Same money. Same options. Different label.

Imagine two people open the same pension-app notification. One version says don't miss out on the tax break; avoid losing it. The other says secure it and benefit from it. Same money. Same deadline. Different label.

One of them tops up. The other swipes it away. The company, looking at the average, decides the wording didn't matter.

It did. Just in opposite directions.

1"Don't miss out" works on some people and backfires on others

A European retirement-savings company — most of its customers in Switzerland — randomly sent its annual tax-deduction reminder in either a loss frame or a gain frame (Bösch et al., Scientific Reports, 2 Nov 2026). One experiment covered 22,919 customers; a second, the next year, covered 18,288. The reminders went out, in the app and by email, at the beginning of November.¹

Across everyone, the two frames came out about even.¹ Underneath that flat average, the quarter most responsive to loss framing were an estimated 9.4 points more likely to contribute, and the quarter most responsive to gain framing were 8.4 points less likely.¹ Among people who contributed, the loss frame is estimated to have raised contributions for 64% of them, by about €737 on average, and cut them for the rest, by about €615 on average.¹

Matching the frame to the person is estimated to lift participation by 51%, and average savings by 18%, relative to sending everyone the loss frame.¹ Estimated, because nobody actually rolled that policy out and measured it. A model predicted what would have happened.

So "no average effect" is not the same as "no effect". Before you scrap a message, a policy or a habit because it "didn't work on average", ask for whom? And when the message is aimed at you, try reading it in the other frame before you act.

2The $5 you'd drive for, and the $5 you wouldn't

Say you're buying a $15 charger and a $125 jacket. The shop assistant mentions the charger is $5 cheaper at their other branch, twenty minutes away. You grab your keys — a third off is a third off. Ten minutes later your partner is at the same counter with the same two items, but the $5 is off the jacket. They shrug and pay. Same $5. Same twenty minutes. Different label on the saving.

A second version of the same quirk: you've lost your concert ticket at the door. Paying for another feels like paying twice for the show. If you'd lost the same amount in cash on the way in, you'd buy the ticket without a second thought.

Both of those are versions of two classic mental-accounting scenarios, and researchers ran seven of them on 5,589 adults from 21 countries (Priolo et al., Journal of Consumer Research, published July 2026).² All seven effects held up in the pooled analysis. Country by country, 133 of 147 effects (90.5%).² The effects were generally smaller than the originals.² Nobody spent real money — these were hypothetical scenarios answered online — but the pattern is stubborn enough to show up across a lot of places.

Rule of thumb: price a saving in dollars and minutes, not as a percentage. "Would I drive twenty minutes for $5 cash in hand?" And for sunk costs: "If I'd lost the cash instead, would I still buy it?"

3The list that chooses for you

Picture a family picking a summer holiday. Someone sends the shortlist: "Noosa, Byron, Margaret River, Port Douglas… or somewhere overseas." Four beach towns each get their own line. The whole rest of the world gets three words. Nobody argues about it, and Noosa wins.

Something similar happens at work. A training-budget form lists five in-house courses individually and puts every outside conference on a single line marked "External (specify)". Nobody hid any option. The grouping quietly told everyone which ones were the normal ones to pick.

Across ten experiments, researchers put six options on a menu and varied only how they were grouped (Tannenbaum, Fox & Goldstein, Organizational Behavior and Human Decision Processes, Nov 2026). People chose from whichever category was listed item by item far more often — shifting choice share by an average of 32 percentage points.³ When real prizes were at stake, one study found a 36-point shift.³ The reason, the authors argue, is that people infer the separately listed options are more popular on average.³

Before you choose from any list, regroup it yourself. Count the categories, not the lines, and ask: if the other group were split out one by one, would I still pick this? If you're the one writing the list, give each serious option the same amount of space.

Putting it together

Three labels, then. A reminder worded as a loss or a gain. A $5 saving stuck to a charger or to a jacket. A menu split one way or the other. Not one of them changed the options. Every one of them changed what some people did.

Which is a little unsettling, if you sit with it. Someone chose the wording on the last email that asked for your money, and the layout of the last form you filled in. Sometimes on purpose, often not. Either way, the label got a vote.

So the habit worth building is a small one. When a choice lands in front of you, look at the label before you look at the options. Then swap it. Loss for gain. Percentage for dollars. One line for five. If your answer changes, the label was doing some of the deciding.