Loss aversion is the tendency to feel a loss more strongly than a gain of the same size, so losing $50 hurts more than finding $50 feels good.
A 2024 meta-analysis of 607 estimates put the average weight at about 1.96: losses count roughly twice as much as equal gains.
In marketing, that is why "don't lose" framing, guarantees and deadlines can move behavior, but A/B tests show the effect depends on context and often backfires.
This page separates what the research shows from what real A/B tests in our test database show.
For related biases, see the psychology hub and the overview of cognitive biases.
Loss aversion examples: from the lab to real A/B tests
Where the idea comes from
Daniel Kahneman and Amos Tversky introduced loss aversion as part of prospect theory in "Prospect Theory: An Analysis of Decision under Risk" (Econometrica, 1979).
Their key claim was that people judge outcomes as gains or losses from a reference point, not as final wealth, and that the value curve is steeper for losses than for gains.
The famous "losses loom twice as large" number comes from their 1992 cumulative prospect theory paper, which estimated a median loss aversion coefficient of 2.25. That estimate was fit to the choices of a small group of graduate students, so treat it as a classic starting point, not a universal constant.
How strong is the effect, really?
The best summary today is a 2024 meta-analysis in the Journal of Economic Literature by Alexander Brown, Taisuke Imai, Ferdinand Vieider and Colin Camerer.
Across 607 estimates from 150 articles, the mean coefficient was 1.955, with a 95% credible interval of 1.82 to 2.10.
The core choice patterns also replicate. A 2020 replication of the original 1979 study, published in Nature Human Behaviour with 4,098 participants in 19 countries, reproduced 94% of the items, with some weaker effects.
There are real caveats, though:
- Small stakes. Eldad Yechiam's review "Acceptable losses: the debatable origins of loss aversion" (Psychological Research) argues that for small to moderate amounts, people often show no loss aversion at all. Most ecommerce purchases are small stakes.
- Overgeneralization. David Gal and Derek Rucker's "The Loss of Loss Aversion" (Journal of Consumer Psychology, 2018) argues the idea has been stretched into a general law of behavior that the evidence does not support.
- Framing is not the same thing. Saying "don't miss out" instead of "get this" is message framing. A meta-analysis of 93 studies by O'Keefe and Jensen (2007) found that for disease prevention messages, gain-framed messages were very slightly more persuasive (r = .03), not less. Loss framing in copy is not a reliable win.
A field example: teacher bonuses
One of the clearest real-world tests is a field experiment by Roland Fryer, Steven Levitt, John List and Sally Sadoff in Chicago Heights schools.
Some teachers got a bonus up front and had to pay it back if student results fell short.
That "loss" version raised math scores by 0.2 to 0.4 standard deviations, while the traditional end-of-year bonus showed smaller, non-significant effects.
The lesson for marketers: loss aversion is strongest when something already feels owned.
That links it closely to the endowment effect.
Loss aversion examples from our test database
Here are A/B tests where a loss frame, or removing the fear of a loss, changed results.
Numbers are as reported by whoever ran the test.
- Fear-of-consequences product copy. Dave Diederen tested four product description approaches: control, fear, value and problem. The fear version, which described the health risks of not using the product, won with +32.75% conversion rate and +41.90% revenue per visitor, about $102,665 a month.
- Pain-point copy. Oliver Kenyon replaced general product claims with copy about soreness, fatigue and chronic pain, and the test added $159K per month, $1.9M a year forecast.
- Framing a benefit as already earned. Dave Diederen changed cart copy from "Free shipping will be applied at checkout" to "Congrats! You unlocked FREE shipping!" The result was +11.08% conversion rate and +13.93% revenue per visitor. Once a shopper owns the reward, leaving means giving it up.
- Removing the risk of losing money. Dave Diederen added a prominent 60-day guarantee badge below the add-to-cart button, and conversion rate rose 14.50%. Visually reported that a '60 Day Returns' line under Faherty's cart checkout button gave +10.5% conversion.
- Showing what you save. Putting the savings in the button ("Add to Cart - Save 50%") gave Dave Diederen +10.22% conversion rate and +14.41% revenue per visitor.
@oliverkenyon
@DaveDiederen@DaveDiederen@DaveDiederen@DaveDiederen
VisuallyNone of these tests isolated loss aversion alone.
Copy tests change clarity, specificity and tone at the same time, so a win is consistent with loss aversion rather than proof of it.
Loss aversion marketing: what to test, and when it backfires
In practice, loss aversion shows up in four kinds of page changes.
Each has tests that won and tests that lost.
| Tactic | The loss it plays on | Won | Lost or backfired |
|---|---|---|---|
| Deadlines and countdown timers | Losing a deal or a delivery date | Benchmade shipping-cutoff countdown, +32.4% revenue per session | Timers in European markets, -35.61% revenue per visitor |
| Guarantees and returns | Losing money on a bad purchase | 60-day guarantee badge, +14.50% conversion | Return policy on product pages, -6.0% conversion |
| Already-owned rewards | Giving up something earned | "You unlocked FREE shipping", +11.08% conversion | No clear loser in our data yet |
| Trials and paywalls | Losing access after using a product | Acuity 14-day Pro trial, +268.14% paid sign-ups | BoldVoice hard paywall, refunds outgrew the gain |
Urgency only works when the loss is real
Countdown timers are the most common loss-aversion tactic, and the most mixed.
Dave Diederen found that a checkout "order reserved" timer for US brands gave +9.50% conversion rate and +4.83% revenue per visitor, with AOV down 4.26%.
The same practitioner reported that across 20+ brands in non-US markets, timers often hurt, with one result of -17.35% conversion rate and -35.61% revenue per visitor.
Kasper Dolk saw the reverse side for a Dutch audience: removing the checkout timer raised conversion 25.61% and revenue per visitor 26.18%, after survey answers said the timer felt like too much pressure.
A real deadline is a different case. Benchmade's countdown to the Christmas 2-day shipping cutoff was tied to an actual loss, missing the holiday, and Visually reported +32.4% revenue per session.
More detail is on the countdown timers page and in scarcity and urgency.
Reducing loss beats threatening loss
Guarantees work on loss aversion from the other side: they take away the fear of losing money.
That tends to be safer than adding threats, but placement and wording still matter.
Clean Commit found a return policy on product pages lowered conversion 6.0%, possibly because it raised the thought of returning before the shopper had decided to buy. Browse more on the money-back guarantees page.
Let people own it first
The strongest loss frames come after a sense of ownership: a trial, an unlocked reward, items saved in a cart.
The Acuity Scheduling result is the clearest SaaS example, though that test also added a card requirement, so it changed the offer, not just the framing. See free trials for more SaaS tests.
Free-shipping and free-gift progress bars use the same idea.
Read the free shipping thresholds page before setting one, because the threshold amount matters as much as the copy.
When loss framing is the wrong choice
- Small, low-risk purchases. Research suggests loss aversion weakens or disappears at small stakes, so a threat-based headline on a $15 product has little to work with.
- Fake deadlines or fake scarcity. Timers that reset, or "only 2 left" claims that aren't true, are dark patterns. They can win a short test and still erode trust and repeat purchases.
- Audiences that distrust pressure. The European timer results show that the same tactic can win in one market and lose badly in another. Segment results by country before rolling out.
- When the metric is too narrow. BoldVoice's paywall with no close button produced more trial starts, but refunds rose by more than the gain. Measure revenue, refunds and churn, not only clicks.
How to test it
Write the gain and loss versions of the same message, keep everything else equal, and run them as a clean A/B test.
For example: "Get clearer skin in 30 days" versus "Stop wasting money on products that don't work."
Pick revenue per visitor as the main metric, and check it holds across devices and countries.
If the loss version wins, you have evidence for your audience, not a rule.
The framing effect page covers how to write paired versions.