A startup offered me 0.4% equity and a 30% pay cut. How do I think about this without lying to myself?
What the lenses had in common. Sharpest when a pattern names a tension you hadn’t fully named yourself — weakest when it summarises something you already know.
Xerox PARC created the future and captured none of it. Between liquidation preferences, two more rounds of dilution and a possible 90-day exercise window, your 0.4% has several independent ways to become nothing — none of them in your hands. Model three exits on paper before the number is allowed to carry any emotional weight.
From Xerox PARC The Ultimatum Game
Moneyball's move is to purchase the underpriced attribute. In a startup offer that's rarely the equity and almost always the scope: what you'll own, who you'll sit next to, what you'll be able to say you built. If the offer wins on those without the options, it's a real offer.
From Moneyball: Oakland's Statistical Edge
Scott reached the Pole and died coming back. Work out the exact month the 30% cut becomes unsustainable, and if it lands inside eighteen months, fix the cash or the fixed costs first. A pay cut you can't survive turns every subsequent decision into a fear-based one, which is precisely when people stay too long.
From Amundsen vs Scott Chanakya's Mandala
Often the sharpest provocation lives inside a single lens, not in the synthesis above. Skim them; one tends to land harder than the others.
People will burn the money rather than accept an insulting split.
Read this lensIn the ultimatum game, one player splits a pot and the other can accept or reject; rejection means nobody gets anything. Rational play says accept any non-zero offer. Real humans reject unfair splits routinely, and are right to — because the split tells you what the other party thinks you're worth, and you will be dealing with that opinion for years. Your 0.4% is not just a number. It is a statement about your standing, and you should read it as one.
Ask directly for the cap table band for your level, and the last preferred share price and total shares outstanding. A company that won't answer those has told you something more useful than the answer would have. Refusal is data.
In the classic experiment, one player splits a pot and the other can accept (both keep their shares) or reject (both get nothing). Rationally any offer beats zero — yet people routinely reject lowball offers, paying to punish unfairness.
The principle: Fairness isn't a soft nicety; it's load-bearing. Offer a deal that's efficient but feels unjust and people will torch the whole thing to deny you the win. Price the insult, not just the economics.
Invent the mouse, the GUI, the future, and let others sell it.
Read this lensXerox PARC invented the graphical interface, the mouse, ethernet and laser printing. Apple and Microsoft captured the value. The researchers weren't wrong about what mattered — they were wrong about who would own it. Joining a startup at 0.4% is a bet that you can create value and also capture a share of it. Those are two different bets, and the second one is decided by the cap table, the liquidation preferences and the next three funding rounds, none of which you control.
Model three exits — $20m, $100m, $500m — with the actual preference stack and two more rounds of dilution. Most people discover their 0.4% is worth less at $100m than they assumed it was worth at $20m. Do this on paper before you feel anything about it.
At Xerox PARC through the 1970s researchers invented the mouse, the graphical user interface, Ethernet, the laser printer, object-oriented programming. Xerox could not figure out how to sell any of it as a non-copier-company. Apple's Steve Jobs took a tour, saw the GUI, and built the Macintosh around it. Xerox watched.
The principle: Inventing the future is not the same as owning the future. The constraint is rarely R&D capability; it is the will and the org structure to ship something the current business doesn't recognise as its own.
Humble preparation beats heroic improvisation.
Read this lensAmundsen reached the Pole and came home; Scott reached it and died on the return. The difference wasn't courage or luck — Amundsen planned for the journey back, laid supply depots, and over-provisioned. Scott optimised for arrival. A 30% pay cut is a supply problem. The question isn't whether you can reach the startup; it's whether you can survive eighteen months there and still get home if it fails.
Calculate the exact month at which the pay cut becomes unsustainable at your current fixed costs. If that month is inside eighteen, either negotiate the cash up or lower the fixed costs first. Never take a pay cut whose failure case you haven't provisioned.
Two expeditions raced to the South Pole in 1911. Amundsen used Inuit-style dog teams, fur clothing, calorie-dense rations, and a meticulous depot strategy. Scott used Manchurian ponies, motor sledges that broke, manhauling, and tea-and-biscuit thinking. Amundsen reached the Pole first and returned without losing a man. Scott reached it second and his entire team died on the way back.
The principle: Glory accrues to whoever planned for the boring problems. Heroic narratives form around the things that went wrong — but the team that didn't need heroics was the one that read the manual before they left.
Buy what the market under-prices, not what the market wants.
Read this lensOakland won by buying on-base percentage, which was cheap because the market priced batting average and appearance. The move wasn't finding good players — it was finding the attribute everyone else was mispricing. Apply it here honestly: the equity is the thing everyone is looking at, and it is therefore the thing most likely to be priced correctly, or over-priced by your own hope. The genuinely mispriced assets in a startup offer are usually scope, title, proximity to the founders and what you'll be able to claim you built.
Write the offer down twice: once with the equity, once without. If the no-equity version isn't a job you'd want, the equity is doing work it shouldn't be doing — it's compensating for a role you don't actually want, and options don't fix that.
The 2002 Oakland Athletics had roughly a third of the New York Yankees' payroll. Beane and his analyst Paul DePodesta noticed that scouts and front offices were over-paying for batting average and visible 'tools' (speed, look-of-an-athlete) while under-paying for on-base percentage — the statistic that most strongly predicts runs scored. They built a roster of cheap players the market dismissed, on the back of a more accurate model. They won 20 games in a row and reached the playoffs.
The principle: Markets price what they understand. The asymmetric opportunity is what they don't. Find the metric everyone else is measuring wrong and you can buy real value at the price of optical mediocrity.
Map every actor as friend, enemy, or neutral — and act accordingly.
Read this lensKautilya's Arthashastra treats alliances as instruments of interest, not affection: understand what each party wants, assume they will pursue it, and structure accordingly. Nothing cynical in it — just clarity. The founders want committed senior people cheaply, which is legitimate. You want upside and experience, which is legitimate. Trouble arrives when one side pretends the relationship is a mission rather than a deal, because then asking for terms feels like disloyalty.
Get the post-termination exercise window in writing before you accept. A 90-day window on options you can't afford to exercise means the equity is theoretical for anyone who leaves. Ten years is the founder-friendly standard; asking for it is a normal, unembarrassing request that tells you a great deal about who you'd be working for.
Chanakya's Arthashastra, written around 300 BCE for the Mauryan court, maps statecraft as a series of concentric rings — your immediate neighbour is naturally your enemy; their neighbour, naturally your friend; the neutral state at the edge can be either. The strategist doesn't moralise about it; they read the geometry and act on it.
The principle: Most disputes happen between adjacent parties — same market, same field, same authority. The classmate, the competitor, the rival sibling. Allies are often one step removed. Before you treat someone as friend or enemy, ask where they sit in the geometry: adjacency tends to produce friction; distance tends to produce alignment. The map predicts the conflict before character does.
Lenses ranked relevant but kept off the exploration — the runner-up reading. Each links to the lens itself, not an applied reading of your question.
Sell the scaffolding, not the building.
Imagine the project has already failed. Now write the obituary.
When confidence becomes a design assumption.
Stop splitting the pie and ask if you can bake a bigger one.
Bind yourself in advance to the decision you'll later wish to break.
Invent your own killer and hide it in a drawer.
Master each mode separately; switch when the moment changes.
It solved the rail map with no brain at all.
drag · scroll to zoom · click anything