When something succeeds, product teams exhibit a certain level of confidence. A checkout process improves conversion by 20%. People return to the app after receiving a reminder email. A small, partially filled progress bar encourages users to complete the task at hand. It is shipped everywhere by the team. Then the numbers from Stockholm, São Paulo, or Seoul arrive, and something is wrong. Not exactly broken. Quietly incorrect.
I’ve seen this occur several times, and the engineers never seem to be as taken aback as they ought to be. Most persuasive software operates under the premise that a nudge is a kind of universal lever. Behavior shifts when you push here. The tech industry adopted the neat vocabulary that Thaler and Sunstein provided, incorporating choice architecture into default toggles, onboarding screens, and those slightly embarrassing “Are you sure you want to leave?” pop-ups. It is believed that human psychology is essentially universal. That might be true on some profound level. It’s also possible that the most variable part is the one around which we really build software.
Think about the language of urgency. Directness is often interpreted as helpfulness in American apps. “Don’t miss out.” “Only 2 left.” “Finish setting up your account now.” The same wording may come across as pushy or even slightly disrespectful in societies that value indirect communication. Speaking with designers who have worked in a variety of markets, it seems that the West’s emphasis on individual success and direct calls to action just doesn’t translate to regions where group harmony and maintaining one’s dignity influence how people react to instructions. A confident nudge may come across as haughtiness in other contexts.
This is a recurring theme in the research. A carefully crafted nudge message was tested in Japan, Canada, and the US in a cross-cultural study published in 2022; the intervention was successful in all three nations. However, the effect size fluctuated. In the US and Japan, women reacted more strongly than men, but there was no discernible sex difference in Canada. Strangely, the pandemic weakened the message in some groups by altering the level of support that people already experienced on a daily basis. Universality was the main discovery. Beneath the segments, the intriguing discovery was that the same words performed noticeably differently depending on who was reading them and where.
Product teams tend to underestimate this. The text is localized by them. For right-to-left languages, they might reverse the layout, translate the strings, and change the currency symbol. The persuasion itself, the underlying social logic of why the nudge is supposed to be effective, is what they hardly ever localize. In some contexts, a reciprocity cue that seems generous might be perceived as an imposition of debt. Social proof, that dependable “join 10,000 others” line, is predicated on the idea that people aspire to be like the group that is presented to them. According to Cialdini’s own observations, we imitate people the most when we perceive them to be similar to ourselves. If you show a wall of American testimonials to a German user, you might have created persuasion that subtly leads them astray.
It’s difficult to ignore how frequently this is discovered after launch, typically as a vague underperformance that no one can adequately explain. The funnel is blamed by teams. More A/B tests are conducted in the same defective frame. Seldom does the deeper problem—that persuasion is cultural before it is technical—find its way into the retro.
This is not just a commercial expense; there is a true cost. Software that assumes what constitutes a good decision is known as nudging software. If you export that without reviewing it, you’re exporting both the product and a worldview. It’s still unclear if businesses will start translating persuasion in the same manner that they translate words. It appears that those who were paying attention were the ones who were burned first.
