Quit Before You're Certain (Would I Start This Today?)
If you're thinking about quitting, you're probably already late — waste is prospective, not retrospective
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 92%
Duke's core quitting claim: by the time you're even considering quitting, you're almost certainly past the point where you should have. Sunk cost, endowment, and identity all bias us toward continuing until quitting 'isn't a decision anymore' — until the money's gone, the blizzard is on you, or you've already fallen in the crevasse. The corrective is a forward-looking test: knowing what I now know, would I start this today? And a reframe: waste is prospective, not retrospective — the real waste is everything you'll pour in from here, plus the opportunity you can't see while you're still holding on.
Origin
Annie Duke, from Quit: The Power of Knowing When to Walk Away. Draws on Richard Thaler's observation that most people won't quit until it isn't a decision anymore, and on Jeff Bezos's 70%-information rule for starting under uncertainty. Illustrated with Stewart Butterfield's shutdown of Glitch, which became Slack.
Core principles
- 01The same uncertainty that makes starting hard makes stopping hard — we want to be sure, and we never will be
- 02Waste is a prospective problem: what you've already spent is gone; what you spend from here is the actual waste
- 03Endowment is worst in product — once you've built a thing you own it, and you value it more than an identical thing you don't own
- 04Quitting carries an identity cost: quit late with a wreckage story and nobody questions you; quit early and everyone does
- 05The unseen cost of continuing is the opportunity you cannot even perceive while your attention is committed
How to run it
- 1
Ask the restart question
Knowing what you now know (the information you didn't have when you started), would you start this today? This is a forecast about the future, not an audit of the past.
Pro tip Note what new information you've learned since starting — the 'if I had known this, I wouldn't have started' items are the decision trigger.
- 2
Do the forward math, with honest assumptions
Project the economics forward without assuming today's conditions hold. Butterfield's break-even math said 31 weeks at current CAC — but he refused the flat-CAC assumption, because saturating the core market forces CAC up.
Pro tip The assumption you're most tempted to hold flat is usually the one that kills the business.
Watch out Extrapolating current unit economics unchanged is the standard way founders talk themselves into continuing.
- 3
Reprice the sunk cost as zero
Explicitly refuse to count what you've already spent. Sunk cost is retrospective; the decision is about resources not yet committed.
Watch out Also check for endowment (you built it, so you overvalue it) and identity (will I look like a failure?) — these will be masquerading as analysis.
- 4
Price the opportunity you can't see yet
Add the cost of your attention being unavailable for other opportunities. You will not be able to see the alternative until you let go — that invisibility is part of the cost of continuing.
- 5
Accept quitting early looks wrong from outside
Expect that when you quit at the right time — before the wreckage — people will ask 'what?'. Quitting only looks justified once you had no choice, which is exactly why almost everyone quits too late.
Pro tip Set kill criteria in advance so the early quit is a policy being followed, not an act of individual nerve.
In the wild
Glitch was a critically adored MMO — Monty Python meets Dr Seuss — with great word of mouth, top investors (Andreessen Horowitz, Accel), $6M in the bank and 5,000 diehard users playing 20+ hours a week. But for every diehard, 95–99 people arrived and left in five minutes. In 2012 they ran six weeks of paid marketing, growing new users 6–7% week over week. That Monday morning Butterfield wrote to his investors and co-founders: 'I woke up this morning with the dead certainty that Glitch was over.' Back-of-envelope math said 31 weeks to break even at current CAC — but CAC had to rise as they saturated the core gaming market. It was not a venture-scale business, and his employees were working for equity that wasn't worth their time. He shut it down. Two days later he noticed the internal communication tool his team had built to make Glitch.
→ Investors rolled their money into the new thing — 'searchable log of all company knowledge' — which became Slack. Butterfield has said he should have shut Glitch down before the marketing push; he needed the push to prove it to himself.
Common mistakes
Waiting until it isn't a decision anymore
As Thaler put it, most people won't quit until it isn't a decision — the startup is out of money, you're already in the blizzard, you've already fallen in the crevasse. Then you 'know for sure', but the option value is gone.
Treating waste retrospectively
The feeling of 'I'll have wasted everything I put in' protects past resources at the price of future ones. If you wouldn't start today, everything you put in going forward is the real waste.
Needing certainty to stop when you never needed it to start
You started under uncertainty — Bezos's 70% rule exists precisely to let you act without being sure. Demanding certainty before stopping applies an asymmetric standard that guarantees you stop late.
Is it for you?
Best for
Founders, product leaders and individuals deciding whether to shut down a project, product, job or relationship that isn't clearly dead yet
Not ideal for
Efforts whose payoff is known to be back-loaded and where the leading indicators are genuinely on track — don't use it as license for serial abandonment
From the transcript
“so as Richard Thor put it most people won't quit until actually isn't a decision”
“waste is a a prospective problem not a retrospective one”
“would I start this today that's a forecast of the future”
“and as smart as Stuart Butterfield is he couldn't see slack until he quit glitch”
“I woke up this morning with the dead certainty the glitch was over”
From the episode
This will make you a better decision-maker
Annie Duke (author of “Thinking in Bets” and “Quit,” former