Quick answer: Choose between monadic, sequential monadic and comparative designs based on whether you need pure reactions or a head-to-head winner. Randomise order, ask the same questions about every option, and measure recognition, appeal, clarity and purchase intent alongside reasons. Read the winner and the barriers together, because a concept can lead overall and still have a fixable weakness.
How you show the options matters as much as what you ask. Each design has its own bias, so match the design to the decision.
| Design | How it works | Best for | Watch-out |
|---|---|---|---|
| Monadic | Each respondent sees one option only | Clean, unbiased read of each option | Needs a larger sample, since each option needs its own group |
| Sequential monadic | Each respondent sees all options one at a time, in random order | Good balance of efficiency and fairness | Earlier options can colour later ones, so randomise |
| Comparative | Respondents see options side by side and choose | Finding a clear winner for packaging | Less depth on each option individually |
For packaging, a comparative step is often natural, because shoppers do compare packs on shelf or on screen. For promotion concepts, sequential monadic is a common choice, because each concept needs to be understood on its own before it is compared.
Preference alone is a thin result. The most useful studies capture several measures for every option, so you can see not only which one leads but why.
Ask the same questions, in the same order, about every option. Any difference in wording can be mistaken for a difference between the options.
Randomise the order in which options appear, and randomise the order of answer choices. Use images of the same size and quality, shown the same way, and check how they look on a mobile screen, since many respondents will see them there. If one design is a current pack and another is new, label them neutrally and avoid wording that signals which is which.
When comparing a current pack with a former or new one, add a recognition question as well as a preference question. People often prefer the familiar option, and recognition tells you whether that familiarity is an asset or a barrier to change.
Promotion concepts usually succeed or fail on friction. Respondents react to minimum spend thresholds, complicated steps, and savings claims that sound too good to be true. Simple mechanics that need no upfront cost tend to be easy to understand, so test them against more complex alternatives on the same scale.
Include a question on what would make the concept more appealing, and a free-text field on what worries respondents. These often point to a straightforward fix, such as lowering a threshold or adding a smaller guaranteed benefit, which can move a middling concept into a leading one.
It also helps to include a short warm-up before the first option, such as a question about how they usually choose in the category. This gives you a baseline for interpreting the results, and it shows whether a concept appeals to heavy buyers, light buyers or both. Without it, a winner among enthusiasts can look like a winner among everyone.
Finally, plan for a follow-up round. A first test rarely produces a finished design. It tells you which option is strongest and what to fix. A smaller second round on the refined version confirms that the fix worked, and it is often quicker than the first because the questionnaire is already built.
Showing options in a fixed order is the most frequent error, because the first option often benefits from being seen when respondents are most attentive. Others include using different image quality for different options, asking leading questions such as which design looks more premium, and drawing conclusions from segments that are too small to read.
A further mistake is testing too many differences at once. If two packs differ in colour, shape, claim and logo, you cannot tell which change drove the result. Where possible, change one thing at a time, or accept that you are comparing complete designs and read the results that way.
Look at the overall winner, then check the diagnostics behind it. A concept that leads on preference but scores low on credibility has a messaging problem. One that scores high on understanding but low on intent may be easy to grasp yet not compelling.
Cut the results by the segments that matter to the business, such as region, age group, language or ethnic group, but only where the sample is large enough to support it. Our guide to product concept testing in Malaysia and the note on testing ad creative before a campaign show how the same logic applies to other assets.
For sample size, a common norm is around one hundred respondents per cell you want to compare, with more for monadic designs. The survey sample size guide explains this in more detail.
We build Vase.ai so that concept and packaging tests move quickly from idea to read-out. You can run the study yourself, with randomisation, logic and image display built in, or with research experts supporting the design, and watch results arrive in a real-time dashboard.
That speed means you can test a refined version after the first round instead of debating it, which is often where the biggest gains come from.
Can I test packaging and promotions in the same survey?
Yes, but keep them in separate blocks and randomise block order. Respondent fatigue builds quickly, so keep each block short and focused.
How many designs can I compare at once?
There is no fixed limit, but three or four is a comfortable maximum for most surveys. More options increase fatigue and dilute the sample behind each one.
Should I show real photographs or mock-ups?
Use the highest-quality visuals you have, treated the same way for every option. Rough mock-ups next to polished designs bias the result towards the polished ones.