Concept testing when the market is one you cannot staff
Concept testing is straightforward when your market is down the corridor. It stops being straightforward when the concept is going to Spain, Mexico or Chile, you have no local research partner, and the budget for a market you are only exploring is zero. This page covers what to ask, in what order, and how six synthetic profiles grounded in one of those markets close that loop in two minutes.
Three questions, not one
A concept test is not measuring whether people like the idea. It measures whether it is understood, whether it solves something for someone, and whether that someone would drop what they buy today to buy it.
Comprehension is the cheapest to test and the most often skipped: if the participant cannot repeat what it is, there is no concept yet, there is a sentence to rewrite. Substitution — what they would stop buying — is the only answer that resembles a business signal, and it is the one almost nobody asks.
The order that protects the answer
Each step contaminates the next, so the sequence is the method, not a preference.
- Show the concept in two sentences, in plain speech, with no benefit list.
- Ask what they understood, in their own words, before any rating.
- Ask what problem they think it solves and who they think it is for.
- Ask what they would stop using if this existed.
- Only then, price — and in a separate round, never in the same sentence.
- Close with what would have to be true for them to try it.
Why the market matters more than the translation
A concept tested against a generic Spanish-speaking persona comes back plausible and useless. The profiles here are built on one market at a time: income bands, retail names people actually reach for, the institutions they cite in an argument about prices, and the current economic and social climate of that country injected before the session.
That is what makes a price read as cheap in one market and offensive in another with the same figure in euros, and it is the part a translated persona cannot fake.
What you get back
A full transcript plus an executive report — headline finding, reading by segment, drivers and barriers each tied to the quote that supports it, prioritised recommendations — exportable to PDF, PowerPoint and CSV. 10 questions cost €7.99, which is 60 answers.
The report is written in the language of your panel, so an English-speaking team gets English analysis while the quotes stay in the participant's own words.
Where this stops
If you need a defensible number — what share of the category would buy, with a margin of error — this does not produce one, and no synthetic panel does. That is a survey on a sample designed to project.
Use this to arrive at that survey with three concepts instead of ten, and with the wording already tightened by six people who tried to misunderstand it first.
Three ways to test a concept in a market you do not staff
| Local fieldwork agency | Online panel | Synthetic room | |
|---|---|---|---|
| Time to first answer | 3–6 weeks | 24–72 hours | Under 2 minutes |
| Cost per round | Thousands | Hundreds | €7.99 |
| Follow-up questions | Yes | No | Yes |
| Local language and references | Yes | Yes | Yes |
| Projectable to a population | No | Yes, with the right sample | No |
Frequently asked questions
You write in whichever language you like and the analysis comes back in the language of your panel. The participants answer in their own register, because that register is what you are testing — a concept that only works when translated into corporate Spanish does not work.
Keep reading
Last reviewed: 2026-09-14. Participants are AI-generated synthetic archetypes, not real people.