Multilingual qualitative research: running one study across markets
By Juno · Published · Updated
Multilingual qualitative research means interviewing people in each market in their own language, while holding every interview to the same research objectives so the results can be compared. Done well, you end up with one study that covers several markets, rather than several separate studies that happen to share a topic.
Language is only part of the job. The rest is keeping the study the same study once it has passed through translation and local habits of speech. Below is how to approach that, and how Juno handles it.
Why interview people in their own language
People explain their reasoning best in the language they think in. In a second language, answers tend to get shorter and more general, and people reach for the words they know rather than the words they mean. Qualitative evidence lives in exactly those details, such as the hesitation before an answer or the example that doesn't quite fit.
There is a sampling reason too. If you only interview people who are comfortable in English, you hear from a narrower slice of each market than the one you are trying to understand. Interviewing in the participant's own language lets the sample reflect the market, not just its most fluent English speakers.
Keeping a study comparable across markets
The most reliable way to keep markets comparable is to build the study from research objectives, not a translated list of questions. An objective such as "understand what makes people trust a new energy provider" means the same thing in Madrid and Melbourne. A word-for-word question often doesn't.
Where you do need exact wording, for a key question you want to compare directly, have it translated properly and checked by a native speaker, then ask it the same way in every interview.
Screening needs the same care. Criteria should be equivalent rather than identical. An income band or a job title can mean quite different things in two markets, so define who you want to hear from in plain terms and let the local criteria follow from that.
Finally, plan the sample per market, not only in total. If one market has too few interviews to show a pattern of its own, you can't say much about how it differs from the others.
Writing questions that travel
Questions that work in one market can fall flat in another, usually because of phrasing rather than meaning.
Avoid idioms. "Are you on the fence about it?" or "Was it a no-brainer?" may translate literally and make no sense, or carry a different tone. Ask the plain version: "How sure are you?"
Avoid category jargon. Words like "omnichannel" or "consideration set" are industry language in any language. If a category term matters to the study, define it in plain words first, or better, ask about what people do.
Watch for concepts that don't map. "Value" or "premium" can mean different things from one market to the next. Asking about a real recent purchase gets you the person's own definition instead of the one in your brief.
Be careful with directness. In some cultures people are less willing to criticise openly, especially to someone who seems connected to the brand. Asking about concrete events ("What happened the last time you tried to cancel?") tends to surface problems that a direct "What don't you like?" would not.
Translation and analysis
There are two broad approaches. You can analyse each market in its original language and translate the findings, or translate the transcripts into one language and analyse them together. Either can work. What matters is that the original words stay available.
Keep original quotes alongside any translation, so someone who speaks the language can check that the meaning survived. Pay particular attention when participants in two markets use different words for what seems to be the same thing, or the same word for different things. Those are often findings in their own right, and a smooth translation can hide them.
When you compare markets, be strict about what counts as a difference. A theme that appears in one market and not another may reflect the market, or it may reflect who happened to be in the sample. Report it as a market difference only when the evidence supports it.
How Juno handles this
Juno interviews each participant in their own language. People take part by text or voice in their browser, with nothing to download. When scoping a study, confirm the languages you need, and the text and voice combination for each, with the Juno team.
A Juno study is built from research objectives rather than a fixed script. Each interview adapts to the person, but Juno works towards the study objectives and asks word-for-word questions exactly as written, so interviews stay comparable across the whole study. Before launch, the design is checked for leading wording and built-in assumptions, and you can rehearse it with a simulated interview to catch weak wording early. Simulated interviews are only used for rehearsal, never as research data.
During interviews, Juno uses neutral, plain-language questions and asks for concrete examples when an answer is vague.
You can bring your own participants through a study link, connect a panel provider you already use, or recruit through verified panel partners with 21M reach across 56 countries.
In the findings, every claim links back to the quote and the moment in the transcript it came from, so your team can inspect the source evidence. Confirm original-language transcript access and any translation requirements with Juno before launch. You can then debrief with Juno, ask follow-up questions about the data and challenge a claim. See our methodology standards.
Frequently asked questions
Should I translate my discussion guide word for word? Usually not. Translate the research objectives and keep word-for-word translation for the few questions you need to compare directly. Have those checked by a native speaker.
Can I compare findings across markets if the interviews were in different languages? Yes, as long as every interview worked to the same objectives and any key questions were asked the same way in each market. Keep the original quotes so differences can be checked against what people said.
How do I avoid questions that don't translate? Write in plain language without idioms or industry terms, and ask about real events rather than opinions in the abstract. A native speaker reading the questions before launch will catch most problems.
Which languages does Juno interview in? Juno interviews people in the participant's own language. Confirm current coverage, and whether you need text or voice, with the team when you scope the study.