Founders tend to worship at the altar of perseverance and grit. After all, startup lore is marked by “slow-bake” stories like Roblox’s or (more recently) Figma’s journey, where it took years of patiently grinding it out to get to that inflection point and outsized outcome. The aversion to quitting is so strong that that merely saying the word is almost like uttering “Macbeth” in a theater — even the mention might invite misfortune. Instead, we opt for softer euphemisms, like “pivoting,” “iterating,” or “going in a different direction.”
So when Annie Duke set out to write her latest book, titled “Quit: The Power of Knowing When to Walk Away,” she knew she was tackling a touchy subject. “I’m on a mission to rehabilitate the word, which has been given the Voldemort treatment. But while my book is called ‘Quit,’ in no way do I want to imply that grit isn't an important trait or a muscle worth developing,” she says. “Grit gets you to stick to things that are worthwhile, but also very hard. And most of the things that we do that are worthwhile and will change the world are going to have periods that are really difficult. Where we go wrong is in thinking that grit is just good — period.”
Earlier this year, we officially introduced Duke as First Round’s Special Partner for Decision Science, sharing her decision-making advice for founders right here on The Review. As a consultant, bestselling author and former poker pro, there are few who are better equipped to teach founders about the art of knowing when to hold ‘em, and knowing when to fold ‘em.
“When we look at success stories that were a long time in the making, there’s a temptation to say sticking to it is just good — full stop. But the problem is that the grit that allows us to power through will also get us to stick to things that aren’t worthwhile,” says Duke. “Success comes from sticking to the stuff that’s working and quitting the rest. That’s why quitting is a skill you need to get good at.”
It's true that in order to be successful at something, you have to stick to it — but that doesn't necessarily mean that sticking to something makes you successful.
In this exclusive interview, Duke unpacks the psychology behind why it’s so hard to walk away. She flags the cognitive biases to beware of, from how identity can become enmeshed in company building, to the common tendency to make incremental changes instead of taking more drastic measures.
Duke shares a set of mental models that will help you approach crucial decisions with a clearer mindset, sharing targeted strategies for goal-setting, evaluating progress, and seeking outside counsel. Her tactical advice is helpful for all the big forks in the road, whether it’s quitting your job, firing an early employee, abandoning a product, changing your go-to-market strategy, or sadly, shutting down your startup.
Prefer to listen along instead? Check out our recent In Depth podcast episode with Duke here.
CALL IT BY ITS NAME: WHY A PIVOT IS A QUIT
Setbacks come in different stripes. In the startup context, sunsetting a product feature that flopped pales in comparison to the difficult choice to wind down the business altogether. But take the “pivot.” Despite its scrappy, entrepreneurial connotations, it’s a quit in disguise, says Duke. “I don't think we should hide behind the term pivoting.”
We tend to think about quitting as leaving the court. But pivoting is still quitting.
“The idea of quitting is such a bitter pill to swallow that we have to take it with a spoonful of sugar,” she writes in her book. “Stripped of its negative connotation, quitting is merely the choice to stop something that you have started. We ought to stop thinking that we need to wrap the idea of quitting in bubble wrap and serve it soft.”
She points to a couple of examples: “Stewart Butterfield was developing an online game and then pivoted to building Slack — that's quitting. He quit developing a game,” she says. “Or take Notion, where the founders have a dramatic story of throwing out the code, downsizing the team, and rebuilding the product from scratch. The key takeaway from these stories is that they weren’t just pounding their head against a product that wasn’t working.”
In those dramatic pivots, where things looked to be on the brink of failure and then founders rise from those ashes into success, there's a lot of quitting that's usually happening.
But the line between when to keep at it and when to throw in the towel isn’t usually clear-cut. Here’s why so many founders opt for the former, even when they shouldn’t: “You have this broader goal of creating a successful startup, and that's what you're abandoning when you shut a company down. You’re admitting that you can't turn it around anymore. That's the moment that you have to say, ‘It was failing before, but now I have failed,’” says Duke.
“Pivots are easier for us to handle because we still have that chance to make the company work and achieve our goal of building a startup — even if we have to abandon the product that we were initially developing.”
As long as that chance exists, it will usually feel too early to give up the cause.
BUT WHY IS IT SO HARD TO QUIT? THE COGNITIVE HURDLES STANDING IN OUR WAY
It’s a familiar scene: A startup’s go-to-market motion just isn’t working. They can’t seem to get customers to sign on. But the founder insists that it’s a hiring issue — if they could just get the right salesperson on board, then all the pieces of the puzzle would snap into place.
“As an outsider, you can see that it’s much more complicated. You need to change the whole GTM motion, or even change the product itself drastically. And external advisors often have this feeling of, ‘Why can’t they clearly see what I can see?’” says Duke.
The issue stems from the expectations we place on our future selves. “We all expect that if we see signals that what we’re doing isn’t working, obviously we’ll then change gears. This is the intuition that enables us to make decisions under uncertainty in the first place,” she says. “When you think about which product you’re going to develop, or the GTM motion you’ll bet on, you’re making those decisions under tons of uncertainty. There’s a lot of stuff we don’t know. And there’s going to be information discovery along the way. We believe that when we discover that information, we will surely quit something that’s not working and change course.”
In fact, the opposite tends to be true.
When we get signals from the world that are negative, that the thing we’re doing isn't working, we actually escalate our commitment to whatever it is that we're doing.
Duke flags a few of the driving forces behind this impulse to double down:
Reason #1: Sunk cost fallacy
“Richard Thaler first identified this effect. When we've put a lot of effort into something, we don't want to switch course because we think that then we'll have wasted all of the time or the effort. If you buy a stock at $50 and it's now trading at $40, you don't want to lose $10. But of course that $10 is already lost — it's already sunk,” says Duke.
Instead, the question should be: Is it worthwhile going forward? “So often founders think, ‘If I stop now, I'll have wasted a year and a half of my time, my employees’ time, my investors’ money. But do you want to waste the next minute? Continuing to let them plug away at something where their equity isn't worth anything is what’s wasting their time.”
Start thinking about waste as a forward-looking problem, not a backward-looking one.
Reason #2: Status quo bias
“Once we’ve established the status quo, like the GTM strategy that we’re currently executing on, there’s a bias that makes it very hard for us to switch,” says Duke. “When we dump that go-to-market motion and try to develop something else, if that new plan doesn’t work out, there’s an asymmetry to how we process that regret. We feel it much more keenly when we start something new. So we'd prefer to make incremental changes, hoping we can still turn it around.”
We're much more tolerant of the bad outcomes that come from sticking to the plan than from switching to something new.
Reason #3: Founder identity
“When you do things that are outside of the mainstream, they become part of your identity — which is particularly true for founders. And walking away from your identity is so incredibly hard,” she says. “The hardest thing to quit is who you are. You will be more likely to reject the facts rather than to update your beliefs, even when the evidence is clear.”
This compounds over time. “The more that we put into something, the more our identity gets tied up with what we're doing, the more endowed we are to it. That makes it harder and harder to quit. You work on something for six months, which makes you more likely to continue working on it because you've already put six months into it. That causes you to put another year into it. And so on, and so forth,” says Duke. “It becomes a self-reinforcing cycle.”
There's a big difference between shutting your company down after six months and shutting it down after nine years. The longer that we continue to do something, the more cognitive debris builds up — and the harder it is to abandon.
Reason #4: Tendency to grade goals as pass/fail.
There is no shortage of goal-setting frameworks to choose from. But Duke finds that most of them have the same problem. “There's lots of evidence that goal setting is good. I'm not going to argue with the body of evidence that says if you create very clear, specific goals it helps people to achieve outcomes more quickly,” she says.
But there’s a sneaky downside that we often overlook. “A woman was running the 2019 London marathon and she started experiencing pain in her legs around mile three. And then, at mile 8, her fibula snapped. And here's the amazing thing — she kept running. She finished the marathon on a broken leg,” says Duke. “This might sound kind of nutty. But four people in that 2019 London Marathon broke something and continued to the finish line. Just a quick Google search will tell you that stories like this are incredibly common. People continue to run, which extends recovery time after the race and may impact their ability to ever run again in the future,” she says.
“Once we set that finish line, we grade it pass-fail. If you quit after mile 20, you failed. If you run 26.2 miles, you passed. Even though running 20 miles is better than not having tried at all — cognitively, not having tried at all feels better to us than having to quit. If you’re 300 yards from the summit of Everest, but you don’t actually summit the mountain — you failed. Never mind that you climbed higher than almost any human being ever has.”
The problem with goals is that once we set a finish line, they’re graded pass-fail. And we will barrel toward that finish line, come what may. The goal itself becomes a fixed object, even though the conditions have changed.
Founders will recognize that this phenomenon applies outside of physical feats, of course. “If nobody's buying the product, and we clearly haven't achieved product-market fit and it doesn't look like it's on the horizon, we'll keep tinkering anyway. We ignore the signals that we’re off track. Because we have a finish line, and it's graded pass-fail,” says Duke.
“One of the things that give startups an advantage is that they're exploring in a way that established companies aren’t able to. Enterprises have an innovation problem. Startups are exploratory. But what we have to realize is that the very act of setting a goal makes you become more and more enterprise-like. You stop exploring other avenues, strategies, products, and motions that you could be pursuing,” she says.
Her advice: Get specific about the “unlesses.” “I'm going to run this marathon unless I break my leg in the middle of it — and then I'll stop. This seems silly, because, obviously, we think, if we break our leg we’ll stop. But you're not going to stop once you're in the middle of it. So when we’re setting goals, we have to set those unlesses in advance, too” says Duke.
Goals are great — as long as you have thought in advance about what would make it so that you wouldn’t pursue that goal anymore.
HOW TO BECOME A BETTER QUITTER: MENTAL MODELS FOR MAKING TOUGH DECISIONS
So how can founders cut through these cognitive challenges and make the right decision? How can you get better at spotting those instances in the moment, saving yourself heartache and dead-ends? Duke shares five strategies for disciplining your decision-making and getting more comfortable with quitting:
Strategy #1: Train the monkey before you build the pedestal.
“My key recommendation is that you need to understand whether or not you should continue as quickly as possible. And the model that I like to use for this comes from Astro Teller, who's the CEO or ‘Captain of Moonshots’ at Google X,” says Duke.
“This group is trying to do super innovative things that have a very high chance of failure. They’re going after huge non-incremental discoveries, and they want to get these big world changing ideas to commercialization in five to 10 years. But even at Google, they have limited resources. You only have a certain amount of time, attention and money that you can devote to things,” she says. “And so Astro Teller is trying to understand if a project is worth pursuing or not, especially in comparison to all the other things that they could pursue. And he’s trying to get to that decision as quickly as possible to reduce those sunk costs and identity issues.”
Teller relies on an unusual mental model to get there. “Imagine that you’ve decided that you want to train a monkey to juggle flaming torches while standing on a pedestal in the town square. If you could accomplish that, you would make lots of money. But when you’re approaching this idea, you should not build the pedestal first. Instead you should make sure that you can actually train the monkey to juggle the flaming torches,” Duke says.
“Because that's the bottleneck. There's no point in building the pedestal if you haven't figured out whether or not you can actually get the monkey to juggle the flaming torches. That's the unknown. You can certainly build a pedestal — in a pinch, you could turn a milk crate upside down,” she says. “Teller says to tackle the hard part first, because everything else is just going to create an illusion of progress.”
If you can’t solve for the hardest part, the part with the most uncertainty, there’s no point in doing anything else.
There’s another problem with starting with the pedestal. “The time, effort and money that you put into building the pedestal starts to create those sunk costs. Now your identity is more wrapped up in what you're doing — even though it's not creating true progress — which makes it much harder for you to actually quit when you figure out that you can't get the monkey to juggle those torches,” says Duke. “Instead you’ll say, ‘I just need to do this one more thing, I'm so close to the breakthrough.’”
For an example that doesn't involve primates, consider the California high-speed rail system. “The idea was to connect San Francisco and Silicon Valley to LA and San Diego. A $9 billion bond was approved back in 2008, and the projections at the time were that it would take about $33 billion to complete by 2020. They approved some track between Madera and Fresno, which was the interior of the line on flat land,” says Duke.
“The problem with this approach is that any track that you build on flat land is, in essence, going to be a pedestal — because you already know that you can build it. It turns out that when it comes to the California Bullet Train, there are actually a couple of monkeys that are really hard to figure out. These monkeys are two mountain ranges: the Tehachapi Mountains and the Diablo Range’s Pacheco Pass,” she says. “And there are massive engineering problems to try to figure out how you blast through those mountains in a safe and cost effective way so that you will be able to complete the line. It’s 2022 and they’ve still hardly built any track. And the latest cost projections are $105 billion, and they still haven’t figured out how to solve for the mountains.”
This is, of course, an extreme example, but terms like “low hanging fruit” are thrown around all the time when building products and companies. “Make sure you're not just collecting debris without actually solving for the hard thing first,” says Duke.
I don't have any problem with tackling the low hanging fruit. Eventually you have to. But you better make sure that you solve for the bottleneck first. Because every bit of low hanging fruit that you tackle creates an illusion of progress and sunk cost problems.
Strategy #2: Temper your projections by relying on base rates.
“For any individual founder working on a problem, the probability of failure is really high. There are founders who pound the same pavement and it just takes a long time before they hit gold, versus others who pound the pavement but ultimately figure out that it isn't working,” says Duke. And founders are an optimistic bunch. Even if growth has stalled out or the product has yet to find its footing, it’s easy to convince yourself that you’ll still be in that select few, that small group that does go the distance.
“The founder is just one individual caught in their own perspective,” says Duke. Thus, much of her advice centers around broadening that window and pulling in external perspectives. One technique is grounding yourself in base rates. “Here’s a simple way to think about base rates: How often does something happen in the situation that you're considering?” Duke shares a few examples:
Stock market: “How often, year over year, in any given fiscal year is the stock market up from where it started at the beginning of the year? Depending on how you calculate it, I believe somewhere around 70-75% of the time the stock market is up. So it doesn’t matter if over the past five years it was up for all of those years. The next year, it should be up about three-quarters of the time — unless something significant has changed.”
Valuations. “If the top SaaS companies are historically valued about 16-20X revenue, and then all of a sudden there's one year where they're being valued at 40-60X revenue, that would be odd compared to the base rate,” she says. “Now, that doesn't mean that that valuation is untrue. It just means that something has to be substantially different in the environment, like how the valuation for retail companies was impacted by the shift from brick and mortar to online.”
Hurricanes. “It doesn't make any sense to look historically over the last 100 years at either the strength or the frequency of hurricanes, because we know that there are big, recent shifts in average temperature and how warm the ocean is, making hurricanes stronger and more frequent now. And so those base rates wouldn't hold because something significant has changed about the world.”
But how is this helpful to a founder trying to assess their startup’s prospects? “People make predictions and forecasts by thinking about them in relation to their own experience. The base rate asks you to take an extra step. Look at the information you have, think about the problem you're trying to solve, and then try to figure out how that generally goes — before you start to think about your own experience. The reason that you have to do that is because our perspective on the world is actually quite riddled with cognitive bias,” says Duke.
Here are a few of these biases that base rates can help combat:
Availability bias: “As an example, if I were to ask you to estimate the probability of a terrorist attack on U.S. soil, your estimate would be much higher than what the actual base rate is,” says Duke. “The reason it just is that it's easier for you to recall, because those things tend to be on the news, and so we think about those as more frequent than they actually are because ease of recall becomes a proxy for our judgments of frequency.”
The planning fallacy. “If I'm trying to figure out how long a project will take, I tend to underestimate how long it's going to take. So I ought to look at the base rate for that.”
The illusion of control. “We think we have more control over outcomes than we do. This is similar to the gambler's fallacy, where we think there's a higher likelihood of things happening than they actually will.”
“If we want to actually discipline those cognitive biases so that our forecasts, predictions and decisions are more accurate, one of the best ways to do that is to look out and see what's going on in the world,” says Duke.
“Returning to valuations, I heard a lot of people in the summer of 2021 talking about how the world is different. They were seeing other founders raise at these crazy valuations. But unless there's some sort of paradigm shift, like that Walmart to Amazon shift, looking up the historical base rate for valuations for companies helps to discipline that desire that we have to think that something special is going on,” she says. “It's not that something special isn't ever going on. But you at least have to disprove that the base rate wouldn’t hold in that situation.”
Strategy #3: Calculate expected value.
If you’re weighing whether to leave the arena or stick it out in hopes of pivoting your way through, here’s Duke’s litmus test: Quit when your expected value goes negative in comparison to other things that you could be doing.
“When we start anything, there's some sort of expected value that you can calculate. Essentially you can think of it as: Does the upside outweigh the downside? When the upside is great enough to outweigh the downside, then we would say we're winning to the decision — we're positive in expected value,” she says. “To be clear, this doesn't mean that the company's never going to fail. It means that if you were to make that decision over and over again, there would be enough upside to compensate for the downside that's associated with it.”
On the flip side, when the expected value is negative, you ought to walk away. “But we can broaden that to other situations. Say you’re gaining ground to your goal, but there are other things that you could be doing which are associated with even more upside. So then it would behoove you to switch to the thing that has even more upside. Stewart Butterfield’s game before Slack is the example here,” says Duke.
“They were gaining lots of customers and doing a big marketing push, but it looked like they were probably going to break even within 31 months, only if they could sustain the growth. What Stewart realized was that they could get to a place where they were going to make money, but it was never going to be a venture-scale business. The expected value was most likely positive, but it just wasn't big enough compared to what he felt like he could accomplish.”
Remind yourself: Life's too short to be working on something that's not worth your time.
Strategy #4: Bring in a quitting coach.
“When you're actually facing questions of ‘Am I going to just throw this product out or am I going to change my strategy?’ it's hard to get these decisions right. One of the best strategies you can employ in order to get better at making these decisions to quit is actually to get somebody from the outside to help you,” says Duke.
“In the book, I talk about Ron Conway, who really prides himself on coaching founders to quit. People would be surprised because he has obviously helped people grit it out, to great success. But one of the things he's proudest of is that when he sees that a company isn’t going to work, he feels an obligation to help free that founder up so they can go work on something amazing,” she says. “It's hard for us to get to those decisions for ourselves, but we can generally see them pretty well from the outside.”
Outside advisors have a much bigger data set than an individual founder. They have access to the base rates, they understand the reference class. They can identify when it’s a dead company walking, as opposed to one that's just going through hard times that they can push through.
“They also don’t have the same sunk cost fallacy as a founder who's been building for a year and a half. So the value is that they're not in it with you,” says Duke. “They can see clearly that if you have employees that are working for very little cash comp and lots of equity, that the moment that you determine that equity isn't worth their time, you should let them go.”
Strategy #5: Set (specific) kill criteria.
If you’re struggling to objectively evaluate your company’s way forward, lean on this method: “There's all sorts of benchmarks that you can set for yourself of what you need to see happen within a quarter within two or three quarters. Those benchmarks become what I call kill criteria, literally criteria for killing a project or changing your mind or cutting your losses. If you miss them, it means you’re going to quit,” she says.
“If you're pre-revenue, that might be about the product: How are tests going? What's your ability to hire and get other people excited? Can you tell the story in a compelling way that gets people to want to give you money?” says Duke.
“Reviewing those criteria on a regular cadence with an outside advisor will help mitigate your tendency to say, ‘No, I know I can turn around. I can tweak the messaging or hire a new salesperson.’ This is the strategy that Ron Conway uses: ‘You're a brilliant person, I have no doubt you can turn around. But what does turning it around look like within what time?’ A simple way to develop kill criteria is with ‘states and dates,” she says. “‘If by (date), I have/haven’t (reached a particular state), I’ll quit.’”
Duke takes us through an example of how she would push a founder to get more specific here: “Say the benchmark is generating net new ARR — how long would we have to wait to see that change occur? If it’s a long sales cycle, your tendency as a founder might be to say it’ll take six months. But the advice-giver could say, ‘Well, it might take six months to close new business, but you should be seeing changes at the top of the funnel if this new salesperson is working. Let’s say that will take a month. What does it look like with quantitative data — what are the numbers that we're going to see there?' Then you can agree to revisit in a month, and if you're not seeing progress, then something drastic needs to change — then you really do need to think about quitting or pivoting.”
Essentially, you need to establish a tight timeline and identify the leading indicators in advance. “What I want to avoid is it's six months from now and they still haven't generated new ARR, but they're telling me again that they ‘just need to tweak things,’” she says.
Get to the leading indicators and actually write down the benchmarks. What are the things that we're going to see that are going to tell us that things have turned around within the next month?
You can also apply this approach at every level of the company. Take this sales example: “Imagine that you are pursuing a sales lead that was generated through an RFP. It's six months later and you’ve lost the deal. Looking back, you realize there were early signals that you were not going to win. What were they?” says Duke. “Maybe you had a few meetings and they couldn't get a decision maker in the room. Maybe the RFP matched a competitor a little too closely. Maybe they just wanted to talk about pricing in the first meeting and didn’t even want a demo. They're probably trying just to use you as a stalking horse to beat somebody else down on price,” she says.
“When you’re actually in the situation it’s hard to see it. When you’re in the middle of a deal, you continue to pursue it and insist that you can still win. Assembling the criteria in advance can help you see if it’s not a deal that’s worth pursuing actively or bringing it further into the funnel. You can have different levels, like automatic kills or ones where you need to seek out further information before making a decision.”
WRAPPING UP: WHAT IT LOOKS LIKE TO BE A COMPETENT QUITTER
How can you tell you’re making progress on your quitting skills as a founder? What are the indicators that you’ve taken Duke’s words to heart?
“There are some surface signals. For example, you’d get efficient at letting people go who aren’t good fits. As soon as you start thinking, ‘Maybe this person isn't a good fit,’ you would sit down with them and figure out what you expect to see in terms of turning it around. And you would then revisit those leading indicators very quickly,” she says.
“In a deeper way, this kind of thinking would be inculturated within the company. When you start a new project, you would be thinking about making forecasts, examining your options and giving your best guess of what the future might hold. You’d be writing those things down so you can go back and revisit the accuracy of those forecasts,” says Duke. “You’d write down your kill criteria and make commitments broadly within the team so you can actually act on them. Then once you’re in flight, you would get on a regular cadence of revisiting those goals.”
To recap, Duke shares some questions for founders to use for these check-ins:
Are these goals still the right goals to pursue, given what we’re trying to achieve?
What are the other options? Are there creative alternatives to what we're trying to do?
How are we going to find out as quickly as possible that the thing that we're doing isn't working?
If I imagine it's six months from now and this project has failed, what were those early signals or leading indicators?
Are we actually tackling the bottleneck, or are we retreating into pedestal building in order to make ourselves feel better?
What does success really look like? What do I expect to see happen?
Is this just hard or is it hard and not worthwhile?