Skip to content

An Idea Is a Maze, Not a Point — Which Is Why the Same One Makes One Founder and Breaks Another

Hand two founders the identical idea and you get opposite outcomes, because an idea is not a point you possess but a maze of decisions with dead ends and hidden doors — and one of them has already walked it.

By Mehdi8 min read
Share
On this page

Hand the identical idea to two founders and you will often get opposite outcomes, and the usual explanations — one executed better, one got lucky, one had more money — miss what actually happened. The real difference is that they were not handed the same thing at all. One received a point: a sentence, a destination, a "wouldn't it be great if." The other received a maze: a branching structure of decisions, each with dead ends and a few hidden doors, most of which look identical from the entrance. The first founder is standing outside reading the map aloud. The second is already three turns in, and knows why the corridor everyone runs down first ends in a wall.

The "idea maze" is Balaji Srinivasan's and Chris Dixon's framing, and it is one of the few pieces of startup vocabulary that earns its keep: a good idea, they argued, is not a lightning bolt but a well-thought-out map of the decision tree, where you can reason about which paths lead where because you have studied who walked them before. I want to take that further, because the frame does more than describe good ideas. It explains why "having an idea" is a category error, why the most common objection to a startup is usually wrong, and it hands you a concrete test you can run on yourself before you commit a year of your life.

The value is in the walls, not the destination

Start with what an idea actually contains, because the naive accounting is where everything goes wrong.

The tourist's model treats an idea as a possession: you either have it or you don't, and once you have it the remaining work is "just execution." Under that model, the idea is a single valuable atom and secrecy makes sense — someone could hear it and take it. This is why first-time founders are paranoid about NDAs and why the paranoia is misplaced. If the idea were a point, it could be stolen in a sentence. It can't, and the reason it can't is that almost none of the value lives in the sentence.

The value lives in the structure the sentence points at. Take any idea seriously and it immediately fans out into a maze: which customer first, which wedge, priced how, sequenced in what order, depending on which assumption being true. Each of those is a fork. Most forks are dead ends — not obviously, or nobody would take them, but subtly, in a way you only learn by walking down and hitting the wall, or by watching someone else hit it. The accumulated knowledge of which forks are dead, which sequences work, and which walls are load-bearing is the idea. The destination is the cheap part. The map of the dead ends is the expensive part, and it is expensive precisely because you can only buy it with research, scars, or lived experience.

This reframes the most common objection a founder hears. "Someone already tried that" is treated as a kill shot, and it is almost always the wrong objection. Someone trying it tells you a path exists; it tells you nothing about which path they took or why it ended where it did. Webvan and Instacart had the same one-sentence idea. Someone tried grocery delivery and it was a generational bankruptcy; someone tried grocery delivery and it went public. The sentence was identical. The maze was walked completely differently — different sequencing of markets, different position on whether to own the warehouses, different bet on gig labor that only became available after a specific cost curve moved. "Already tried" collapses a maze into a point, which is exactly the error the whole frame is warning you against. The right question is never did someone try it. It is which cell did they die in, and do you know why.

Founder-market fit is a starting position inside the maze

If the value is accumulated path-knowledge, then the founder who has lived the problem does not start at the entrance. They start halfway in, holding a partial map they acquired for free — by suffering the problem before they ever decided to build for it.

This is the mechanism underneath founder-market fit, and it is why founder-market fit predicts outcomes better than product-market fit: it is a leading indicator precisely because it measures how far into the maze you begin. A generic founder enters at cell zero and pays full tuition on every dead end. A founder who spent five years inside the industry has already watched three companies charge down the obvious corridor and splatter against the same wall, and they know which wall and why. They have not skipped the maze. They have pre-walked a chunk of it on someone else's payroll.

I felt this concretely building Kommerce, an operating system for cash-on-delivery commerce in low-trust markets. The one-sentence idea — "software to help COD sellers run their business" — is a point anyone can hold, and plenty of well-funded teams have held it. The maze is where it gets lethal. The obvious first move is to optimize the checkout, because that is what commerce software does in trust-rich markets. That corridor is a dead end here, and you only know it is a dead end if you have watched the actual failure mode: in COD, the money does not exist until a stranger opens their door days later and decides to pay cash or refuse the parcel. The load-bearing assumption is not conversion; it is the delivery-success rate and the working capital trapped in parcels in transit. A founder who has lived that starts several cells past the one where outsiders confidently wall themselves in. Nobody could have handed me that by explaining the idea. It was only legible from inside.

The point generalizes past my case. Lived problem-knowledge is not a credential you cite; it is a coordinate. It tells you and everyone else how deep into the structure you are standing before the building begins.

A good pivot is a step, not a teleport

The maze also fixes the most abused word in startups. A pivot is not "we changed our idea." Under the maze frame, a good pivot is moving to an adjacent cell you can see from where you are standing — one you can reach because the walk that got you here revealed a door invisible from the entrance. A bad pivot is teleporting to a completely different maze and starting over at its entrance, throwing away every dead end you had already mapped and calling the amnesia "fresh energy."

Slack came out of a game studio: the adjacent cell was the internal communication tool the team had already built and mapped while failing at the actual product. That is a step. The founders could see the new cell because they were standing next to it. Contrast the pivot that abandons the entire hard-won map to chase a hotter market the founders know nothing about — that is not a pivot, it is emigration to a new maze where they are, once again, a tourist at the entrance. This is the same failure in different clothing: most pivots fail because they keep the wrong thing, discarding the validated path-knowledge and carrying forward the artifact. In maze terms, they keep the code they built in the old maze and throw away the map, when the map was the only asset that survived the move.

How to tell a walker from a tourist

All of this converges on a test you can run in a single conversation, and it is brutally discriminating. There are two ways to talk about an idea, and they sound nothing alike once you know what to listen for.

The tourist pitches the destination. They describe how great it is at the end of the maze — the TAM, the vision, the world where they've won — and they hand-wave the path with words like "and then we scale" or "once we get distribution." Ask them why the obvious version fails and they don't have an answer, because to them the obvious version is the idea. Ask them who tried this before and they say "nobody, that's the opportunity," which is almost never true and always tells you they haven't looked.

The walker does the opposite. They can explain, unprompted, why the obvious approach is a dead end — and they get specific about the wall. They can name the failed predecessors and dissect them: not "grocery delivery is hard" but "this company burned itself owning warehouses before demand density existed, that one underpriced the labor, this one nailed the ops and starved on the wrong first market." And they can articulate the one specific, non-obvious bet they are making — the door they believe is real that the consensus reads as a wall. Destination-talk is the tell of the entrance. Dead-end-talk is the tell of someone inside.

The concrete move follows directly, and you should run it on yourself before you commit, while it is still cheap. Map your maze on paper. Write down who tried the adjacent paths and exactly why each one ended where it did — and if you cannot fill this in, stop, because you have not done the research, you have had a feeling. Name your load-bearing assumptions, the two or three claims that if false collapse the whole structure. Locate at least one non-obvious wall: a place the market confidently expects a door and there is none, or expects a wall and you believe there is a door. Then state, in one sentence a competitor could not have written, why the obvious version of your idea is a dead end. If you can write that sentence, you are inside the maze. If you can only describe how good it is at the exit, you are standing at the entrance holding a map, mistaking it for the territory — and so is the next founder handed the same idea, which is exactly why only one of you will get out.

An idea is not the thing you have. It is the thing you have walked.

Frequently asked questions

Isn't this just a fancy way of saying 'execution beats ideas'?
No, and the difference is the whole point. 'Execution beats ideas' treats the idea as trivial and the doing as everything, which is why it collapses into a platitude. The maze frame says the idea is not trivial — it is dense with information you can acquire before you execute anything. What separates founders is not raw effort applied to a known plan; it is how much of the maze they have mapped, which failed predecessors they can name and dissect, and which non-obvious wall they have located. You can out-execute a rival on a path that was a dead end the whole time. Knowing the path is dead is the asset, and it is upstream of execution, not the same thing.
How do I map the maze if my idea is genuinely new and has no failed predecessors?
It almost never does, and 'no predecessors' is usually a research failure rather than a fact about the world. If nobody visibly tried your exact thing, widen the frame: who tried to solve the same underlying customer pain by other means, who owned an adjacent workflow and chose not to extend into yours, which enabling condition (a cost curve, a regulation, a behavior) had to change for this to be viable now. The relevant predecessors are rarely companies with your exact pitch; they are everyone who stood in an adjacent cell and turned away. If you truly cannot find a single one, treat that as a red flag about your search, not a green light about your originality.
Doesn't demanding a fully mapped maze just reward over-analysis and punish fast movers?
Mapping the maze is cheap and fast relative to building — it is weeks of research and conversation, not a paralysis ritual, and it is the opposite of over-analysis because it is aimed at a decision. The goal is not a complete map; complete maps do not exist and the maze reconfigures as you move. The goal is to have walked far enough in that you can state why the obvious version is a dead end and name the specific non-obvious bet you are making. A founder who can do that in three weeks moves faster afterward, not slower, because they stop paying tuition on dead ends the market already priced.

Filed under Business & Strategy. How durable advantage is actually built — and lost.

Essays like this, in your inbox.

Thoughtful essays. No spam. Unsubscribe anytime.

Business & Strategy

Your Byproduct Is More Defensible Than Your Product

The most defensible revenue a company has is usually the exhaust its product throws off — data, audience, trust: near-zero for you to accumulate, your entire operating history for a rival to reproduce.

8 min read
Business & Strategy

The Offer Is the Strategy: You Can't Out-Spend a Commodity

An ad is a multiplier, not an engine. It amplifies whatever your offer already is, so a commodity offer times a big budget is just an expensive failure — and the real work is upstream of the spend.

8 min read