What Could Go Right
Run devil's advocate on the upside, because you can only lose your money once.
- Difficulty
- Easy
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 92%
Gurley traces this frame to Benchmark partner Bruce Dunlevie, who picked up the phrase after reading Matt Ridley's The Rational Optimist and started using it in partner meetings. The input is any deal or bet where the payoff is wildly asymmetric — in venture, as Gurley puts it, you can only lose one time your money, but in a case like Google you make ten thousand times your money. The mechanism is a deliberate inversion of the devil's advocate: instead of stress-testing the bull case, the room stress-tests the bear case, asking what would have to happen for this thing to break out. The output is a decision process that stops treating a well-argued no as a win. Gurley is blunt that the trap is real — it is very easy to get into a trap in venture where getting a no right feels like a win, and it is just not, because the job is to find the outliers. The origin is his Google miss in 2002: search looked dead (Yahoo down from 80 to 10, Excite going bankrupt), two PhD founders insisted on running the company, and the checklist said no. A partner asked Larry Page what it would take to close; Page said 120 pre; the deal eventually got done at 80. Gurley's counterfactual is that they should have offered 150.
Origin
Benchmark partner Bruce Dunlevie began using the phrase in partner meetings after reading Matt Ridley's The Rational Optimist, around the period following Benchmark's decision not to chase the 2002 Google round. Gurley adopted it as the standing counterweight to venture's structural bias toward saying no.
Core principles
- 01In a portfolio where the downside is capped at 1x and the upside can be 10,000x, the cost of a false negative dwarfs the cost of a false positive.
- 02Most investment discussion defaults to finding reasons to say no, because a no feels like a demonstration of rigour.
- 03Correctly identifying a negative is not a win; the job of the fund is to find the outliers, not to avoid the losers.
- 04The corrective is to deliberately argue the positive case: what would have to be true for this to break out enormously?
- 05Gurley calls his Google pass the biggest mistake of his career, and the two best VCs of the era both said yes to the same deal he declined.
- 06The frame does not license investing in everything — you still go broke doing every deal — it rebalances a room that is structurally biased toward no.
How to run it
- 1
Establish the payoff shape before the debate
Write down the realistic downside (usually 1x your capital) and the realistic ceiling. If the ratio is thousands to one, declare that this decision is governed by asymmetry, not by probability of success.
Pro tip Say the ratio out loud at the top of the meeting so it frames every subsequent objection.
- 2
Ask the positive devil's-advocate question first
Before anyone lists red flags, force the room to answer: what would it take for this thing to break out, and can you imagine it being really, really big? Gurley describes this as being devil's advocate on the positive side.
Pro tip Assign the upside case to the most skeptical person in the room, not the sponsor.
Watch out If the sponsor argues the upside alone, the room will discount it as advocacy.
- 3
Separate mechanism objections from pattern objections
Distinguish reasons the business cannot work (unit economics, no distribution) from reasons it merely looks unusual (first-time founders, unfashionable category). Only the first class should kill a deal.
Watch out Two PhD founders insisting on being CEO was a red flag by pattern, not by mechanism — and that was Google.
- 4
Price the asymmetry into the offer
When the ceiling is extreme, do not haggle on entry price. Gurley's regret is that when Page named 120 pre, the right move was to counter at 150 to guarantee the deal closed.
Pro tip Ask the closer in your group what it takes to get this done right now, then beat it.
Watch out This applies only to genuine asymmetric bets; applying it broadly is how you overpay across a whole portfolio.
- 5
Audit your nos, not just your yeses
Periodically revisit passed deals and ask whether the rule that produced the no was flawed or whether the miss was collateral damage from an otherwise sound rule.
Pro tip Keep the reasoning, not just the verdict — you cannot audit a rule you never wrote down.
In the wild
In 2002 Google presented to Benchmark with roughly 25 employees. External conditions looked terrible for search, the founders had never been CEOs, and the checklist produced a list of reasons to decline. Benchmark did not chase. John Doerr and Michael Moritz — the two best-regarded VCs of the era — locked hands and said yes.
→ Gurley calls it the biggest mistake of his career, and the episode became the origin of Benchmark's standing what-could-go-right question.
A newly hired Benchmark partner was told to bring in every company he met. The group said no to company one, no to company two, and he grew frustrated. On company ten, the reaction flipped entirely — the room told him to close it immediately.
→ The partner absorbed the collective pattern recognition through repeated exposure rather than through a rulebook, and the group kept its bias toward the breakout case.
Common mistakes
Treating a correct no as a win
Getting overly jazzed about correctly identifying a negative feels rigorous but produces nothing. In a power-law portfolio the scoreboard only counts outliers found, not losers avoided.
Setting hard rules in a dynamic system
Gurley notes that many variables are dynamic and none are constant, so the minute you set a very hard rule you may be setting yourself up for a mistake. Rules should be loose pattern recognition, constantly re-checked.
Using the frame to justify doing every deal
The asymmetry argument does not mean unlimited yeses. Gurley is explicit that you cannot do every investment or you go broke; Benchmark's real constraint is board seats, not capital.
Is it for you?
Best for
Anyone allocating capital, headcount or attention in a power-law domain where a single hit pays for every miss.
Not ideal for
Decisions with capped upside and uncapped downside, where the loss from a bad yes is unbounded.
From the transcript
“you can only lose one time your money and yeah in a case like Google you make what ten thousand times your money and that…”
“I think Bruce had just read the rational Optimist which is a Matt Ridley book and he started using a phrase at our partner meeting…”
“just not that big a deal it's not the job the job is to find the outliers”
“thinking Devil's Advocate in your mind right and Devil's Advocate on the positive side like what would it take for this thing to break out”
From the episode
#651: Legendary Investor Bill Gurley on Investing Rules, Finding Outliers, Insights from Jeff Bezos and Howard Marks, Must-Read Books, Creating True Competitive Advantages, Open-Source Strategies, Adapting Mental Models to New Realities, and More