First Principles Thinking: How to Solve Problems Instead of Copying Solutions

In 2008, SpaceX had lost three rockets in a row. Elon Musk was down to the last reserves of his eBay payout, Tesla was burning cash, and the consensus view in aerospace was that launch costs were fixed — $65 million minimum, maybe $100 million, and that was just the physics of the thing. Every serious person in the industry, from Boeing engineers to Lockheed procurement officers, would have told you the same. Rockets are expensive because rockets are expensive.

Musk did something almost nobody in the room was doing. He went to the periodic table. He looked up the cost of aluminium, titanium, carbon fibre and copper at commodity prices. He calculated the total raw-material cost of a Falcon 9 rocket: about 2% of the sale price. The other 98% was labour, tooling, overhead, profit margin, and — this was the expensive part — the cost of making something once, by hand, with bespoke processes designed by bureaucrats who'd never considered reuse.

That's first principles thinking. Strip the problem to its atoms. Build back up. Don't assume anything that can't be physically justified.

The Analogical Trap Most People Work Inside

The opposite of first principles is reasoning by analogy. It's how most decisions actually get made in large organisations. "The last company did X, so we should do X." "Our competitor is launching a chatbot, so we need a chatbot." "Everyone uses HubSpot for sales, so we should use HubSpot for sales."

Analogical reasoning is fast, cheap and usually good enough. It's also the default mode that locks whole industries into broken patterns for decades. Print newspapers reasoning by analogy about digital distribution in 2005. American car companies reasoning by analogy about electric vehicles in 2012. High-street retail reasoning by analogy about e-commerce in any given year since 1998. In each case, the industry's best-paid executives used perfectly good analogies to stay wrong.

First principles thinking is slower, costlier and harder to do in a committee. It's also how nearly every meaningful breakthrough in business and engineering has actually happened. Musk with SpaceX. Jeff Bezos deciding, from first principles of e-commerce economics, that AWS should become a standalone business. Jim Simons at Renaissance Technologies deciding that markets could be modeled mathematically when every serious fundamentals investor said they couldn't.

The Five Steps — and Where People Go Wrong

1. State the actual problem

Most people skip this step and pay for it later. The problem you're actually solving is rarely the problem stated in the meeting invite. "How do we improve customer retention?" is not a problem. "Why are customers leaving in month three, and what's the single biggest driver we haven't addressed?" is a problem.

A good test: if you can answer the problem with a generic best-practice, it's not phrased at the right resolution. Go deeper.

2. Identify your assumptions

This is where first principles work earns its keep. List, on paper, every assumption embedded in your current thinking. Then ask, for each: is this a law of physics, or is it a habit someone inherited from a previous situation?

Most assumptions, when you inspect them, turn out to be inherited. "We can't change the commission structure mid-year because salespeople will leave" — is that a law of physics, or is it what the last CRO told everyone because he didn't want to fight the battle?

3. Decompose to fundamentals

Break the problem into its smallest, indivisible components. For SpaceX it was raw materials. For a retention problem, it might be: acquisition cost, product-value-delivered, pricing, onboarding quality, customer expectation vs reality. None of these are opinions. They're the atoms the problem is actually made of.

4. Rebuild from the atoms

Now reason up, not down. Given only the fundamentals, what's the best solution you can construct, with zero reference to what others are doing? This is the step that tests whether you're actually thinking from first principles or just hunting for confirmation. If your answer matches the industry consensus, you haven't done the work. Go back to step three and look harder.

5. Sanity-check against the real world

First principles thinking can produce beautiful, logical, wrong answers. The physicist Richard Feynman had a line about the difference between the mathematics of a problem and its physics. Your rebuilt solution has to survive contact with reality — regulations, human behaviour, messy customer preferences, the competitor who will react.

This is where the Musk process actually gets its edge. He rebuilds from first principles, then spends a brutal amount of time stress-testing the rebuild against what engineers actually know about materials, weather, payload constraints. The first principles are the skeleton. The stress tests put the flesh on.

Where First Principles Thinking Is Actively Wrong

Here's the nuance most evangelists for this approach skip. First principles thinking is expensive and it's not always worth it. In some domains — most of them, actually — analogical reasoning is faster and just as good.

Ordering office supplies? Don't derive from first principles. Picking a payroll provider? Don't derive from first principles. Choosing a tax structure for a small LLC? Analogy is fine. These are commodity decisions where the cost of being wrong is low and the solutions are well-understood.

First principles thinking earns its cost in three situations: when the stakes are high, when the industry consensus is old, and when you're doing something with significant downside asymmetry. If all three apply, you'd be mad not to reason from fundamentals. If none apply, you're probably overthinking and you should just copy what works.

The Mental Habit That Beats the Technique

Past the technique, the real shift is a habit of questioning assumptions as a default posture. Charlie Munger, in his 2007 USC speech, called this "inverting, always inverting" — not first principles exactly, but a close relative. Instead of asking how to succeed, ask how to fail. Instead of asking what the right strategy is, ask what the disastrous strategy is, and then systematically avoid it.

This posture is simpler than the five-step process and arguably more useful. If you build a habit of, every week or two, picking one assumption in your working life that "everyone knows" is true, and spending 45 minutes pressure-testing it with a notebook and a coffee, you will, over five years, accumulate a small pile of genuine insights. Most of your peers will not. That pile compounds.

The Uncomfortable Part

First principles thinking produces answers that make you look strange in meetings. You'll find yourself arguing against positions held by everyone senior to you, with reasoning that's internally consistent but sounds, on the surface, like nonsense. This is the price. Musk spent the first three years of SpaceX being dismissed as a dot-com guy who didn't understand rockets.

The defence is brutal preparation. When you bring a first-principles answer to a room, you need to have done the math. Not "I think the unit economics work" — "the unit economics work because raw materials are X, direct labour is Y at the volume we're proposing, and the sensitivity analysis looks like this." Bring receipts. First principles thinking without receipts is just being contrarian.

Get that part right and you'll find, over time, that the rooms change. The people who used to dismiss your thinking start quietly stealing it. The decisions you've been arguing against start quietly reversing. It's slow and it's worth it.