Mental Models on the Entrepreneurial Journey
#1: First Principles, Analogical Reasoning, Scientific Testing and Systems
--
Introduction: Company-building and Mental Models
The typical accelerator organizes its program around the ideation of a business idea, followed by the incubation phase and scaling.
The generation of ideas is usually accompanied by heightened excitement and impatience. An idea never looks more perfect than until it is put into practice, in the incubation phase, where you first get into contact with reality through testing and experimentation. Iterating fast and cost-efficiently has no recipe, it is an art form to master on the go until you can sell to a broad market and scale operations.
The underlying logic of ideation, incubation, and scaling has proved as a reliable way to systematize building companies and has been adopted, with minor adjustments, from some of the most prominent and successful accelerator and company-building programs globally.
In addition to the hands-on advice from industry experts and startup veterans, this categorization of phases also inherits and teaches a new way of thinking that is essential to entrepreneurial success. Each phase emphasizes a distinct pattern of thinking and evaluating information to deal with the messy and chaotic environment you plunge yourself into when you start a company.
Patterns of thinking that guide an approach to solving problems are often called mental models. These models are simplified representations of the most important parts and relationships in a given situation, an internal representation that is good enough to support problem solving.
Mental Models can take various forms and the most useful ones provide a mix of algorithmic structure and creative flexibility to approach a variety of problems. The following series of articles aims to give an in-depth investigation of some of the mental models that are useful for successfully ideating, incubating, and scaling a venture company.
One of the main criticisms of mental models (and the articles about them) is that they are rarely actionable since most of them are descriptive. It is intellectually stimulating to scan your environment (or the market you are targeting) for appearances of dynamic equilibria, feedback loops, etc. However, that is seldom enough to give you any practical advice on how to use that knowledge to your advantage. Therefore, my focus in this series is on mental models as methods of thought with proposed frameworks for a particular phase in the company-building process rather than on the accumulation of descriptive models from disciplines like physics, chemistry, or economics.
After the first overview in this article, you will find links for the deep-dives at the end. Let’s go down the rabbit hole of mental models.
Ideation of Ideas: First Principles vs. Analogical Reasoning
“An idea is nothing more nor less than a new combination of old elements, and the capacity to bring old elements into new combinations depends largely on the ability to see relationships“— James Webb Young
Retrospectively, humans tend to organize the past into tidy narratives. Groundbreaking ideas are viewed as lightbulb moments of clarity and discovery. Suddenly, the light switch is turned on, and a unique connection of neurons firing for the first time in sync lets a revolutionary idea pop into consciousness.
The idea of mental models stands contrary to that illusion. Ultimately, it proposes that even in the messy, wild combinatory process that we call creativity, we can use models for thinking to impose regularity on phenomena otherwise hard to understand, structure, and interpret.
One such way to support your idea generation process, that reportedly some of the great entrepreneurs of the last century have used, is reasoning from first principles. The core idea is to break down problems into their foundational propositions, assumptions, or even spare parts.
With the original problem laying bare before you in its most basic components, you can re-assemble the parts in a new way, configuring unseen connections and thus forming previously unaccessible solutions.
Even if the theory is intuitive, several questions arise for applying first principles in the startup context: What is the big thing exactly that we should try to break down to generate ideas: an existing customer problem, product, or even just an aspirational goal that is not yet actively pursued because no one thinks is possible? What is the appropriate level of depth for breaking down a domain (Elon Musk reportedly uses the laws of physics)? I will discuss different ways to answer these questions for application in the deep dive.
While first principles are usually thought of as the gold standard for idea generation, what comes way more naturally to us is reasoning through the creation of metaphors, a process I will summarize as analogical reasoning in this series.
Analogical reasoning, in its essential form, means encoding regularities You assert that something you observe in A is similarly true for B.
In venture capital, you often hear propositions such as “we are the Uber of X”, a metaphor expressing the statement that features of the Uber business model and industry apply to the new venture; it establishes a mapping between both concepts that instantly increases understanding. Analogical reasoning, thus, is applied every time a venture brings a new technology or trend from one industry to another.
Some analogies are naturally more useful than others, for both generating an idea and communicating the essence of it to potential investors. As we will see, the extent to which a formed metaphor generates insights in a business context is mostly related to the accuracy of the mappings between the two entities chosen. Less formally stated, you have to form good metaphors to get good ideas, and we will dive deeper into a potential framework on how to do that.
First principles and analogical reasoning, in idea generation, serve the purpose of initiating a flow of thought and support creative insight. Especially when looking back at how an idea has formed, the boundary between both models is blurry. Is Space X really the result of first principles or an analogy mapping a traditional transportation service from the ground into space?
There is no objective way to answer this question as it clearly depends on perspective. However, staying true to the topic of mental models and adopting a mindset of pragmatism, this should not matter after all. If the whole point of mental models is to support your decision making, giving both models a shot and then arguing about the specific origin of your idea later, preferably after your IPO or monster funding round, should be the way to go.
Calibrating the Setup: Scientific Thinking
“It does not make any difference how beautiful your guess is, it does not make any difference how smart you are, who made the guess, or what his name is — if it disagrees with experiment, it is wrong” — Richard Feynman
When you calibrate a model, you update the value of the model’s parameters through inferences as new information becomes available. Bayes theorem accomplishes this by weighing new evidence against prior probabilities, hopefully paving a way to refine the model to more accurately represent the relationships it tries to replicate or predict.
It might seem strange to propose that the exciting, rocket-ship-like trajectory of the next hot startup should be related to the previous two sentences that jumbled together various abstract terms most of us hated in math classes. But if you peel back the layers of making progress towards the objective truth, or in the business context, the external reality of what the market wants, you will find that it often resembles exactly those principles. And wherever there are principles that (almost) accurately illustrate part of an overwhelming complexity in a simplified and compressed form, the resulting mental model can enable better problem solving.
Incubating a startup is similar to a modeling approach in that it also has parameters that are configured for which the optimal input values have to be found. Where regression has independent variables, your startup has a product, a marketing strategy, distribution channels, and infinitely many other parameters which stand in relation to another. Finding the optimal input values to best fit the model to the underlying phenomenon might be analogous to iterating products, strategies, etc. as to best fit the opportunity in the market.
As an example, a specific composition of channels, messages, etc, will combine to the optimal customer acquisition strategy for a given market environment, thus resulting in the optimized input variable for your (startup) model.
If finding and adjusting these input values is the engine for startup success, scientific testing (and thinking) is the gasoline.
The classical scientific method is structured in distinct stages that taken together might form an iterative cycle. First, you make an observation and ask questions about its origin and causal mechanisms that led to its unfolding. Through this thought process, you develop a hypothesis, a proposed explanation for the observation. A hypothesis about how something works allows you to make a prediction and test if that prediction will come true in reality. Results of testing will either falsify or support (but not confirm) your hypothesis and serve as new evidence to refine it and start the process again.
The relevance of this method for surviving and thriving in an environment dominated by uncertainty can hardly be overstated. Brilliant works like the entrepreneurship classic “The Lean Startup” by Eric Ries center around its application for managing a startup. The book even claims that as little as 5% of entrepreneurship is the big idea that is so often at the focus of attention, while the remaining 95% entail the ability to rigorously measure, confront the hard truths, and iterate effectively.
Applying scientific thinking in the incubation phase includes unpacking the idea you have come up with, transforming it into falsifiable hypotheses, and finding the sweet spot between iterating fast and gathering valuable intelligence. The goal is not to question your idea in total (at least not right away) but to let empirical evidence shape its edges.
After the imaginative and exciting ideation phase, the process of then finding the setup of key variables is usually less pleasant. It can feel like drinking your morning coffee and thinking it will be a great day before stepping out into cold rain on a commute with way too much traffic. The mental model we will dive in deeper and try to complement with a bayesian mindset in the separate article can help to see these things in a rather probabilistic perspective to take some of the hardship out of it.
But in any way, at the end of incubation, when the model is tuned, what awaits is Product-Market fit, a popular and elusive term introduced and re-interpreted many times by some of the smartest people in the industry.
Trying to define a generalizable point at which a startup has reached PMF is debatable, it rather can be thought of as a flexible corridor that might differ in its position and size depending on business models and industries.
Nevertheless, reaching PMF for everyone implies the next transition on the entrepreneurial journey: the experimental scientist becomes the system engineer.
Engineering the Hockey-stick: Systems
“Reality is made up of circles but we see straight lines” — Peter Senge
The formal definition of a system as “a set of things working together as parts of a mechanism or an interconnecting network” makes the concept itself not very information-bearing: seemingly everything can qualify as a system once it has various components that interact to form a whole. Similarly, approaching an ever-expanding list of to-dos and little disasters to fix with concepts like feedback loops, bottlenecks, margins of safety, or redundancies, on the first view, is counterintuitive.
However, this kind of thinking is deeply immersed into the idea of scaling and the famous hockey-stick curve, even if the terms used differ. As a quick side note, any iteration done in the incubation phase has already introduced a feedback loop and any form of strategy executed can be formulated as an internal system setup. But the relevance becomes even more striking if one looks more closely into Product-Market Fit (PMF) and what it means for the mode of thinking from an entrepreneur’s perspective.
The concept of PMF was introduced by Marc Andreessen as “being in a good market with a product that can satisfy that market” and since then has been subject to many new framings and interpretations. What most of these definitions stress is the alignment between a product’s value proposition and market needs.
Thus, when adopting our previous analogy, you could say that PMF is the point at which your product variable is accurately calibrated to the external reality that is your target market.
Naturally, once the testing and iteration cycles have led to the desired result, the focus shifts from learning to systematizing. This shift, marked by PMF as the tipping point, has special meaning not only from a mental models perspective but also for signaling to investors. Experimentation is costly because falsified hypotheses usually result only in gained knowledge and not in gained customers or revenue. In contrast, reaching PMF optimally means that new capital is directly converted into measurable top-line results instead of iterations, which makes an investment far less risky and thereby more attractive.
From an entrepreneur’s perspective, system thinking as a mental model becomes valuable because it shifts your perspective for problem solving and decision making from a micro to a macro lens.
Instead of asking how you can tweak some product features to better appeal to customers, you are concerned with coordination and management as to best extract value from your gained knowledge. If problems occur, instead of attempting to fix one isolated part, you focus on the underlying system and thereby hopefully prevent future problems as well.
An intuitive example of the benefit of applying this mental model is considering growth through the lens of feedback loops. A feedback loop is a closed chain of causal connections that either has a stabilizing or a reinforcing effect on a system. Reinforcing feedback loops are self-enhancing, they mimic the compound interest that pays on your savings (well, in a healthy economy) and that keeps getting more powerful with each loop. Considering the above quote from Peter Senge, when you metaphorically x-ray the hockey stick, a curved line with a steep upward slope, you will see that it is made up of many small reinforcing feedback loops.
Thus, re-framing a typical question in scaling like “how to create customer growth” to a system perspective becomes “how to engineer reinforcing feedback loops in customer acquisition”.
Equivalently, optimizing marketing becomes identifying bottlenecks in acquisition pipelines. Production and operations topics extend to discussions regarding redundancies and margins of safety. Even if some of these re-framings might not yield surprisingly different answers than your first intuition, regularly applying them re-centers your focus and prevents the tendency to micromanage and lose efficiency in decision-making as a company grows in size and complexity.
While ideation and incubation were concerned with creative insight and rigorous calibration, respectively, scaling is about engineering long-term benefits and longevity. A system well-engineered and maintained runs smoothly, accumulates deliberately, and after some time, as juggernaut systems like Amazon or Facebook can attest, becomes nearly impossible to stop.
>> Let’s connect on LinkedIn <<
Series Links:
#1: Mental Models on the Entrepreneurial Journey
#2: Ideation of Ideas: First Principles → Coming Soon
#3: Ideation of Ideas: Analogical Reasoning → Coming Soon
#4: Calibrating the Setup: Scientific Testing → Coming Soon
#5: Engineering the Hockey-stick: Systems → Coming Soon
