The Sunk Cost Fallacy

Sunk costs reflect investments that can no longer be recovered. The classic example of this is spending money to repair an old car. Every time you pay for repairs, you are incurring an additional sunk cost. When considering whether or not to sell your car, you should not rationally consider past sunk costs as the only thing that matters is optimizing present/future utility. That is to say, the money you spent repairing the car should not factor in to whether you ought to keep the car. Instead, you should consider future potential costs, and whether selling the car to buy a new one would be a better decision than keeping it. However, this is not how people behave in real life. Most of us experience the sunk cost fallacy/effect, where we consider past investments when making decisions about the present and future. For example, if you spent $10,000 repairing your car over the last year, you may feel like selling the car (especially if it’s for less than $10,000) would be wasteful. Thus, you may instead continue to invest in a bad investment to avoid the negative feelings associated with abandoning it. Research suggests that the sunk cost fallacy is driven by negative affect (emotion). For example, Dijkstra & Hong (2019) had participants in a study read sunk cost scenarios, and asked them to rate how likely they would be to choose the sunk cost option versus the normative correct option on a scale of 1 to 6. When participants reported feeling more negative affect, they were more likely to prefer the sunk cost option. This is in line with evidence from neuroscience that suggests sunk costs elicit the dorsal anterior insula which is associated with anticipatory negative affect (Fujino et al., 2016), various regions of the anterior cingulate cortex related to emotion regulation (Fujino et al., 2016; Haller & Schwabe, 2014), and the dorsolateral prefrontal cortex associated with avoidance of wastefulness which is a social norm (Haller & Schwabe, 2014). Further, Hafenbrack et al. (2014) found that mindfulness meditation, which reduces negative affect and focuses attention on the present, can reduce susceptibility to the sunk cost fallacy.

A meta-analytic review of 45 studies and 100 data samples by Roth et al. (2014) found that the sunk cost fallacy is robust with a weak to moderate effect on decision making, but that its influence differs depending on context. For example, Liang et al. (2014) found that sunk costs influenced decisions about bed and breakfast accommodations (that required a deposit) more so than a membership to a website that lets you read unlimited books (that required a monthly fee). Roth et al. (2014) also found that as we get older, we are less susceptible to sunk costs, which lines up with past findings by Strough et al. (2008). Strough et al. (2008) suggest this may be due to a relationship between the sunk cost fallacy and loss aversion, as loss aversion is driven by our negativity bias which decreases as we age, and can shift to a positivity bias as we get older. Roth et al. (2014) also found that as the time between an initial payment (e.g. sunk cost) and consumption increases, the sunk cost effect decreases. They in turn describe how this might be relevant to businesses in the service industry, where usage of tickets may decrease with time. Thus, one might predict if tickets are sold one year ahead of time, less individuals will attend the event than if tickets were sold one month ahead of time.

When and How to Build Sunk Costs into your Marketing Strategy

There are two key times that you want to consider sunk costs. The first occurs when there is a natural sunk cost associated with your product or service. Sunk costs in these situations can often help increase profits. For example, many website developers sell WordPress sites that are relatively cheap to create (as they are often built off a pre-existing theme). However, WordPress sites need to be maintained and updated regularly, so every couple of months they contact customers asking for a fee to update the site. Overtime, these costs add up, and a poorly designed and expensive website has become a sunk cost. Developers take advantage of this and begin to charge higher prices for updates. Customers who have now invested considerably more than they had expected into the site become adverse to switching to a new site, thinking it will just lead to the same experience. Thus, website development has very much transitioned to a model that leverages the sunk cost fallacy for profit, due to the natural way WordPress and other content management systems are set up. For most businesses a normal HTML website works just as well as a WordPress website, without the need for ongoing maintenance. But without additional maintenance fees, an HTML website will seem more costly upfront. It’s important to keep in mind that if a product/service involves a sunk cost, it often drives out alternative options that have upfront costs. As a result, business models that involve sunk costs can slowly come to dominate an industry. Video games are the most salient example of this. A decade ago, you would pay around $60 CAD for a video game which would include all relevant content for the life of that game, save for the occasional expansion pack. However, the industry realized that it was more profitable to give away a game for free, and charge for loot boxes, cosmetic items, or have gacha systems (gambling for characters/items). With these systems, most customers would end up spending considerably more per year on a game, than the $60 upfront cost. Further, they would see all the purchases as sunk costs, and continue to play the game as well as make future purchases, leading to many individuals spending thousands of dollars on a single game. Today, this model has become so normalized that companies sell a game upfront for around $80 CAD, and then have lootboxes, cosmetics, or gacaha systems ontop of that. Further, many games launch with missing content and require addon purchases to complete the game. In environments like this, you should seek to understand the impact of potential sunk costs in your industry, and leverage sunk costs when possible (assuming it’s ethically acceptable). Keep in mind that consumers often get upset if they start having to pay more for the same products/services, so there are tradeoffs that need to be considered. Thus there can be brand-related costs involved with being the first in an industry to switch to a sunk cost model of profit. Natural sunk costs can also hinder profit in some industries. Consider cars or mobile phones that are often costly, paid off over time, involve additional accessories, and can require ongoing maintenance. A car or mobile phone company likely wants customers to purchase new products (e.g. to upgrade), but their customers have sunk costs in respect to their existing technologies. In this situation, marketing strategy should focus on creating value propositions that help consumers overcome their sunk cost fallacy. Okada (2006) suggests that to accomplish this, products need to be perceived as dissimilar, such as through the addition of new/additional features rather than just upgrades to existing features.

The second key time you may want to consider sunk costs is in respect to the creation of loyalty and membership programs. These types of programs can take a number of different forms, such as loyalty points that you accumulate for free or discounted goods, membership benefits like special sales/rates and early access to merch drops, special deals like a free item for every ten purchases, or restricted access to stores or events. As such these programs act as a sunk cost in two distinct ways. First, some of them require an upfront cash payment or annual fee to retain a membership, creating a sunk cost that can drive sales. Second, even if free, by accumulating points or rewards, these programs create a sunk cost that you may lose out on if you switch to an alternative brand. For example, if you have 5000 points, and 7000 points gets you a desired item, the sunk cost fallacy may kick in, even if alternative choices are more economically rational (e.g. you would save more money, even when you consider the existing sunk cost). Various studies have found these programs to be quite effective. For example, Chaudhuri et al., (2019) examined 322 new loyalty programs and found that on average they increased both sales and gross profit in the short term (1 year) and long term (3+ years). Similarly, as a result of the sunk cost fallacy, Eason et al. (2015) and Iyengar et al. (2022) found that paid loyalty programs and subscriptions (e.g. Amazon Prime) often lead to increased purchase intentions, and Ashely et al. (2016) found that they improve attitudes toward the brand. As a result, paid membership programs can help you determine who your best, most profitable, and most loyal customers are. For example, a lab study by Jang et al. (2007) found that 20 to 28% of those in a paid restaurant membership condition chose to stay loyal even when a better alternative choice became available. However, paid and free membership to loyalty or subscription programs likely differs between industries and contexts. Thus, it would be wise to research your industry prior to implementing one of these programs, and consider how sunk costs may influence customers, as well as what competitors are offering.

Practical Examples of The Sunk Cost Fallacy

When To Bundle Products and Services

If your product or service can be bundled (e.g. selling 4 tickets separately or selling a bundle of 4 tickets) there are considerations in relation to sunk costs that you may want to make. Soman and Gourville (2001) carried out a number of studies examining the relationship between prince bundling and the sunk cost fallacy. In one study, they introduced a scenario where the participant was asked to imagine going on a 4-day ski trip where the final day has bad weather. Those who were told they purchased a bundled 4-day pass were less likely to ski on the 4th day than those who were told they purchased 4 separate 1-day tickets. They then looked at four real-life Shakespearean theatrical performances, and found that the more tickets an individual purchased, the less likely they were to use them. This pattern demonstrates that perception of a direct relationship between a purchase and an event elicits a greater sunk cost effect. When products are bundled, individuals experience what Soman and Gourville (2001) call transaction decoupling, where individuals dissociate the costs from the benefits. Put another way, when you buy a single ticket, its relationship with the event is more salient (eliciting the sunk cost effect) than when you buy a set of bundled tickets. Thus, if you want to elicit the sunk cost effect, you should avoid bundling goods. For example, if you sell chocolate bars, selling a single bar might lead to a higher chance of it being consumed (due to sunk cost considerations) than selling your bars in a bundle. On the other hand, if you were selling extremely expensive chocolate bars the opposite may occur. Selling a single unit might elicit sunk cost considerations, which in turn might lead a consumer to save the expensive bar for a special occasion. Thus, it might be most profitable to sell normal chocolate bars alone, but expensive chocolate bars in bundles, to drive consumption and increase the chance of a repeat purchase. This demonstrates the importance of considering how the sunk cost effect might influence consumer decision making in respect to what you are selling. And you might find that different target markets are impacted by sunk costs in different ways. For example, if consumers in our chocolate bar example were extremely wealthy, they might not differentiate between normal and “expensive” bars, suggesting that single sales might be more profitable than group sales, as the wealthy may not have any incentive or inclination to save a conventionally expensive bar in response to sunk cost considerations. If you think that you would benefit from a consumer not using a purchased good/service, you may want to bundle the goods or services you are selling. For example, if you offer Pilates training services, bundling ten 1-hour sessions may result in customers being less likely to use up all sessions (as the sessions are perceptually decoupled from the sunk costs), giving you more free time for the same profit. We have oversimplified this explanation a bit, so we suggest reading Soman and Gourville (2001) if you are considering whether or not to bundle a product/service. Further, there are other variables in addition to sunk costs that are context-dependent and need to be considered. For example, research has found that stockpiling products that are convenient to consume (e.g. chocolate) might make them more salient and thus lead to more consumption (Chandon & Wansink, 2002).

Creating Sunk Costs in SaaS through Data/Project Storage

If offering software as a service, allow your users to store their data or projects on your server. But use a unique format that is not transferable to other platforms, or difficult/time-consuming/expensive to transfer. This way, the time spent creating projects and data stored on your platform becomes a sunk cost. Thus, individuals will be less likely to unsubscribe, as they will experience negative emotions related to losing all the work they have completed. And if you charge a low price each year, they will likely find it more emotionally convenient to keep paying, especially if they are subscribed through their business. In respect to this strategy, keep in mind that the research on considerations of time investment are somewhat inconclusive. Soman (2001) found that the sunk cost effect did not occur for considerations in respect to time investment unless an equal monetary cost was made salient. However, Navarro and Fantino (2009) found a sunk time effect across multiple studies. For example, in one study, nearly half of the students who spent a lot of time working to complete a 500-piece puzzle and were given the opportunity to quit decided to persist in finishing the puzzle. Further, Magalhães and White (2016) perform a review of studies on both humans and animals, and suggest that the sunk time effect is apparent not just in humans but also in non-human animals, suggesting an evolutionary origin rather than a cultural one. Anecdotally however, if you play videogames, we likely don’t have to convince you of the sunk time effect. Much like with the previously mentioned puzzle experiment, most gamers can likely recall at least one game in which they continued to play even when it wasn’t fun, simply because they invested time into building/leveling/gearing up a character, and not because they enjoyed the game. To be safe however, you could offer a manual file export service and charge a high fee (e.g. $500). This would make the financial cost of giving up the existing time investment salient. A smart competitor may offer the same service to transfer your files to their system for free or for a low cost by outsourcing the labour. Thus, you may want to include legal restrictions against third-party access to accounts in your terms of service (but speak with a lawyer about your options first).

Using Emails to Make Sunk Costs Salient

If you are in an industry that has a series of sunk costs like education, you can use emails to make those sunk costs more salient. For example, you might send a congratulations email to students for every course completed, that mentions the number of courses they have already completed. This makes existing sunk costs salient, and gives you an opportunity to drive future sales towards completing a certificate, while at the same time inducing positive emotions by celebrating your customer’s accomplishments.

Game Design and Sunk Costs

If you are launching a game that utilizes a gacha (gambling) system, loot boxes, or some other ongoing financial investment in character or account development, you can incentivize sunk costs at launch. By offering huge sales or deals and allowing players to get key limited time units easily that will be useful for quite a while into the future (albeit not without further investment) you can get a large amount of players to invest in their new accounts. Then, months down the line, you can start charging more for the same opportunities at account growth (e.g. make it more expensive to acquire desired characters) or add additional features that require additional investment to keep an account competitive and up to date. Many people who invested in their account at the start of the game when costs were low, now see those investments as a sunk cost, which helps them justify present purchases even if they have less utility than previous purchases, and creates pressure to continue playing even if faced with alternative choices that would lead to greater utility (e.g. better/cheaper games). This is currently how most mobile games operate. Most gamers hate this model and wish it would be banned, but video game companies recognize the profit opportunities and have lobbied hard to prevent governments from imposing restrictions (see Singletary Jr., 2019). Further, keep in mind that this model promotes and normalizes forms of gambling towards children and thus is arguably unethical.

Sunk Costs as Grandfathered-in Pricing

Some companies allow their users to keep an original subscription rate that no longer exists. This combines loss aversion with the sunk cost effect to keep customers subscribed even if they no longer use the service. Customers recognize that unsubscribing would cause them to lose a good deal. And they view their past payments towards the subscription service as a sunk cost in respect to their current special rate. This is especially true if they don’t use the service often. Consider an individual who has remained subscribed to a streaming platform for three years but has only watched one season of one show. If they are grandfathered into a lower rate than new subscribers, they likely view their ongoing subscription as a sunk cost that allowed them to maintain that rate (as they weren’t actually getting any value out of the subscription through usage). Thus, if they cancel their subscription, it feels like all the money spent was wasted as they had only ever watched one show. However, if they continue to subscribe, they can psychologically justify the past sunk costs as allowing them to keep the lower rate in case they want to watch shows on the platform in the future. These customers may be ideal, as they continuously pay a fee but use minimal to no resources. Thus, even though they pay less than new customers, they may be considerably more profitable in the long run. It can be tempting to increase subscription costs across-the-board to meet short-term targets/metrics, especially when annual bonuses are involved. But long term considerations in respect to the effect of sunk costs should always be considered.

Research Example of The Sunk Cost Fallacy

Bidding $2000 for $20

Keith Murnighan (2002) tells the true story of an in-class demonstration in an executive masters’ organizational behavior class of 70 students in 1998. And this story is amazing, so rather than read our summary below, we suggest you just read the actual article. Murnighan (2002) demonstrated the dollar auction in his class, where he pulled out a $20 bill and said he would auction it off. The dollar auction has a number of key rules. Bidding starts at $1, and only goes up in increments of $1. Players are warned before the auction ends. There is no collusion allowed. The highest bidder (the winner) pays what they bid and gets the $20. And the second highest bidder pays what they bid last but doesn’t get the $20. So if player A bid $8 and player B bid $7, player A would pay $8 and get $20 (a $12 profit), and player B would pay $7 and get nothing (a $7 loss). In this first auction, a winner ended up paying $54 for a $20 bill (a $34 loss) with the second highest bidder paying $53 (a $53 loss). Murninghan (2002) then pulled out a second $20 and auctioned it off again. He claims he regularly does this, sometimes up to four times, as students often overbid higher amounts in follow-up auctions than in the first. In this second attempt, the bidding reached $100 with two remaining students competing with each other. At this point, Murnighan changed the increments to $5, until the bids reached $400, at which point he changed the increments again to $10. Once the bids reached $700 he changed the increments to $20, until the bids hit $1200 at which point he changed the increments to $50. The bidding went on until the two remaining participants reached $2000 (a $1980 loss for the “winner”) and $1950 (a $1950 loss). Because the increments were greater than $20 at this point, the winner actually lost more money than the second place bidder. Murnighan (2002) didn’t force them to pay the money back, he collected $50 from each, and combined with the first auction collected a total of $187 that went to charity. Murnighan (2002) suggests that emotions played a huge role in these individual’s decision-making processes. Which, as discussed on this page, may drive the sunk cost effect. Thus, it’s almost certain that the sunk cost effect played some role in these extreme bids, demonstrating the power of our heuristic processes.

Works Cited

Ashley, C., Gillespie, E. A., & Noble, S. M. (2016). The effect of loyalty program fees on program perceptions and engagement. Journal of Business Research, 69(2), 964-973. https://doi.org/10.1016/j.jbusres.2015.09.001

Chandon, P., & Wansink, B. (2002). When Are Stockpiled Products Consumed Faster? A convenience-salience framework of postpurchase consumption incidence and quantity. Journal of Marketing Research, 39(3), 321-335. https://doi.org/10.1509/jmkr.39.3.321.19111

Chaudhuri, M., Voorhees, C. M., & Beck, J. M. (2019). The effects of loyalty program introduction and design on short- and long-term sales and gross profits. Journal of the Academy of Marketing Science, 47(4), 640-658. https://doi.org/10.1007/s11747-019-00652-y

Dijkstra, K. A., & Hong, Y. (2019). The feeling of throwing good money after bad: The role of affective reaction in the sunk-cost fallacy. PLoS ONE 14(1), e0209900. https://doi.org/10.1371/journal.pone.0209900

Eason, C. C., Bing, M. N., & Smothers, J. (2015). Reward me, charity, or both? The impact of fees and benefits in loyalty programs. Journal of Retailing and Consumer Services, 25, 71–80. https://doi.org/10.1016/j.jretconser.2015.04.001

Fujino, J., Fujimoto, S., Kodaka, F., Camerer, C. F., Kawada, R., Tsurumi, K., Tei, S., Isobe, M., Miyata, J., Sugihara, G., Yamada, M., Fukuyama, H., Murai, T., & Takahashi, H. (2016). Neural mechanisms and personality correlates of the sunk cost effect. Scientific Reports, 6, 33171. https://doi.org/10.1038/srep33171

Hafenbrack, A. C., Kinias, Z., & Barsade, S. G. (2014). Debiasing the mind through meditation: Mindfulness and the sunk-cost bias. Psychological Science, 25(2), 369-376. https://doi.org/10.1177/0956797613503853

Haller, A., & Schwabe, L. (2014). Sunk costs in the human brain. Neruoimage, 15(97), 127-133. https://doi.org/10.1016/j.neuroimage.2014.04.036

Iyengar, R., Park, Y.-H., & Yu, Q. (2022). The impact of subscription programs on customer purchases. Journal of Marketing Research, 59(6), 1101-1119. https://doi.org/10.1177/00222437221080163

Jang, D., Mattila, A. S., & Bai, B. (2007). Restaurant membership fee and customer choice: The effects of sunk cost and feelings of regret. International Journal of Hospitality Management, 26(3), 687-697. https://doi.org/10.1016/j.ijhm.2006.06.001

Liang, A. R., Lee, C., & Tung, W. (2014). The role of sunk costs in online consumer decision-making. Electronic Commerce Research and Applications, 13(1), 56-68. https://doi.org/10.1016/j.elerap.2013.10.001

Magalhães, P., & Geoffrey White, K. (2016). The sunk cost effect across species: A review of persistence in a course of action due to prior investment. Journal of the Experimental Analysis of Behavior, 105(3), 339-361. https://doi.org/10.1002/jeab.202

Murnighan, J. K. (2002). A Very Extreme Case of the Dollar Auction. Journal of Management Education, 26(1), 56-69. https://doi.org/10.1177/105256290202600105

Navarro, A. D., & Fantino, E. (2009). The sunk-time effect: An exploration. Journal of Behavioral Decision Making, 22(3), 252-270. https://doi.org/10.1002/bdm.624

Okada, E. M. (2006). Upgrades and new purchases. Journal of Marketing, 70(4), 92-102. https://doi.org/10.1509/jmkg.70.4.092

Roth, S., Robbert, T., & Straus, L. (2014). On the sunk-cost effect in economic decision-making: A meta-analytic review. Business Research, 8, 99-138. https://doi.org/10.1007/s40685-014-0014-8

Singletary Jr., C. (2019, May 24). Video game lobbying group trashes proposed anti-loot box bill. Digital Trends. https://www.digitaltrends.com/gaming/esa-loot-box-ban-bill-reponse/

Soman, D. (2001). The mental accounting of sunk time costs: why time is not like money. Journal of Behavioral Decision Making, 14(3), 169-185. https://doi.org/10.1002/bdm.370

Soman, D., & Gourville, J. T. (2001). Transaction decoupling: How price bundling affects the decision to consume. Journal of Marketing Research, 38(1), 30-44. https://doi.org/10.1509/jmkr.38.1.30.18828

Strough, H., Mehta, C. M., McFall, J. P., & Schuller, K. L. (2008). Are older adults less subject to the sunk-cost fallacy than younger adults? Psychological Science, 19(7), 650-652. https://doi.org/10.1111/j.1467-9280.2008.02138

Important Links

Discover More Knowledge Contact Us