Once you've paid for something, that payment is sunk—it's gone, and no future action can recover it. Rational decision-making says sunk costs should be irrelevant. Your decision to continue should be based on future costs and benefits, not on what you've already spent. But your brain doesn't work that way. The sunk cost fallacy is the compulsion to throw good money after bad because you've already thrown bad money—and your brain refuses to acknowledge that the past spending is gone forever.
You buy a movie ticket for $15. Halfway through, you hate it. Rationally, you should leave—the $15 is sunk, and staying costs you an hour of wasted time. But you stay because you already paid. You don't want to "waste" the $15. Of course, you're wasting something worse (your time), but the sunk cost fallacy makes the past spending feel more real than the future cost.
Amazon Prime weaponized this brilliantly. Once you've paid the annual fee ($139), every service feels "free" because you've already sunk the cost. You use Prime Video more, Prime Music more, Prime shopping more—not because these services improved, but because the $139 feels like it needs to be justified through usage.
Arkes & Blumer (1985) conducted a foundational experiment on theater-going behavior.1 They tracked actual ticket purchasing and attendance:
Same availability, same price point for additional shows, different sunk costs. The $15 ticket holders attended 25% more shows because the sunk cost was larger. They felt compelled to use what they'd paid more for.
The mechanism is psychological: sunk costs create a sense of obligation. Your brain treats the past spending as an unresolved debt. Using the product or service is how you "complete" the transaction—not economically (the transaction is already complete), but psychologically (you need to justify the spending).
This is why people finish bad meals at expensive restaurants. This is why people stay in unsatisfying subscriptions they've paid for upfront. This is why people keep investing in failing projects at work—they've already sunk time and money, and abandoning the project feels like abandoning the investment.
Sunk costs don't just make you use what you've paid for. They make you invest more to justify the past investment. This is called escalation of commitment, and it's where sunk costs become genuinely dangerous.
You've invested $10,000 in a failing business venture. Rationally, you should cut losses. But your brain says "I've already invested $10,000, so quitting would mean that investment was wasted." Instead of quitting, you invest another $10,000 to "turn it around." Now you've sunk $20,000, and the trap deepens.
Shotton points to Amazon Prime as the elegant version: once you've paid, every additional service feels justified by the sunk cost, so you use more services, which makes the annual fee feel more justified, which makes renewal feel more justified. The sunk cost locks you in, and each use compounds the psychological justification for the subscription.
This is why software companies push annual billing instead of monthly. A monthly subscription, you can quit anytime without loss. An annual subscription, the $139 fee sinks in January, and by December you've used enough services to feel like you "got your money's worth." The sunk cost prevents churn.
This is critical for implementation: not all costs that feel sunk are actually sunk. There's a difference between:
True sunk costs: Money already spent that can't be recovered (theater tickets, annual subscriptions, non-refundable deposits). These are genuinely irrelevant to future decisions, but your brain treats them as relevant.
Future costs disguised as sunk costs: Time, energy, or opportunity costs that continue to accrue if you stay in the situation. If you stay in an unsatisfying job "because I've already invested 5 years," you're not protecting a sunk cost—you're incurring a future cost (wasted time that you could spend elsewhere). Your brain frames it as "protecting the past investment," but the actual loss is future opportunity.
Brands that exploit sunk costs carefully distinguish these. Amazon Prime uses true sunk costs (annual fee that's gone) to lock you into future services. They don't emphasize future costs (continued subscription = continued lock-in). Instead, they emphasize the justification of the past cost (you've already paid, so use the services).
Step 1: Create a sunk cost barrier Shift from variable costs (pay-per-use) to fixed costs (upfront, large payment). Annual subscription > monthly subscription for lock-in, because annual fees feel sunk in a way monthly fees don't.
Step 2: Make future services feel "free" by framing them as justification of the sunk cost Once someone's paid $139 for Prime, every additional service (Video, Music, Shopping) feels "free" because it's not an additional cost—it's usage of the already-sunk $139. You're not selling the additional services. You're selling the feeling that using them justifies the sunk cost.
Step 3: Ensure the sunk cost is large enough to create obligation, but not so large it triggers regret Too small a sunk cost ($5 annual fee), and it doesn't feel sunk. Too large a sunk cost ($1,000 annual fee), and the person regrets it and cancels. The sweet spot is large enough to feel like an investment, small enough to feel like reasonable spending.
Step 4: Offer easy additional services that leverage the sunk cost Once the sunk cost exists, any additional service is easy to add because it feels "free." Amazon Prime Video, Music, and Shopping all leverage the single $139 annual fee. The additional services aren't sold—they're offered as ways to justify the existing sunk cost.
Step 5: Structure renewal to feel like continuing the commitment, not making a new decision Auto-renewal is the technical implementation. Psychologically, renewal feels like continuing a commitment ("I've already invested in Prime, so I'll continue") rather than making a new decision ("Should I pay $139 for Prime?"). The sunk cost from the previous year carries psychological weight into the renewal decision.
Escalation of commitment has limits. If losses become too visible—if you're losing $1,000 per month and the sunk cost is $50,000, eventually people admit the loss and quit. The trap works because the sunk cost feels larger than the ongoing loss. But if the ongoing loss becomes undeniable (business failing visibly, service completely unused), eventually people overcome the sunk cost trap.
Also, sunk costs only trap people who have a choice. If you're locked in by contract (can't cancel without penalty), you're not victim of sunk cost fallacy—you're victim of literal lock-in.
Psychology → Loss Aversion: Sunk costs create a loss-aversion trap. You've lost the money (past), so you make suboptimal future decisions (stay in bad situation) to avoid admitting that loss. Loss aversion amplifies the sunk cost effect: the pain of admitting you wasted $139 feels worse than the benefit of canceling. Loss Aversion explains why the psychological pain of acknowledging sunk costs exceeds the rational pain of incurring them.
Behavioral-Mechanics → Illusion of Progress: Sunk costs pair with illusion of progress: as you use a service more (to justify the sunk cost), you feel like you're "getting value," which reinforces the decision to keep paying. The progress you're tracking (usage) is an artifact of the sunk cost, not evidence of value. Illusion of Progress explains why increased usage feels like increased value.
Behavioral-Mechanics → Present Bias: Sunk costs are a form of present bias: the immediate need to justify the past spending (felt presently) overrides the rational future calculation (whether the service will actually be used). You prioritize the immediate psychological relief of feeling like the past spending was justified over the future decision of whether to continue.
Sharpest Implication: Sunk costs are a form of self-imposed lock-in. Companies don't have to trap you with contracts—they just have to make you sink a cost that your brain refuses to admit is gone. Once you've paid, you become your own jailer, making suboptimal decisions to justify the sunk cost. The implication: you're easier to lock in through psychology than through contracts. A $139 annual fee that feels sunk is more effective lock-in than a contract that forces payment, because you want to justify the sunk cost.
Generative Questions: