Made with 🧠 and 🫀 by Youssef Bouksim

Back to library
🖼️

Framing Effect

The same information, presented differently, produces completely different reactions. It is not what you say -- it is how you say it. And how you say it changes what people feel, decide, and do.

5 min readUX · Product · Marketing

“90% fat-free” and “10% fat” are mathematically identical. But they feel completely different. The first sounds like a healthy choice. The second sounds like something to avoid. The yoghurt hasn't changed. The number hasn't changed. Only the frame has — and with it, the entire perception.

Tversky and Kahneman documented this systematically in 1981. They gave participants a choice between two medical programmes to fight a disease expected to kill 600 people. Programme A would “save 200 people.” Programme B had a “one-third chance of saving all 600 people, and a two-thirds chance of saving no one.” Most people chose A. When the same programmes were reframed — A would result in “400 people dying,” B had a “one-third chance nobody dies” — most people switched to B. The outcomes were identical. The frames were not.

The mechanism behind this is loss aversion: losses feel roughly twice as painful as equivalent gains feel good. Framing something as avoiding a loss activates a stronger emotional response than framing it as achieving a gain — even when the underlying reality is the same. Designers who understand this can use framing to communicate genuine value more compellingly. Those who abuse it use it to exaggerate urgency, manufacture scarcity, and mislead users about what they're actually getting.

✦ Three types of framing that matter in product design
✓
Gain vs loss framing. “Save $20” and “Don't miss out on $20” describe the same saving — but the second activates loss aversion. “Cancel anytime” and “No commitment” say the same thing — but one frames the subscription as something you control, the other as something that doesn't trap you.
✓
Attribute framing. “95% uptime” and “down for 438 hours per year” are the same number. “Charges 1% per transaction” and “costs just $1 on a $100 sale” are the same fee. How you present an attribute determines whether it sounds impressive or alarming.
✓
Risky choice framing. When faced with uncertainty, people are more likely to take a risk to avoid a loss than to achieve a gain. “Your files are at risk without a backup plan” prompts more upgrades than “Get peace of mind with a backup plan” — because risk-avoidance motivation is stronger.
“The same information, framed differently, is not the same information. The frame is part of the message.”
— Tversky & Kahneman, Science, 1981

The same facts — feel the difference yourself

Every pair below contains exactly the same information. The numbers, the product, the offer — all identical. Only the frame changes. Click through each pair and notice which version feels more compelling, which feels more urgent, and which feels more honest.

9:41
Framing Compare1/6
Gain vs loss
A
Frame A
Save $29 a year by choosing the annual plan.
B
Frame B
Don’t lose $29 a year by staying on monthly.

A product page — two frames, same features

Below are two landing pages for the same cloud storage product. The storage capacity, price, features, and uptime are identical. What changes is which facts are surfaced, which are hidden, and whether they are framed as gains or losses. Both pages are technically accurate. One is designed to inform. The other is designed to convert at any cost.

Frame A — Honest, balanced
cloudsafe.app/pricing
CloudSafe Pro
2 TB of secure cloud storage
Automatic backup, file versioning, and 256-bit encryption. Available on all your devices. $9.99 per month, billed monthly. Cancel anytime.
No credit card required
99.5%
Uptime — about 44 hours downtime per year
30 days
Version history — restore any file from the last 30 days
1%
Transaction fee on additional storage purchases
2 TB storage (expandable)
256-bit AES encryption
Up to 5 connected devices
30-day version history
Priority customer support
Automatic nightly backups

Facts presented clearly: uptime with its downtime equivalent, version history with its limit, transaction fee disclosed upfront. No urgency. No scarcity.

Frame B — Manipulative framing, same facts
cloudsafe.app/pricing
Free trial ends in 2 days — do not lose your files
Your files are at risk
Stop losing files forever — protect 2 TB today
Every day without a backup is a day your memories, work, and data could be gone forever. CloudSafe Pro users never lose files. Do not be the person who wished they had signed up sooner.
Usually $19.99/mo
Limited offer: only $9.99/mo — save 50% today
Only 14 spots left at this price
99.5%
Industry-leading uptime guarantee
Files protected — never lose anything again
$1
Tiny fee on a $100 purchase — barely anything
Never run out of storage again
Hackers can’t touch your files
Sync across all your devices
Restore files before they’re gone forever
Never wait on hold again
Sleep soundly — backups run while you rest

Same product, same price — but the fee becomes ‘barely anything,’ the 30-day limit becomes ‘never lose anything,’ a fake original price manufactures a discount, and fear replaces facts.

Frame B is not just aggressive marketing — several elements cross into dark pattern territory. The fake “Usually $19.99/mo” is a fabricated anchor (the product never sold at that price). The “files protected -- never lose anything” stat is misleading — version history is 30 days, not infinite. The “Only 14 spots left” is fabricated scarcity for a digital product with unlimited capacity. These are not frames — they are lies wearing the clothes of framing.

The honest version of loss framing does exist. “Your files are not backed up — one hard drive failure and they are gone” is a true statement that uses loss framing to motivate genuine action. The difference is that the threat is real, the product genuinely addresses it, and no other facts are distorted to make the case.


Applying this to your work

Framing is unavoidable. Every product decision about what to call something, in what order to present options, and which numbers to surface is a framing decision. The question is whether you make those decisions deliberately and honestly, or by default and in the product's interest at the user's expense.

The honest framing rule: the frame you choose should make it easier for users to understand the real value, not harder for them to see the real limitations. “95% uptime” is fine — it is the real number. “95% uptime” while hiding that it equals 438 hours of downtime per year is selective framing designed to obscure. The difference is what you choose not to show.

✓ Apply it like this
→Use gain framing for genuine benefits: 'save $29/year' is cleaner and more motivating than the same number presented as a monthly amount.
→Use loss framing only when the risk is real and the product genuinely addresses it -- 'your data isn't backed up' is honest if the threat is true.
→Present both sides of a stat when the less flattering side is material -- uptime percentage alongside what it means in hours of downtime.
→Choose frames that help users make better decisions, not frames that prevent rational evaluation.
✗ Common mistakes
→Manufacturing urgency that does not exist -- countdown timers with no real deadline, scarcity claims for digital products with unlimited supply.
→Selective stat presentation -- showing '99.5% uptime' while hiding the downtime equivalent, or 'never lose anything' for a product with a 30-day limit.
→Loss framing based on manufactured fear -- 'your files are at risk' when there is no specific threat, just a product trying to sell a solution to a problem it invented.
→Feature copy designed to obscure rather than inform -- 'hackers can't touch your files' says nothing about what the encryption actually covers.

Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453--458. Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. Levin, I. P., & Gaeth, G. J. (1988). How consumers are affected by the framing of attribute information. Journal of Consumer Research, 15(3), 374--378.