The IPIP-NEO

My product is a free personality test: the IPIP-NEO, a public domain Big Five inventory with 300 questions in the long form and 120 in the short one. https://ipipneo.com/en

Before I put it in front of anyone I did something most people would call a mistake: I localised all of it into 32 languages. Not the buttons. The whole thing, including thirty facet write-ups, five long domain articles and the interpretation texts, plus the questionnaire itself.

Why I did it:

1. This niche has a shape. Search demand for "big five test" is enormous in English and almost unserved in twenty other languages, where people land on machine translated pages that read like nonsense.

2. The content is the product. If the differentiator is depth, then depth in one language is one market and depth in thirty two is thirty two.

3. It gets harder later, not easier. Every new page multiplies by the number of locales, so the cheapest moment to be multilingual is before you have many pages.

The cost was real. Verification scripts per locale, a glossary, gender neutrality rules for languages that inflect for the reader, and about three times the calendar time I planned.

What I still do not know is whether the non-English traffic converts at all, and I will not know for months. If you want to see the depth, the long version is here: https://ipipneo.com/en/ipip-neo-300

Has anyone here launched multilingual from day one? Did the extra locales pay for themselves, or did the English version end up carrying everything anyway?