Shoppers can tell when a website was only translated, not localised. More than half have walked away because of it
A new survey suggests businesses expanding into multiple languages may be solving the wrong problem. The Liferay 2026 Website Localization Report, conducted in September 2026 through the polling platform Pollfish, surveyed 800 U.S. adults who had used a website, app or digital service in a language other than English over the past year. Among them, 56 per cent said they had decided against buying from a company because its site felt poorly adapted to their language, and 90 per cent had noticed at least one localisation problem somewhere online in the past year.
Good translation isn’t the fix
The more striking finding sits a layer deeper. Even among consumers who rated a brand’s translation as good or excellent, 54 per cent had still abandoned a purchase over the broader localisation experience. Only 23 per cent said non-English websites usually felt like they’d been written specifically for them.
“You can translate every page and technically have everything working, but that doesn’t mean the experience feels natural to the person using it,” said Bryan Cheung, co-founder and CMO at Liferay, which commissioned the research. “Translation gets the content into another language. Localisation is what makes it feel natural to the people using it.”
Consumers also don’t think they’re being fooled by machine output. 82 per cent said they’re confident they can tell when content has been machine-translated rather than written or reviewed by a fluent speaker. On the research team’s own numbers, that confidence may be well placed: only 20 per cent of respondents credited AI alone for a strong localised experience, against 43 per cent who pointed to human involvement as the deciding factor.
There’s a commercial upside buried in the same data. 48 per cent of respondents said they had chosen one company over a competitor specifically because its website or app felt better adapted to their language and culture, meaning the same gap that costs sales for one brand is a genuine point of difference for whichever competitor closes it.
A U.S. study, an Australian problem
The sample here is entirely American, so the exact percentages shouldn’t be read onto the Australian market. But the underlying dynamic is arguably more relevant here than in the U.S. Close to a fifth of Australians, nearly 4.5 million people, speak a language other than English at home, and migrants are estimated to own about a third of the country’s small businesses, according to figures cited by the Australian Small Business and Family Enterprise Ombudsman. For businesses selling to Australia’s own multicultural population, let alone anyone exporting into Asian or European markets, the distinction this survey draws between translating a page and actually localising it is a live one, not a theoretical one.