How to Extract Structure From the Mess of Life
And Why It Matters to Startup Founders and Anyone Trying to Change the World
They say intelligence is what you do when you don’t know what to do.
How do you take a mess and break it down into discrete parts you can try to understand? Extracting structure from a formless heap is the first step toward solving a problem. It’s really trying to figure out the terms by which you should pose the question. It’s the beginning of your theory of the problem. What is in front of me and how should I begin to think about it?
What are some problems that need you to extract structure?
Research and inquiry. For these, you need to formulate and reformulate questions, and determine what kind of information would count as good, bad or indifferent, and then figure out how to get it.
Reasoned argument. Anybody who wants to communicate a problem and potential solution in order to rally support has to show what is happening, why it’s bad, and how it might be solved.
In a way, the first bullet point describes how to persuade yourself of something, the second describes how to persuade others.
People pay a steep price when they are unable to extract structure and apply it to search and persuasion. In the first case, they end up not knowing why they chose a painfully suboptimal result; in the second, they fail to gain support for what they want and believe in.
Fund-raising and job-seeking require one or both activities — search and persuasion — and failing at either will impact your ability to make a living, to be fulfilled on the job, and to build something greater than yourself.
I’ve lived, witnessed (and sometimes participated in!) egregious examples of people not extracting structure from the mess of life. Venture capitalists see a lot of decks and emails from people trying to persuade someone to give them money. Many are not driving home the points they need to raise funds, while many others are predicated on assumptions that a better search would have refuted or refined.
But I’ve also seen people miss critical chances to persuade others to change their behavior when lives were literally on the line; i.e. there were problems that resulted in death, and convincing others to make a different decision could have reduced the death.
So I’m not talking about how to write an idle essay where a college professor, who pulls their salary from your tuition, gives you a grade that won’t get them in trouble with the dean. This is high stakes, and if you want to succeed at anything, you should master these skills. (If you are still in school, you should try to write more compelling essays, knowing that life will demand it later…)
Since this post looked at search, I’ll focus now on making a persuasive argument. Warning: Obvious points below! But for those who were never exposed to high-school debate frameworks, or didn’t have a strong composition instructor, or who have simply forgotten, maybe these ideas will be useful.
So let’s talk about how to extract structure when making a compelling argument designed to convince someone of something. What can you convince them of? Three things:
Something is the case - Here you must demonstrate that something is happening, or happened. What are the facts of the matter? While we all know that predicting the future is hard, it is equally difficult to predict the present, or the past; that is, to know what actually happened. Establishing that requires evidence and argued proof, and even then your certainty will often settle below 100%. Call this descriptive.
Something is good or bad - Here you need to convince people that something which happened or could happen is, say, a problem. This is a separate claim than that something is the case. Everyone in America agrees that things like immigration and abortion are actually happening, but they differ radically in the value of those activities. Why should they think that something is bad or good? This is sometimes a moral claim (it’s just wrong), and sometimes a pragmatic one (this will hurt you/us if it continues). Call this normative.
Something should be done - After you have established that something is happening and it’s a problem, you often have to propose a solution. This type of claim argues “here’s what we should do about it.” Often a solution will be one point in a solution space containing many alternatives. Why is your solution the best? Why do you think your team can build it better and faster than any one else? An investor may well be convinced of the problem, and even of the solution, but hesitate about why a certain team should be backed, or whether any given solution is defensible and differentiated over the long-term (i.e. hard to copy and offer at a lower cost). Call this prescriptive.
Most of the time when people communicate, all three of these aspects are mixed up and sometimes they are unspoken. If someone says “I’m working on sustainable manufacturing,” one piece of their the unspoken argument is that “sustainable manufacturing is good,” the normative piece, and another piece is that “most manufacturing is not sustainable,” the descriptive piece. If they are saying this to you at some founder-investor event, then the prescriptive piece is “you should invest in my startup to solve this problem.”
It’s good to boil things down to a compelling elevator pitch, and to do that, you have to leave several pieces of your argument implied. But as soon as you have time to expand your presentation on, say, a 30-minute Zoom call, then you should be ready to argue and provide evidence for each piece of it.
You may find some investors willing to take pieces of the argument on faith, but you will surely find others who want you to step them through the facts, the judgment and why the proposed solution is the best. Investors who back founders purely because they’ve wrapped themselves in an ideal may never raise their second fund. The cleantech debacle of the late 2000s and early 2010s is an example of what results from that sort of unquestioning bubble.
As with most problem-solving, it’s good to figure out what your goal is and then walk backwards. Your goal in making an argument in a pitch deck is to get someone to believe something, to conclude that something is true. What is that conclusion?
Prescription: You should invest in startup Z
What claims underpin that conclusion?
Fact: Startup Z has found a real problem
Fact: The market to solve that problem is large
Judgment: The team to solve that problem is better than any other
Judgment: The solution they propose is better than other approaches
Fact: The solution is technically possible
What evidence underpins each of these claims?
Each claim demands its own proof, or ensembles of data points to establish the facts of the matter, and the judgment that say one technical approach is better than another (e.g. because it costs less and integrates well!).
So you have this fractal structure of claims underpinned by other claims underpinned by evidence at each leaf in the tree, the evidence itself being an argument that something is the case. Each claim and proof will hook into some other part of the argument like the loops of a knitted sweater (this analogy gives an intuition as to why “text” and “textile” are etymologically related). Arguments are woven, and the best ones are woven very tightly.
If you’re ever trying to improve an argument, ask yourself how any particular claim, piece of evidence, or sentence, hooks into the sentences that come before and after, and how that group of sentences fits into your larger flow from start to finish. Diagram it on paper! (The brain is no place for serious thinking, as someone once said.) Some of the strongest hooks are causal: This is true because that is true. A pyramid of causes is how you walk someone through your vision in a compelling way. A series of broken links is a sign that you have not thought deeply enough, or need more practice with your pitch.
Any of these claims and proof points can also be weak points, or steps in your argument that don’t stand up under scrutiny. The nature of investor due diligence is to examine each claim and the underlying evidence, and figure out if it makes sense. Most startups never get past very shallow due diligence like “we don’t believe in the market”, because deep due diligence is incredibly time consuming. VCs are looking for reasons to say “no” early, so that they can go deep and get to yes with a pitch that ticks all their boxes.
The most important thing you can draw from this post is that a structured argument can and should be extracted from the mess of data at your disposal, an argument which connects established facts with the conclusions that will trigger action and agreement by the people you present to. If you’re not doing that, you’re sabotaging your own chances to make something happen.
I hope that the work of extracting structure will reveal to you some new questions that you need answers to so that you can be more certain of your own course, whether it is building a business or saving lives (or both!). And once you have extracted that structure and resolved your own uncertainties, I hope you convince the world to believe you.