This is part of the series Startup that Failed .
The Moment I Knew
Month 6. I was debugging a HIPAA compliance issue at 2 AM that would affect maybe three users.
And I remember thinking: This isn’t going to work.
Not the bug. The whole thing. The startup. The vision. The 18-month plan I’d laid out so carefully.
I knew. Deep in that part of your brain that’s honest when the rest of you isn’t. I knew the math didn’t work. I knew the market didn’t care. I knew I was building something impressive that nobody wanted.
And then I closed that thought, opened another terminal window, and kept building for another 12 months.
Welcome to the sunk cost fallacy, the startup founder edition. Where the admission price is your nights and weekends, and you keep paying long after the show has ended.
What the Textbooks Say (And What They Don’t)
Here’s the economics 101 version: The sunk cost fallacy is when you continue something because of previously invested resources, despite new evidence screaming that you should stop.
The rational thing to do? Evaluate decisions based purely on future costs and benefits. Ignore what you’ve already spent. It’s gone. It’s sunk.
Research by Kleinberg et al. (2021) even shows that people who know about this bias still can’t escape it. Even sophisticated agents — people aware of their cognitive biases — exhibit the same sub-optimal behavior when their own sunk costs are involved.
But here’s what the textbooks don’t tell you: In the moment, ignoring sunk costs feels exactly like quitting.
And everything in startup culture tells you that quitters don’t win. That the difference between success and failure is just one more sprint. That you’re probably right on the edge of a breakthrough.
So you keep going. Because walking away means admitting you wasted all that time. All that money. All those weekends.
The Compounding Investment That Wasn’t
By month 6, I had 450 hours into this thing and about $20K in infrastructure costs. Design partners were “excited.” I had a roadmap. I had momentum.
By month 12, I’d doubled down. 900 hours. $40K. A full year of promises to people who kept saying “this is great, just add X and we’ll definitely switch.”
By month 18, I’d hit 1,500+ hours (probably closer to 2,500 if I’m honest with myself, but I didn’t track it rigorously because I didn’t want to know). $60K total spend. A product that was genuinely impressive from a technical standpoint.
Revenue? About $300/month from fewer than 50 users. Lifetime total: ~$3,600.
The ROI was -94%. And that’s just the financial calculation.
At every checkpoint, the rational analysis was identical: The market isn’t responding. This isn’t working. Cut your losses.
But the emotional analysis got louder each month: You’ve already put in X. You can’t quit now. Just a few more features. Just a bit more polish.
This is what Erik Eyster and colleagues (2021) call “ex post rationalization” — we rationalize past choices, even when we recognize them as mistakes, because admitting the mistake means admitting waste. And I had a lot of waste to admit.
The Math I Should Have Done on Day One
You know what I never did? I never sat down and calculated what “success” actually meant in concrete terms.
I had vague ideas. “Build a profitable SaaS.” “Help dental practices modernize.” “Maybe get acquired someday.”
But I never defined the minimum acceptable outcome. Never set a financial threshold. Never decided what ROI would make this worth the opportunity cost.
Here’s what that calculation should have looked like:
My time is worth roughly $100/hour at my day job. So 1,500 hours = $150,000 in opportunity cost, conservatively. Add the $60K I spent, and I needed to generate $210K in value just to break even on a rational cost/benefit analysis.
At $300/month MRR, that’s… 700 months to break even. Or about 58 years.
Yeah.
But I didn’t do this math upfront because I didn’t want to face it. Instead, I told myself stories: “This is about learning, not money.” “I’m building something meaningful.” “The revenue will come once it’s feature-complete.”
All bullshit. Just elaborate excuses to avoid setting a hard number that would have forced me to quit at month 6.
The Persistence Trap
Here’s where it gets complicated: Startup culture glorifies persistence. And for good reason — most successful founders have stories about pushing through when things looked hopeless.
You’ve seen the quotes. “Winners never quit.” “It’s always darkest before the dawn.” “Most startups fail right before they would have succeeded.”
Paul Graham has written about how the most dangerous time for a startup is when things look bad but aren’t actually hopeless yet. When founders give up too early, right before the breakthrough.
The problem? This advice is indistinguishable from the sunk cost fallacy when you’re living it.
When you’re six months in, exhausted, with no revenue and dwindling enthusiasm from your “design partners,” how do you know if you’re about to break through or just throwing good time after bad?
You don’t. You really don’t.
The Grit Paradox
Angela Duckworth’s research on grit shows that persistence in the face of difficulty is one of the strongest predictors of success. Her book Grit: The Power of Passion and Perseverance is full of examples of people who succeeded because they didn’t quit.
But here’s the nuance that everyone misses: Grit is about long-term goals, not specific paths.
My goal was: Build a successful business.
My path was: This specific dental practice management software.
I confused the two. I showed tremendous grit toward the wrong thing. I persisted on a path that wasn’t working when I should have been pivoting to something else entirely.
Grit means refusing to give up on your goals. It doesn’t mean refusing to give up on your first (or second, or third) approach to achieving them.
I wish someone had told me that at month 6. Or maybe someone did and I wasn’t listening. Probably the latter.
When Pushing Through Makes Sense (And When It Doesn’t)
Here’s the counter-argument I kept making to myself: “But what about all the founders who pushed through when things looked terrible?”
It’s a fair point. The Startup Genome Report found that successful startups often go through a “trough of sorrow” where metrics look bleak right before they take off.
The difference? Those founders had signals. Maybe weak, maybe small, but real evidence that something was working. User retention was climbing. Word of mouth was happening organically. People were coming back daily.
I had none of that. I had design partners who gave great feedback but wouldn’t pay. Features that impressed people who wouldn’t actually use them. Compliments that didn’t convert to customers.
If the only positive signal is that the product is technically impressive, you’re building a resume project, not a business.
What I Should Have Done (But Didn’t)
The rational move at month 6 was straightforward: Stop building and force the decision.
I should have said to my design partners: “We’re launching in 2 weeks. Here’s the pricing: $200/month. Who’s in?”
Not “thinking about it.” Not “excited to try it.” Who’s actually paying?
If I got 5-10 paying customers from that group, I’d know the pain was real enough. If I got zero, I’d know to pivot or quit.
Instead, I kept building. “Just need to finish the mobile app.” “Just need insurance integration.” “Just need to polish the onboarding.”
I convinced myself that the lack of commitment was a feature problem, not a market problem. That once the product was “ready,” the customers would materialize.
They didn’t. Because they were never going to. The pain wasn’t acute enough to justify switching, no matter how good the product was.
The Hidden Cost Nobody Talks About
The real cost of the sunk cost fallacy isn’t the money or even the time. It’s what you don’t do with them.
1,500 hours is enough time to become expert-level in a new technology. To build and validate 3-4 different MVPs with proper kill criteria. To contribute meaningfully to open source projects that could have advanced my career.
$60K is seed funding for multiple properly validated experiments. Or a year of runway to do this full-time with a co-founder.
18 months is enough time to test a dozen different ideas if you’re willing to kill them quickly.
But I spent it all on one idea that I knew wasn’t working at month 6.
The opportunity cost of pushing through exceeded the sunk cost of walking away. By a lot. But I couldn’t see that because I was too busy protecting what I’d already invested.
The Framework I Wish I’d Had
Looking back, here’s what would have saved me 12 months and $40K:
Before starting, write down your kill criteria. Not vague feelings. Specific numbers. “If by month X I don’t have $Y in MRR with Z paying customers, I stop. No exceptions.”
Mine should have been: “If by month 6 I don’t have $2,000 MRR from at least 10 paying customers, I either pivot dramatically or quit entirely.”
Set those criteria when you’re excited and rational, before you’re exhausted and emotionally invested. Because once you’re in the thick of it, you can’t trust your judgment anymore.
Put a monthly check-in on your calendar. Not a “how’s it going?” check-in. A “do the numbers match the kill criteria?” check-in. If no, what changes? If you extend the timeline, what’s the new kill criteria?
I never did this. I just kept building and hoping the market would change its mind.
Ask yourself the fresh-start question. At each decision point: “Ignoring everything I’ve invested so far, is this still the best use of my next 100 hours?”
Not “I’ve invested 900 hours, should I invest 100 more?” but “If someone presented this opportunity fresh today, would I take it?”
By month 8, my honest answer was “hell no.” But I didn’t ask the question. I just kept building.
The Relief of Quitting
Opportunity cost: incalculable. But you know what the real cost was? Learning to distrust my own judgment.
Because I knew at month 6. I knew it wasn’t working. And I ignored that knowledge for a full year because admitting I was wrong meant admitting I’d wasted all that time and money.
That’s the real trap of the sunk cost fallacy: It teaches you that your internal warning systems can’t be trusted. That the obvious answer can’t be right because you’ve already invested too much.
When I finally stopped at month 18, you know what I felt?
Not failure. Not disappointment. Relief.
Because continuing something that isn’t working is exhausting in a way that failing fast isn’t. Every day you’re lying to yourself about the trajectory. Every conversation with your design partners is another exercise in self-deception. Every new feature is another brick in a building nobody’s going to live in.
Quitting, when done at the right time, isn’t failure. It’s strategy. It’s refusing to compound bad decisions. It’s choosing future opportunity over past investment.
The entrepreneurs who succeed aren’t the ones who never quit. They’re the ones who know when to quit the wrong thing so they can start the right thing.
Your Number
Whatever you’re building right now, you need a number. A specific, concrete, non-negotiable threshold that triggers a hard stop or pivot.
Not “I’ll know it when I see it.” Not “I’ll give it a bit more time.” A number.
“If by [date] I don’t have [specific metric], I stop.”
Write it down. Show it to someone who will hold you accountable. Put it on your wall.
Because in the moment, when you’re exhausted and invested and convinced you’re almost there, you won’t be able to trust yourself. The only thing that will save you is the number you wrote down when you were still rational.
Set your number before you start. Check it monthly. Honor it when the time comes.
Future you is counting on it.
Next in the series: How AI made me dumb, and how I’m recovering again by learning to think for myself.