Robots and Bicycles

Robots automate. Bicycles accelerate. Which best describes your market offering? 

In conventional economic wisdom, the role of technology is to enhance productivity: machines allow us to create more output with less labor. 

Closer to the ground, we can see two ways this productivity manifests. One mechanism is the (near-)complete replacement of variable labor. These are the presses, looms, and robots that comprise the modern factory. These machines often require some human staffing, but the necessary skill sets are not the same ones replaced by the machine - much artisanal craft is eliminated.

A second mechanism for improving productivity is the bicycle. In a previous era, an essay like this would have pointed at the lever, which allows the application of mortal strength to massive endeavors. But a lever is just mechanical. The bicycle has become part of many humans' lives, starting in childhood. Learning to ride a bike is a rite of passage that reveals the amplification of our natural talents. We were faster and could go farther. The two-wheeled marvel would reward us for working out muscles. Our muscles would get larger, and we achieve ever-higher speeds and distances. The bicycle is not merely a classical machine: it brings joy. 

From the start of the industrial revolution to maybe thirty years ago, the primary focus was on the productivity of bodies., In particular, the advent of factories caused a massive displacement for even skilled workers. The effort required to produce goods went down on a unit basis: the very definition of productivity. 

The information technology revolution changed that. The focus went from giving us leverage on our bodies to leverage on our brains. Software ate the world by taking over this previously human domain. Robots and bicycles form a helpful dichotomy for looking at what software does for our minds. 

Robots remove concerns. When we have a robot working for us, the mental load lightens. For example, automation can remove much of the "work" of paperwork. Terminals present information relevant to our workflows. Computer networks route the relevant data to other workers and on-demand storage without any human intervention - all at the speed of light. 

Tools have taken over transcribing the thoughts of professionals (e.g., shorthand dictation), "typing," and other tasks previously assigned to secretaries and clerks. Automatic reminders replace our need to check a calendar, and instantaneous information retrieval means we don't have to memorize. 

Bicycles extend our capabilities. Steve Jobs famously described the personal computer as a "bicycle for the mind," allowing us to accomplish more than we ever have before:

As entrepreneurs and investors in these frothy times, we would be well-served to ask whether we offer our customers robots or bicycles. Each offers a different economic and personal relationship.

Robots are loveable for what they take away from our concern. We tell Google to find information for us, and we hand off communication to email. Any number of automatic systems respond to the first order customer requests, so they get their answers faster and without requiring interruption by employees - a win-win. 

We fear robots for the same reason. At the start of the 19th century, craftsmen losing business at their workshops formed a secret society, the "Luddites," whose name we invoke today about those who oppose technological transition. They brought us another word, based on their direct action against the machines: throwing their cheap shoes - sabots - into the gears, they gave us "sabotage." 

I recently spoke with a consultant in Singapore who reported that cleaning crews were performing almost-as-crude actions on information processing units on the night shift. One wonders if they were former clerical or white-collar workers displaced by the changes discussed above. 

Bicycles are loveable for how they make us feel and where they let us go. These accelerants make us more remarkable than what we would be unassisted, and the more we use them, the higher the achievement. 

They also have some strategic implications that are worth a moment's meditation. First, one can only ride so many bicycles because they return value proportional to my time investment. Since I have only so many waking hours, I am limited in which I choose to ride. 

In contrast, I can utilize as many robots as I like, composing an environment where an army of helpers removes my concerns. As this system of helpers grows, managing the robots will itself become a concern. But that cost will lag the scale of benefit. 

A significant number of the new software offerings I see in the market - mainly SaaS - are bicycles. They ask people to spend time in their app, and the language of a lot of product management encourages this path. They want the customer to "love" their app. And so they focus on making a great experience for "using" it. These are appropriate for a relationship with a bicycle. The more you love it, the more you ride it, and presumably, the service enables more value creation.

Zoom back from the perspective of a single product to a marketplace full of vendors and products. As a customer, I can only ride so many of them because they require my time investment. As such, I want a ride to enable my most important work. 

What is the most important work? Geoffrey Moore gives us a language for thinking about this: core and context. In his framing, "core" is creating differentiated value that wins customers. For a law firm, this might be legal research and argument. Context is, quite literally, everything else, no matter how mission-critical it might be. 

A bicycle should maximize return on my limited time. As a result, I will pick the concise set of tools that best accelerate my "core." If that applies to one customer, expand that to the whole market. The consequence is limited units of bikes sold - a short multiple over the number of customers. The supply shape could be a small number of SKUs sold to significant markets or a large number sold to micro-segments. Either way, the customer's attention - the finite number of items one can focus on in the limited hours in the day - constrain the market's total size. 

Compare this with the market for robots. The necessary job of a robot is to remove a concern from my ongoing life. When it does this, it delivers value for me. I judge a robot based on the importance of that task and whether I worry about it anymore. A robot fails in its job when I have to refocus my attention on that task. 

As a result, robots take on tasks from my context rather than my core. Selling me a robot that does my job will challenge my identity and get much more scrutiny for quality, and I am looking for reasons to say no. On the other hand, sell me a robot that takes care of tasks that I previously spent time on but were not my special sauce. That robot I will hire for one or more of those tasks. 

So I can have a fleet of robots working for me, taking care of all those ancillary tasks in my context. I happily hire more of them to extend the list of jobs I don't need to worry about and focus on my core. 

From a market sizing point of view, robots scale. I can sell many different robots to the same customer to further eat away at those contextual concerns. Eventually, offering a robot suite or broader service to take a big piece of context away will remove a management headache from the customer.

SaaS companies offering robots need to make sure they completely take away a concern - and that concern can be relatively narrow. Further, those robots need to be composable so that task-handlers can collectively handle context. 

The smell for effectiveness will be the presence and robustness of APIs. APIs are how we manage that composition at a low level. To be effective, a robot offering needs a comprehensive API, including both commands - usually REST or GraphQL - and event triggering via webhooks. These probably also get wrapped up as integrations for the automation providers, such as Zapier and Integromat.

Customers can and do love their robots, but their relationship will be anti-engagement. They will depend on their robots to take care of context, but they will not want to "use" them. Flipping the script on traditional UX measures will help vendors capture the correct value proposition, driving sales. 

Many SaaS vendors that I have seen and spoken with are in the misfit stage. They sell bicycles for context jobs and wonder why users don't stay. They sell robots that don't altogether remove the concern and wonder why they are not more successful. 

Figure out which type of product you offer. Measure its value creation and market potential in the appropriate framing. When you start from this viewpoint, the following steps will reveal themselves. If more founders do this, they will have more significant successes, the market will expand, and we will enhance customers' lives, reaching for ever-higher productivity. 

Many thanks to the members of the community who gave me excellent feedback on this essay prior to posting. Errors are of course entirely my own.

Photo by Nicholas Ng on Unsplash