Product teams often treat copy as finishing, something to adjust once the feature is ready. In research, I see the opposite: if people don’t recognize the word, they don’t recognize the feature.
“Match” is not “connect”
Gameram is a social app where gamers find people to play with, available across the Americas. I ran research in English, Portuguese and Spanish, and with Latin American users the product’s wording needed adjusting. We adapted it for the region: “match” became “connect”.
The wording change contributed to a 20% increase in conversion in Latin America.
A translation can be correct and still not be the word people would use. That gap rarely shows up in a copy review. It shows up when you talk to the people who use the product, in their own language.
Terminology that didn’t match
At Materialize, I led the research to move in-person courses online within four months. One of the findings was a set of terminology mismatches. Fixing them significantly increased enrollment.
Again, nothing new had to be built. What already existed started being found and understood.
What I take from this
- Test words with the people who will use the product. Translation and localization are different steps, and only the second one goes through real people.
- Treat copy as a hypothesis. A word change can be tested like any other product change.
- Look at language when adoption is low. Before redesigning a screen, ask whether people understand the name of what they’re looking at.
A word is the smallest possible change to a product. Sometimes it’s the one that pays off most.
Some details have been omitted for confidentiality.