How do you test a concept before you put it in front of real people?
A worked example from a real simulation run on one proposition, with the full sample report to explore.
Test one concept at a time with a simulated persona panel that answers the same questions on clarity, relevance, credibility and objections. The report shows which claims confuse people and which objections to answer before fieldwork. It helps you improve the concept first; it is not a score that says the concept will sell.
A concept is an idea you can describe to a stranger in a few sentences: what it is, who it is for and why it matters. Concept testing checks whether that description works before a team spends money on prototypes, campaigns or a large survey.
When should you test a concept?
Test when the idea is specific enough to describe and a costly choice is coming: which proposition to launch, which claim to put on a page, which product to build. Early is cheapest, when the answer can still be “not yet” and the changes are small. If you do not yet know who the audience is, start with audience discovery.
What does a sample concept test report look like?
The example below is a real run on one proposition: an automation partner that offers one company, one contract and one number to call, with a free on-site assessment as the first step. Four simulated personas from operations and executive roles each answered the same questions twice.
One-contract automation partner: does the concept land?
Check whether an end-to-end automation partner (one company, one contract, one number to call) is clear, appealing, credible and valuable enough to earn a free on-site assessment.
Proceed to a controlled pilot with explicit success measures and risk controls.
- Personas4
- Runs each2
- Simulations8
- Study typeConcept test
This is easy to understand and relevant because coordinating multiple automation vendors can consume a lot of operational time. I would take the free assessment, but I would need proof of capability, transparent costs, and clarity about ongoing support before committing.
Add a clear sample assessment showing the recommended solution, expected ROI, implementation timeline, pricing, and ongoing support commitments.
A simulated example for a fictional company, run on synthetic personas from a stated catalogue. It estimates how a defined audience might respond to the material shown; it does not measure real people, and company names are invented.
The overall signal is strong, 67.2 out of 100, and clarity leads at 76.2: all eight runs rated the concept 4 out of 5 for clarity. The lower scores are the useful ones. Credibility and differentiation were each rated 3 out of 5 in all eight runs, and six of the eight said they would “probably” choose it, conditional on proof. The concerns were consistent: unclear scope, no indication of cost and the risk of relying on one provider. One persona rated need relevance 2 out of 5 because the persona was retired, which is a reminder to check who is on the panel before reading an average.
How do you read the report?
Every persona-research report follows the same order. These six steps walk through the example above.
Start with the decision brief
The ring is the overall signal: 67.2 out of 100, labelled “Strong signal”. It combines the dimension scores with equal weight, so read it as a direction, not a verdict. Beside it sits the line the evidence supports acting on.
In the report: Decision briefFind the dimension pulling the score
Concept clarity is the strongest dimension at 76.2; Purchase intent is the weakest at 60.4. An average can hide one weak dimension, and that is usually where the next round of work belongs.
In the report: Performance across dimensionsCompare the personas
The same questions go to every persona, so you can see whether a finding holds across roles and company sizes or belongs to one group.
In the report: Overall and by personaRead findings with their prevalence
Each finding states how many of the 8 simulations it appeared in, next to a quoted simulated answer. A point raised in 8 of 8 runs carries more weight than one raised once.
In the report: Reported findings, prevalence and simulated voiceCheck who was simulated and why
The panel is 4 personas drawn from a catalogue with stated filters, each answering 2 times. The report lists every profile and why it was selected, so you can judge whether the right people were in the room.
In the report: Who was simulated—and whyFinish with the limits and the next step
Every report ends with what the evidence cannot show and the real-world checks to run next. The simulated answers are hypotheses to test with real people, not findings about them.
In the report: Limitations and appropriate use · Where to go next
Can you explore the whole report yourself?
Yes. The frame below holds the complete sample report, including the answer for every persona and run.
What would you do after a result like this?
The recommendations tell you what to fix in the concept before fieldwork:
- define exactly which technologies, project stages and support services are included
- give indicative pricing and explain the contract structure
- add documented return on investment and case studies from comparable customers
- clarify ongoing support terms, accountability and escalation
Change the concept, then test the version you plan to keep. Move the survivors to interviews for depth and a survey for scale.
What can a simulated concept test not tell you?
It cannot forecast adoption or set a price people will pay. Language models acting as simulated participants have reproduced the broad direction of published experiments, but only qualitatively (Horton et al.), and AI-generated “users” tend to approve most concepts and give shallow, over-optimistic answers (Nielsen Norman Group). Treat a strong signal as a reason to proceed to real research, not as a reason to skip it.
A simple sequence
- Write the concept in three sentences: the problem, the solution and the audience.
- Fix the questions and the criteria before you run it.
- Read clarity and objections first, then relevance and credibility.
- Change the concept and test the version you plan to keep.
- Take the survivors to real people.
Frequently asked questions
How is a concept test different from a message test?
A concept test looks at one proposition as a whole: what it is, who it is for and why it matters. A message test compares wording or framing. Test the concept first, then compare messages for the version you plan to keep.
Should I show several concepts at once?
For a pure concept check, one at a time avoids comparison effects. If you need to choose between concepts, use a preference comparison with the same questions for each option.
What should I do with a concept that tests badly?
Look at the specific objections and clarity problems first. Rewrite the claim that confuses people and test again. A weak result on one wording is not proof that the idea is wrong.
Sources
Sources for the facts on this page, last checked October 17, 2026.
- Horton, Filippas and Manning, Large Language Models as Simulated Economic Agents: What Can We Learn from Homo Silicus? (arXiv 2301.07543) checked October 17, 2026
- Nielsen Norman Group, Synthetic Users: If, When, and How to Use AI-Generated “Research” (2024) checked October 17, 2026
Related guides
- What questions should you ask in a concept test?A concept test asks the same questions about every concept: appeal, relevance, uniqueness, believability, purchase intent, price and why. A ready-to-use set.
Run research end to end. Keep the knowledge working.
One AI-native operating system for market research and insight professionals. Study design, evidence generation, agents, institutional knowledge, delivery and action, all in one place.