Quick answer: The best market research platform for concept testing depends on how much of the work your team wants to do. DIY platforms suit teams with research skills who test often, assisted platforms add expert questionnaire review and analysis support, and full-service agencies suit complex, high-stakes programmes. Match the service model to your team capacity and the size of the decision.
Most lists of concept testing platforms compare features. Yet two teams can buy the same platform and get very different results, because one has a researcher who can write a balanced questionnaire and the other does not. The service model behind the platform often matters more than the feature list.
This guide compares the three main service models for concept testing and helps you choose based on your team, your timeline and the decision at stake.
What are the three service models?
| Model | Who does the work | Typical examples | Best for |
|---|---|---|---|
| DIY | Your team writes, launches and reads the test on a self-serve platform | SurveyMonkey, Qualtrics with a sample source, self-serve panel platforms | Teams with research skills running frequent, simple tests |
| Assisted | Your team leads, with expert review of the questionnaire and help reading results | Panel platforms with research specialists, such as Vase.ai or Milieu Insight | Brand teams that test often but want a safety net on design |
| Full-service | An agency designs, runs, analyses and presents the study | Ipsos, Kantar, NielsenIQ | Complex, multi-market or high-stakes concept programmes |
Our wider comparison of DIY and full-service market research covers the general trade-offs. Here we focus on what each model means specifically for concept tests.
When does DIY concept testing work well?
DIY works when someone on the team knows how to write neutral concept descriptions, choose between monadic and sequential monadic designs, and set up rotation. It is ideal for frequent, low-risk tests, such as screening early ideas or comparing headline wording. The main risk is a biased questionnaire nobody catches, so build a simple peer review step into your process.
When is an assisted model the better choice?
The assisted model suits most brand teams in Southeast Asia. Marketers own the brief and the concepts, while a research specialist checks the questionnaire for leading wording, logic errors and missing diagnostics before launch. After fieldwork, the team reads results on a dashboard and can ask for help interpreting segment differences.
This keeps tests fast and affordable while protecting quality on the parts that are easiest to get wrong. It is especially useful when concepts need translating into Bahasa Malaysia, Bahasa Indonesia, Thai or Chinese, as an expert can flag where the translated wording changes the meaning.
When should you go full-service?
- The concept programme spans several markets and needs harmonised design and analysis.
- You need advanced methods such as volumetric forecasting, conjoint or in-home product testing.
- The decision involves major capital, such as a new factory line or a national launch.
- Leadership expects a presented, agency-branded recommendation.
- Your team has no capacity to brief, review and interpret in-house.
How do you decide which model fits your team?
- Count your tests. If you will run many small concept tests a year, DIY or assisted models give you speed and repeatability.
- Audit your skills. Be honest about whether anyone can design a fair concept test. If not, choose assisted or full-service.
- Size the decision. The bigger and less reversible the decision, the more design support you need.
- Check the timeline. If you need results within days of approving the questionnaire, rule out models that cannot deliver that.
- Plan for mixing. Many teams use DIY or assisted tests to screen ideas, then a full-service study for the final go or no-go.
What does a mixed approach look like in practice?
Imagine a Malaysian beverage brand with twelve early product ideas. The team screens all twelve with quick assisted concept tests, reading appeal, uniqueness and purchase intent among category drinkers. Four ideas clearly stand out. The team refines those four, retests them with fuller diagnostics, and narrows to two.
For the final two, the stakes are higher because a winner means new production tooling. Leadership asks for a deeper study with volumetric estimates, so the team commissions a full-service agency for that final stage. Each model does the job it is best at, and the expensive study only runs on concepts that have already earned it.
The key to making this work is consistency. The team used the same core measures, the same audience definition and the same concept format across every stage, so results from the quick tests and the agency study could be compared directly. Without that, each stage would have started from scratch and the screening work would have been far harder to defend.
If you are comparing specific platforms rather than service models, our checklist on what to look for in concept testing platforms covers features, panels and outputs.
Where does Vase.ai fit?
We build Vase.ai so teams can run concept tests DIY on the platform or with our research experts reviewing the questionnaire and supporting analysis, drawing on a verified panel of Southeast Asian consumers with AI-validated responses.
Either way, results land in a real-time dashboard as fieldwork runs, so your team can watch concepts separate, cut by segment and move to the next round without waiting for a final deck. Full-service add-ons such as survey building and scripting or a full research report are there when a test needs more support.
Frequently asked questions
Is DIY concept testing reliable?
It can be, if the questionnaire is neutral, the design is balanced and the panel is verified. Most problems come from leading concept wording or missing rotation, so a second pair of eyes on the questionnaire makes a real difference.
What does an assisted research platform include?
Typically a self-serve survey platform plus access to research specialists who review your questionnaire, flag design issues and help you interpret results. You keep control of the brief and timing.
Can I switch models between studies?
Yes. Many teams screen ideas with DIY or assisted tests and use full-service research only for final, high-stakes decisions. Keeping measures consistent across stages makes results easier to compare.
