My “what does size mean?” adventure
Technological disruption has been a recurring theme in my career and in the instances where an organization I worked for struggled to respond to such change, its “size” was blamed; smaller organizations (e.g. start-ups) are seen as being more nimble. As I started thinking about this claim an observation about Toyota, a company that I had previously explored, came to mind:
Toyota, which isn’t exactly small, inspired the Lean movement as well as Agile. At the time of the publication of The Toyota Product Development System, Toyota’s time-to-market was 9 months shorter than that of its competitors.
This is one data point but it was enough for me to start questioning the connection between measures of “size” such as revenue or number of employees and outcomes such as agility. In fact, I have worked in organizations considerably “smaller” (by traditional measures) than Toyota that produced simpler products, but struggled to get a product in-market in 15 months (Toyota’s average time to market). So my next thought was that “size” depends on organizational structure or how technology is used, but I wasn’t happy with the imprecision of this idea. The above line of thought shaped a set of questions I ended up exploring in a series of essays:
What is organizational size anyway? How do we define it? What factors contribute to size? How do these factors shape organizational outcomes?
The below is my first essay exploring the above questions.
World Ward-2 — Political Science — Complexity Theory
In the below paragraphs, I explore how the scale (the forest or the trees) we organize collective action at, impacts organizational outcomes; namely the complexity an organization can handle. This capacity for complexity enabled me to develop a working definition of “size”.
To develop our understanding of scale and its interaction with complexity, I explored the below historical event:
How did the German army manage to defeat the combined forces of Britain, France, and Belgium in 6 weeks during World War-2?
The inspiration for using this question comes from James Q. Wilson’s Bureaucracy, which opens with the same line of inquiry; Wilson uses the above-mentioned battle (the Battle of France which took place in 1940) to highlight how differences in the “organization” (how collective action was coordinated) of the German and French armies was a major factor in the collapse of France.
I will be using the description of the differences in “organization” between the two armies to introduce the interplay between scale and complexity based on my reading of Yaneer Bar-Yam’s work on complexity. The main resource used here was his book Making Things Work.
“Organization” as the key explanatory factor
In this essay, I rely on Wilson’s assertion that “organization” was the key explanatory factor of the outcome of the Battle of France. Below I summarize Wilson’s refutation of some common misconceptions about the German success (I thought we’d get these out of the way first):
- Was Germany’s attack a surprise? It wasn’t. Germany had already invaded Poland and Norway, and intelligence gathered by France pointed to an imminent attack.
- Was the German army bigger? No. the German army had fewer soldiers and it did not have as many tanks as France. The best French tanks were larger and more powerful than the German ones (they didn’t have radio capability, however, which meant it was difficult to coordinate their actions).
- Was technology a deciding factor? As late as 1943, a typical German infantry division had more horse-drawn vehicles than motor-driven ones (it required twice as many tons of hay and oats as oil and gasoline). Additionally, key breakthroughs in the battle were achieved by foot soldiers; they paddled rafts across the Meuse river and then had to climb up steep banks or dodge enemy fire across open fields.
- Was political fanaticism a factor? There is disagreement amongst historians about the role played by Nazi ideology in motivating soldiers.
- Was the German strategy superior? It wasn’t; it was very risky and had many vulnerabilities that were never exposed.
World War 1 & Trenches
To understand the organization of France’s army in World War-2, we need to examine the widespread adoption of trench warfare in World-War-1. By the start of World War-1 firepower technology ( machine guns and artillery) had become more advanced than mobility technology (e.g. tanks). This meant that you could protect a position by shooting at an enemy (that had to cross an open field on foot) from a stationary position and therefore put an emphasis on defense and the development of trenches.
The below video provides a more visual summary of the development of trenches (watching the 3rd segment — “Why were trenches used?” — of this video suffices for the purposes of this essay).
What France learned from World War-1
The below video from the movie War Horse shows how difficult it was to advance on enemy positions supported by machine guns/artillery. This was the main French takeaway from World War-1; they set up their army to fight a stationary war of attrition; they organized around a squad whose task was to fire, support, and defend a machine gun. This was reflected in the training of soldiers which put little emphasis on maneuver. This also led to large investments in advanced “trenches” like the Maginot Line. Most importantly, their assumption of a static war led to a hierarchical command structure that didn’t enable fast decision-making (decisions took days).
Germany’s context led it to different conclusions. The treaty of Versailles, which brought World War-1 to an end, restricted the size of the German army, which meant that they couldn’t afford a stationary war of attrition. They focused on figuring out how to overcome the trenches and came up with two solutions:
- Make soldiers bulletproof (by putting them in tanks) and;
- Make soldiers hard to shoot at by focusing them on infiltration as opposed to open-field battle.
The Germans decided to focus on infiltration warfare (Blitzkrieg); finding weak points in enemy lines, cutting off communication lines, and attacking machine guns from the rear. So Germany organized its army into small units armed with light weaponry capable of independent action. Officers were delegated substantial decision-making powers and incentives/evaluation systems were designed to reward fighting prowess and risk-taking. Responsibility for success and failure was shared by officers and foot soldiers and there was mutual respect between the two groups. Furthermore, the command system at the higher levels was focused on the “what” leaving the “how” to the lowest possible level of decision making. You can watch the “speed of decision making” segment of the below video to get a sense of the German army’s MO.
In the below paragraphs, we’ll use the above example to link “size” to complexity and to explore the interaction between scale and complexity.
Zooming In and Out
Imagine the French army generals using an interactive map to decide on how to organize their army; they would have zoomed out to a coarse scale, defined the enemy as a big unit, and concluded that stationary machine guns are a good option for aiming at a big target acting coherently (i.e. advancing on a trench). Additionally, they would have reasoned that a big unit cannot change direction easily, and didn’t make the speed of decision-making a priority in the design of their command structure.
The Germans on the other hand zoomed in. They didn’t want to give the French one big unit to shoot at and designed their organization at a finer scale. The independent units (coordinated by the “what” set by generals) were able to make decisions based on local knowledge (e.g. find enemy weak points) and act accordingly.
Size and Complexity
A simple way to think about the relative complexity of systems is to compare the amount of information it takes to describe them (for example: as measured by the number of characters). The French army was designed to act as a large coherent body protecting a position against another large coherent body. The resulting organization had limited options for action at a fine scale; machine guns were pointed in the same direction, at the field of battle (with the infantry servicing them). The preceding sentence is a reasonable description of the system, and its brevity suggests that the system had a low degree of complexity.
In contrast, the German army had more options for action at fine scales as it comprised a network of units that could make decisions based on the imperatives of the situation each faced. In order to describe the system, you would need to describe the possible actions for the individual units as well as how the units might interact. The resulting system, therefore, had a higher degree of complexity.
Once the Germans stopped acting like a big uniform unit, the French army struggled to cope; big machine guns aren’t particularly useful for preventing enemy infiltration. The mismatch in complexity at a fine scale directly impacted the outcome of the Battle.
In modern times, matching the complexity of military forces to their environment (including the opponent) is standard practice. The US Navy for example is organized to deal with a low complexity environment (Open water with a limited number of enemy forces) and is organized at a large scale (Big ships). On the other hand, the US marines who deal with more complex littoral environments (e.g. different places for the enemies to hide) are organized at a finer scale.
What does this all have to do with “size”?
It is my impression that when we talk about “size” (e.g. that organisation is clunky and slow) what we really mean is capacity for complexity; if there is a mismatch between the complexity of a system and its environment, there is trouble. An organization is “too big” if its complexity is lower than that of its environment.
What does scale have to do with “size”?
Let's visualize the relationship between scale and complexity in the below figure. Observe how simple behavior can maintain the same level of complexity as you increase scale. In the case of the French army, as you zoom out, you see more and more machine guns pointed in the same direction. You could aggregate their behavior as if the individual machine guns add up to a bigger one.
On the other hand, fine scale complexity is difficult to maintain as scale increases. In the case of the German army, coordination of an increasing number of units gets harder and harder as you zoom out. Note that the complexity mismatch at a fine scale inhibited the French army’s ability to deliver large scale behavior.
The point is that you need to manage the tradeoff between complexity and scale, ensuring there is no mismatch with the environment as you zoom in and out. The”Size” of an organization needs to be right at all scales.
(Note: above image isn’t based on data and is based on a similar one in Making Things Work)
An interesting example to explore here would be Airbnb vs. your big brand Hotel chain. Hotel chains are managed at a large (coarse) scale; this means that the design of hotels (e.g. the lobby, the room, the dining areas) tend to be very similar across locations. Airbnb on the other hand limits large scale design to things like finding and booking accommodation, where you can increase scale without losing complexity. It leverages complexity where it matters; the places you find on Airbnb are small and owned by individuals and therefore tend to be unique (lots of options). The cost of managing such a complex portfolio would be prohibitive for a big brand hotel (you can’t increase scale without losing complexity). The Airbnb model seems to be better where unique experiences (complexity) are required by customers (the environment).
So when we think about “size” we need to think about the complexity of the environment and how an organization addresses that level of complexity. In the next essay, I explore how the hierarchical and networked modes of organization impact a system’s potential complexity. I’ll be using examples from my own experience; a failed consulting project and a weird company with 5 CEOs.
In future essays, we discuss some sophisticated examples of how organizations such as Amazon and Spotify use a loosely coupled architecture/organization to create an organization that can act with high complexity at different gradations of scale.
If you liked this essay and would like to hear more from me, please follow me on Twitter.