Sensory Market Research: Methods and Comparisons
Sensory market research tests how people experience physical products. Learn how to choose a method and distinguish similar-enough, preferred and approved.
Sensory market research studies what people notice and like when they taste, smell, touch or otherwise experience a physical product. It can help evaluate a reformulation or choose what to develop. This guide explains how to choose a test, define acceptable change and keep similarity, liking and product approval apart.
Example: replacing a peach drink’s flavor
Tern Orchard is a hypothetical producer of a clear, still peach drink. It needs to replace a flavor ingredient while keeping the drink its customers know.
For an ingredient replacement, agree how much sensory change is acceptable and which error risks the test should control before tasting. A preference ranking or inconclusive difference test answers a different question.
Tern’s current flavor supplier will stop supplying its blend. R&D has a replacement candidate for the single-serve bottled drink and wants to preserve its pale-gold appearance, peach aroma, sweetness balance and mouthfeel.
The technical brief describes a homogeneous finished drink: flavor is dispersed and there are no fruit pieces. The reference uses the current blend; the candidate uses the proposed replacement. R&D specifies the other ingredients, processing and packaging conditions to be held comparable. The sensory specialist must still check whether the supplied materials and serving conditions suit the proposed method.
Urgent supply needs do not define acceptable change. Agree what “keep the product familiar” means, which sensory qualities matter and how large a perceptible difference would be too much.
For Tern, approval needs evidence that the candidate is within the agreed meaningful-difference boundary at the agreed error risks. Approve that rule before the provider chooses the sample size and runs the test.
Approval covers the supplied products and tested conditions. To cover other batches or later shelf life, include suitable samples and a design for them. One fresh lab sample cannot cover all future production.
A nonsignificant difference is not proof of similarity
The lab proposes: “If the difference test is not statistically significant, we will recommend releasing the replacement flavour.” Tern declines that release basis and asks for a design intended to show sufficient similarity against the agreed criterion. The study changes before fieldwork rather than receiving a qualification at the final presentation.
ASTM E1885-25's public Significance and Use section distinguishes the two objectives. A difference test begins by assuming no perceptible difference and asks whether that assumption can be rejected. A similarity test begins by assuming a meaningful difference and asks whether the evidence supports rejecting that condition. The hypothesis and chosen error and perceptibility parameters must fit the objective.
Ask how the design limits both risks: approving too much change and withholding approval from a similar-enough candidate. Agree those risks and the meaningful-difference boundary with the owners. The number of assessors and analysis must support the rule, not just the provider’s usual panel size.
Ask who will taste the samples, how their relevant sensory ability is checked and what familiarization they need. Limit conclusions to those assessors and that task. A consumer liking panel is not the same as a panel chosen for this sensory comparison.
Accept a proposal that explains the design and its limits. Hold the preservation decision if the provider offers only a difference screen or preference ranking. The substitution still needs the evidence Tern specified.
Use samples that represent the proposed change
Send finished reference and candidate products with formula and batch identities. Agree production, storage, age and preparation. An old reference and fresh candidate may differ in more than the replacement flavor. Test the change you actually intend to approve.
The public abstract for ISO 4120:2021 describes a forced-choice triangle method for perceptible difference or similarity with homogeneous products. It flags limits from strong carryover or lingering flavors and says the comparison does not identify the attributes responsible for a difference. That is a method to discuss with a sensory specialist, not a complete testing protocol supplied by the abstract.
Ask the specialist to justify the method for these drinks, including consistency, fatigue, adaptation and carryover. Change an unsuitable design before testing. Serving order alone is not enough, and a waiting interval from another product is not a universal rule.
Code samples to hide supplier and formula identities while keeping appearance visible under agreed conditions. Pale-gold color is part of Tern’s goal. Hiding it with lighting or opaque containers would remove something the test is meant to assess.
If color is hidden to answer a narrower question, say so and study appearance separately. A triangle test also cannot identify which aroma or texture caused a difference; use suitable further work to find that.
Approve only the substitution and conditions tested
Before testing, agree how the report will connect products, assessors, conditions and results to the decision rule. It should describe what was completed, important departures, uncertainty and whether the similarity criterion is supported.
These are possible decisions, not measured results:
Return | Next action |
|---|---|
The agreed similarity criterion is supported. | Consider sensory approval for this substitution and its tested conditions. Keep other release requirements separate. |
Sufficient similarity is not demonstrated. | Hold preservation approval. Resolve whether the issue is the candidate, unsuitable conditions or inconclusive evidence. |
A perceptible change is established. | Rework the candidate, or explicitly change the business question to acceptance of a changed product. |
R&D owns sensory approval alongside the technical release process. Tell supply and manufacturing which ingredient and conditions it covers. Safety, stability, specification and manufacturing approval remain separate. A sensory result cannot supply them.
Tern is buying a test for a defined approval rule, not just a tasting. If it cannot get that evidence in time, manage the supply problem without calling an easier result approval.
A changed drink needs a consumer-acceptance question
A detectable change may still be liked. If Tern is willing to sell a revised drink, test acceptance with actual consumers and products in a suitable drinking setting. That is different from showing an ingredient is interchangeable.
Our food market research guide connects the occasion, promise and actual eating or drinking experience. Use that research to check expectations and choose which product to develop.
A difference alone does not show rejection or justify launch. Liking does not satisfy the earlier preservation rule either. Record the new objective and keep the other evidence the product decision needs.
Familiar replacement or a new variant?
Physical testing may leave questions about how to explain a revised drink, compare offers or choose an audience. Build those alternatives from the product facts and sensory evidence.
At Sapien, we build simulated audiences using purchase, behavioral and research data. Customer evidence and earlier studies can add context. We compare responses, explain group differences and ask the same simulated audience follow-up questions.
Use a simulated comparison to develop the offer and customer questions. It is not tasting, proof of sensory similarity or manufacturing approval. Our concept testing guide connects offer and audience choices with what to do next.
Book a demo to compare truthful descriptions of the revised drink and explore which audience would consider it, using the physical findings your team can provide.
