder if people cheat about how much they exercise. They do. I have seen many people shaking their Fitbit devices vigorously to make it up to the daily steps goal. You can even purchase devices for steps faking if shaking a device is too much of an exercise. People cheat for different reasons. Some want to win contests or impress others, others want to prove that they comply with prescribed exercise routines, and many just lack the motivation to exercise but do not want to acknowledge their weakness of the will. Life coaches keep people honest. And this is exactly what Kim’s app does. It recognizes the physical movements associated with each exercise, assesses their correctness, and calculates the goal achievement. Companies in Korea provide the app as part of their benefit packages to help employees stay fit and healthy.

Min Kim is not the only entrepreneur who has packaged a services-as-a-product. Software chatbots also provide services-as-products. The men-less service is significantly cheaper. Min’s app retails for $1.50 per month, while the average hourly rate for a fitness coach is $20. By making coaching affordable and exercise convenient, Min is driving more people to do it. Similar to location-based apps, Min’s app can be used anywhere and anytime. Affordable and scalable service is at the core of the proliferation of this new technology and data-driven business model.

Opportunity Matrix for the Man-Less Business Model

The man-less business models, regardless of whether they are PaaS or SaaP, are built entirely on data and analytics. The opportunities can be mapped to two dimensions—the autonomy of the machine and the completeness of the tasks that it does. Autonomy is the degree to which the machine can make decisions without human assistance. The second dimension reflects whether the machine does the entire job by itself, or a subset of the tasks within a larger process. In the latter case, the machine is a component within a more complex system. The matrix in Figure 10.1 shows the different categories within the two dimensions.

Figure 10.1 Monetization opportunity matrix: Man-less model

The autonomous car and the autonomous ship will perform the entire job without human assistance. While we are not 100 percent there yet, the end goal is a completely autonomous and independent machine. A fully autonomous car will take passengers from point to point safely without a human driver, and a fully autonomous ship will complete a mission without a crew and without being operated by a remote control. The market and the monetization opportunities are huge as autonomous machines provide a 24/7 service without the human cost. This in turn allows companies to scale the service to many more users and deliver it at any time. Giving autonomy to machines has economic, social, and moral consequences and we are not yet fully aware of all issues or the solutions to the known issues. Since the business opportunities exist and the expected profits are huge, the transition to man-less machines will only continue to accelerate despite the concerns.

Today many retailers have deployed store robots to monitor various aspects of the retail operations. These awkward-looking “employees” can perform many tasks—monitor the available stock on the shelves, store cleanliness, security, and much more. But the store robot cannot do it all. Hence, the need for a hybrid workforce. The adoption of these machines does not depend solely on the cost savings. It will depend on how they fit with other employees and customers. Awkward machines can turn customers off and send them to more traditionally managed stores. The winners in this space will invent new human machine interaction paradigms that will make the autonomous machines more acceptable to humans.

Shadow Robot Company (shadowrobot.com) is a UK-based company that has been working for years to create the perfect robotic hand. Such technology can do many useful things. The robotic hand can pick up strawberries or roses, play ping pong, and even replace the human hand in the case of full or partial hand loss. The robotic hand has a lot of built in intelligence. Like the human hand it has not only dexterity, but also sensitivity that allows it to determine if a fruit is ripe or not, if the grip is too strong to break an object or too weak to hold it. However, the hand is only a component of a larger system. All the information collected by the hand feeds the “brain” of a larger autonomous process that completes the entire job. The same applies for chatbots as they perform limited functions within larger systems. While the robotic components do not have complete machine autonomy, as they cannot complete an entire job, their intelligence is limited to the task at hand. The monetization of components depends on the demand and adoption of larger systems and is analogous to the economics of parts suppliers in the automotive industry.

A surgical robot, like the DaVinci robot and its competitors, can perform with its mechanical arms all the tasks that a surgeon does—cutting, stitching, and so on. But despite its mechanical complexity, the robot lacks full autonomy and requires a human surgeon. I asked Dr. Stoyanov who heads the computer vision lab at the University College London to explain to me why we needed this expensive equipment in addition to the surgeon. It appears that we have added to the cost of the surgeon the cost of the robot and raised the overall cost of surgery. He told me that there were many reasons. I asked for the simplest benefit that anyone unskilled in the art of robotic surgery will understand instantly.

“Imagine an open-heart surgery,” he started. “It is a big opening. There is a lot of blood and exposed flesh. This invites many bacteria to feast on the patient that can cause deadly infections.” I got it right away. The three small keyholes through which the robotic arms perform the surgery eliminate the banquet opportunities for the bacteria. Of course, there are other benefits such as the real-time feedback to the surgeon from the analysis of the streamed visual data that reduce deadly mistakes. In time this data and analysis will allow the surgical robot to become more autonomous. The market for augmentation of human skills with robotic equipment is as huge as the market for amplifying the human physical skills with industrial equipment that caused Ricardo’s initial fascination with machines.

Today, all four quadrants are blue oceans for innovation and new business models development, and yet many people fear automation.

What Should We Be Afraid Of?

There is a lot of anxiety and social fear about automation. The topic is so broad and complex that it will take decades of research and practice to resolve it. But in general, humanity has resolved many such dilemmas in the course of action, that is, as the events occur, and not beforehand in an academic style policy discussion. From this perspective, we should take a pragmatic approach to our fears. We should be afraid of the present negative trends and not of future imaginary effects. It is the present that shapes the future, and if it takes the wrong course the negative effects are likely to occur. From this perspective, we should not fear meaningful automation; we should be afraid of meaningless automation.

The pursuit of meaningless automation starts with the digitization of activities and processes without measurable improvement the human condition, the completion of tasks, or some other outcome. As engineers try to show improvement of outcomes in failing automation projects, they frequently begin to constrain the human activity in order to improve the machine productivity. This is the situation when machines and computer programs become “prescriptive,” that is, they dictate to human beings how things should be done instead of facilitating the doing of things. This can easily be seen in the case of chatbots that force us to go through a meaningless sequence of questions and answers to get to the point. Human conversations can take many paths, but chatbots often constrain us to a single path that wastes time and frustrates us. Why should I have to answer 20 questions before getting connected to a live representative?

In his book “The Glass Cage: Automation and Us,” Nicholas Carr discusses at length the drawbacks of “prescriptive” automation. He summarizes the negative effects in the following way:

The danger looming over the creative trades is that designers and artists, dazzled by the computer’s superhuman speed, precision and efficiency, will constantly take it for granted that the automated way is the best way. They will ignore the tradeoffs that the software imposes on them without considering them. They will rush down the path of least resistance, even though a little resistance, a little friction, might have brought out the best in them.9

New ideas and innovation come out of the friction and difficulties that we encounter as they trigger learning. The danger of meaningless automation is that by constraining the process it forces us not to learn but to submit to the machine. To avoid the Ricardo’s trap, automation should inspire and force new and higher level of learning.

---