The framing effect: why "95% of happy customers" converts better than "5% disappointed"

A surgeon tells a patient that an operation has a "90% success rate." Another surgeon, for the same operation and the same statistics, says there's a "10% risk of death." Both sentences describe exactly the same rate — yet in studies on medical decision-making, patients agree to the operation far more often when it's presented the first way. This isn't a lie or crude manipulation: it's the framing effect, one of the most robust findings in decision psychology. On a landing page, the same mechanism kicks in every time a satisfaction rate, a price, or a guarantee can be phrased as a gain or a loss: the framing you choose, with identical information, moves conversion.

The founding experiment: the Asian disease problem

The framing effect was formalized in 1981 by Amos Tversky and Daniel Kahneman in a landmark paper, "The Framing of Decisions and the Psychology of Choice", published in Science. Their most-cited experiment presents participants with a fictional scenario: a disease threatens 600 people, and a public health program must be chosen. One group receives a gain frame ("program A saves 200 people"); a second group receives a loss frame, mathematically identical ("program A lets 400 people die"). The program is exactly the same in both cases — only its wording changes. The result: a clear majority picks the safe option when it's framed as gains, and the risky option when the same option is framed as losses. The authors conclude that people don't react to raw information, but to how it's presented relative to a reference point.

What the framing effect is not

It's tempting to confuse framing with loss aversion, but the two describe distinct mechanisms. Loss aversion says a €20 loss weighs psychologically more than a €20 gain — it's an asymmetry of value. The framing effect doesn't compare an actual gain to an actual loss: it shows that one and the same reality, reworded in gain language or loss language, changes the decision even though nothing objective has changed. The second phenomenon explains why the first is so easy to exploit in copywriting: since losses weigh more heavily, framing information as loss-avoidance ("don't miss out", "your spot isn't guaranteed") pulls a different — and often stronger — lever than simply framing a rate positively.

Application #1: satisfaction rates and product attributes

This is the most direct application on a landing page: every statistic can be framed as a gain (what works) or a loss (what fails), and the choice is never neutral. A study by Irwin Levin and Gary Gaeth (1988) illustrates this with an example that became a marketing classic: consumers rated the same ground beef presented as either "75% lean" or "25% fat." The information is identical down to the percentage, but the beef framed as "75% lean" was rated noticeably better — until participants actually tasted the product, at which point the framing effect faded against direct experience. The lesson for a landing page: a number presented positively ("96% on-time deliveries") reads more favorably than its negative equivalent ("4% delays"), but that advantage doesn't excuse the need for tangible proof — a case study, a testimonial, or a product trial — that backs up what the framing promises.

Application #2: framing the price

• Monthly framing — "€8 a month" frames a subscription around a small, easy daily trade-off, while "€96 a year" frames the same amount as a single, heavier expense to evaluate all at once; which one to use mostly depends on context, covered in the article on monthly vs. annual pricing.
• Avoided cost rather than spend — "save €200 this year" frames the purchase as a gain, while "don't overpay €200" frames the same amount as an avoided loss; the second wording, closer to loss aversion, is often the more persuasive one for an audience already convinced of the product's value.
• Crossed-out price — showing a higher reference price before the discount frames the final price as a gain relative to that reference point, a mechanism close to the anchoring effect, but it only works if the crossed-out price stays credible.

Application #3: framing the guarantee and perceived risk

A money-back guarantee can be framed at least two ways for the exact same promise: "try it risk-free, refunded if it's not for you" emphasizes the absence of possible loss, while "you have 30 days to change your mind" emphasizes the gain of freedom. Both framings are honest, but they speak to different hesitations: the first reassures a visitor afraid of losing money, the second speaks to a visitor afraid of committing. On a page with high hesitation (expensive purchase, long commitment), loss-avoidance framing tends to remove the friction better, precisely because loss aversion amplifies its effect.

The limit: framing is not distorting

The framing effect works because the information stays, in theory, strictly equivalent from one wording to the other — 95% satisfied and 5% disappointed describe exactly the same rate. The technique becomes a problem the moment framing hides part of the information rather than rewording it: showing "from €29" for an entry price almost nobody actually pays, or a satisfaction rate computed on an unrepresentative sample, isn't framing anymore — it's deception, the same line drawn in the article on dark patterns. The simplest test: if the visitor, on learning the complementary figure (the 4% of delays, the 5% who were disappointed), would feel misled by the original wording, the framing has crossed its legitimate limit.

LanderKit templates (€89 each, €229 for the full 10-pack) keep satisfaction rates, guarantees, and pricing blocks fully editable, so you can test different framings without touching the code — notably on social-proof-driven pages like Coach & Consultant or E-commerce Product. This bias pairs naturally with loss aversion for word choice, and with the three-tier pricing table for visually framing the middle offer.