Sensory analysis is the evaluation of a food or beverage by sight, smell, taste, touch, and hearing. ISO 5492:2008, Sensory analysis — Vocabulary, defines the terms for that evaluation. The test to run depends on the question: whether two samples differ, how a sample is perceived, or whether the intended consumer likes it.

What sensory analysis is

The Institute of Food Technologists Sensory & Consumer Sciences Division describes its subject as the measurement, analysis, and interpretation of human beings' reactions to the characteristics of foods and related materials as perceived through the five senses. Ohio State's Sensory Evaluation Center quotes a 1975 formulation of that idea: a scientific method used to evoke, measure, analyze, and interpret responses to products as perceived through the senses of sight, smell, touch, taste, and hearing. The center lists product development, ingredient replacement, and shelf-life studies among the jobs the work supports.

ISO 5492:2008 applies to industries concerned with the evaluation of products by the sense organs. Its catalog groups the terms under general terminology, the senses, organoleptic attributes, and methods. ISO/TC 34/SC 12 states its scope as standardization in food sensory analysis, including vocabulary, general directives, selection and training of assessors, and the methodology of the tests.

ISO 6658:2017, Sensory analysis — Methodology — General guidance, gives general guidance on the use of sensory analysis. It describes tests for foods and other products, and some general information on statistical treatment of results. The catalog states that these tests are generally for objective sensory analysis. It describes a hedonic test as one that determines acceptability, or preference among two or more products, for a specified consumer population, and it points hedonic guidance to ISO 11136.

ASTM Committee E18 on Sensory Evaluation was formed in 1960. Its standards are published in the Annual Book of ASTM Standards, Volume 14.03. The committee scope covers standards and technical publications for the sensory evaluation of products, materials, information, and services, including terminology and the design, conduct, analysis, and interpretation of tests.

Discrimination tests

A discrimination test asks whether a perceptible difference, or a similarity, exists between samples. Liking is a separate question, covered with the hedonic scale below. Three methods have their own ISO standards: the triangle test, the duo-trio test, and the paired comparison. Each catalog names a change in ingredients, processing, packaging, handling, or storage as an example of when the method is used. Each also names selecting, training, and monitoring assessors.

Triangle test. ISO 4120:2021, Sensory analysis — Methodology — Triangle test, specifies a forced-choice procedure for whether a perceptible sensory difference or similarity exists between samples of two products. The catalog states that the method applies to a difference in one attribute or in several, including when the nature of the difference is unknown. It does not give the size or the direction of the difference, or name the attribute responsible. It applies only if the products are homogeneous. The catalog states that the method is statistically more efficient than the duo-trio test, and that it has limited use with products that show strong carryover or lingering flavors.

Duo-trio test. ISO 10399:2026, Sensory analysis — Methodology — Duo-trio test, specifies a forced-choice procedure for a perceptible difference or similarity between two products, including when the attribute behind the difference is unknown. The catalog states that the method is statistically less efficient than the triangle test and easier for assessors to perform, and that it applies only if the products are fairly homogeneous. It describes two forms. The constant-reference technique is for when one product is familiar to the assessors, for example a sample from regular production. The balanced-reference technique is for when one product is not more familiar than the other.

Paired comparison. ISO 5495:2005, Sensory analysis — Methodology — Paired comparison test, describes a procedure for a perceptible difference or a similarity in the intensity of a sensory attribute. The catalog states that the test is also called a directional difference test or a 2-AFC (alternative forced choice) test, and that it is a forced choice between two alternatives. It can state the direction of the difference on the attribute under study. It does not give the extent of that difference. A match on the attribute under study leaves every other attribute untested. The method applies only if the products are relatively homogeneous. Before the test, the supervisor has to know whether it is one-sided, with a direction expected, or two-sided, with the direction unknown.

Descriptive analysis and the Spectrum method

Descriptive analysis asks what a sample is like, attribute by attribute, and how strong each attribute is. ISO 13299:2016, Sensory analysis — Methodology — General guidance for establishing a sensory profile, gives guidelines for that process. A profile can cover a product evaluated by sight, odor, taste, touch, or hearing, including a food, a beverage, or packaging. The catalog lists uses that show up in brand work: developing or changing a product; defining a production or trading standard in sensory terms; defining a reference fresh product for shelf-life testing; studying and improving shelf life; comparing a product with a reference or with other products on the market or in development; relating perceived attributes to instrumental, chemical, or physical measures, or to consumer acceptability; and characterizing off-odors or off-tastes.

ISO/TC 34/SC 12's scope includes selection and training of assessors. That scope does not state a panel size or a number of training hours. Ask the partner what training covers for this product category.

The Spectrum method is one descriptive approach. It is also called the Sensory Spectrum method. The University of Arkansas System Division of Agriculture describes the Sensory Spectrum method as an objective method for describing the intensity of attributes using references for those attributes. In the hops study that page reports, assessors scored aroma on a 15-point scale anchored to reference foods. That scale belongs to that study. The Center for Dairy Research at the University of Wisconsin–Madison describes descriptive analysis as trained panelists quantifying appearance, flavor, and texture, and states that its panel uses a hybrid of QDA and Spectrum methodology. Ask which descriptive method, which references, and which attributes the ballot will use for this SKU.

Affective tests and the 9-point hedonic scale

An affective test asks about liking. ISO 11136:2014, Sensory analysis — Methodology — General guidance for conducting hedonic tests with consumers in a controlled area, describes approaches for measuring, in a controlled area, the degree to which consumers like or relatively like products. The catalog states that the tests collect responses to questions, on paper or on a screen. Behavioral measures, such as how much is consumed, are outside its scope. ISO 6658 directs that hedonic question at a specified consumer population.

The nine-point hedonic scale is a category scale for degree of liking. A history of food-acceptance research, from the University of Nebraska–Lincoln digital commons, states that Peryam and Pilgrim (1957), "Hedonic scale method of measuring food preferences," in Food Technology, is the paper most often cited for the scale, and that it is not the first paper on it. A Cal Poly thesis citing that paper describes a balanced scale with a neutral center and four phrases on each side: dislike extremely, dislike very much, dislike moderately, dislike slightly, neither like nor dislike, like slightly, like moderately, like very much, and like extremely. A Virginia Tech enology note says a hedonic scale may use an odd number of categories, from five to nine, from "dislike extremely" to "like extremely," with a neutral midpoint. Nine points is a common written form. It is not the only ballot that note describes.

A liking score records degree of liking. A discrimination result records whether samples could be told apart. A profile records attributes and intensities.

When a brand uses each test

Product development. ISO 13299 names developing a product, and comparing it with a reference or with similar products, among the uses of a sensory profile. That is the descriptive job: the appearance, aroma, flavor, texture, and other attributes the formula is supposed to hit, and how intense each one is. ISO 11136 is the liking job, with people from the specified consumer population, in a controlled area. A triangle, duo-trio, or paired comparison test answers a different question: whether two versions can be told apart.

Reformulation. The ISO 4120, ISO 10399, and ISO 5495 catalogs each name a change in ingredients, processing, packaging, handling, or storage as a case for asking whether a perceptible difference results, or whether the products are similar on the question the method asks. A paired comparison asks about one named attribute and the direction of that difference, not the size. A triangle or duo-trio test can be used when the attribute behind a change is unknown beforehand. A profile, under ISO 13299, is the catalog's tool for mapping which attributes changed and for relating them to consumer acceptability. A hedonic test asks whether liking changed for the specified population. An ingredient swap, a process change, and a cost-driven formula change are all still one of those three questions. Ask which question this project needs answered.

Shelf life. ISO 13299 names two shelf-life uses of a profile: defining a reference fresh product, and studying and improving shelf life. The discrimination-test catalogs name storage among the changes they can examine. A hedonic test can ask whether acceptability for the specified consumer population still holds at a later pull. None of these catalog pages states a number of months, a pull schedule, a panel size, or a failing score. Ask what the fresh reference is, which attributes or liking question will be tracked, and what result would end the commercial life of this SKU.

Questions to ask a co-packer or formulation partner

NXTCPG's published Source & produce work is ingredient sourcing, co-manufacturer selection, and quality and scale readiness. The published process builds the sensory target together with the ingredient system, the production path, and unit economics. The site does not state a sensory panel, a method, a panel size, a turnaround, or a price. Who manufactures, who packs, who owns the formula, and what a pilot must show are in how to choose a co-packer. Whether the plant's program meets current good manufacturing practice is a separate question, in what cGMP is.

  • What decision does this test support: a difference, a description of attributes, or liking by a specified consumer population?
  • Which written method will be followed? Name the document and the edition, for example ISO 4120, ISO 10399, ISO 5495, ISO 13299, ISO 11136, or an ASTM E18 standard.
  • Who are the assessors for this SKU: trained assessors, or consumers from a stated population? What does training cover for this product category?
  • For a discrimination test, is the question difference or similarity? For a paired comparison, which attribute, and is the test one-sided or two-sided?
  • For a descriptive test, is the method a Spectrum method, another descriptive method, or a hybrid? Which references anchor the scale, and which attributes are on the ballot?
  • For a hedonic test, what is the scale, who is in the consumer population, and is the test in a controlled area as ISO 11136 describes?
  • For shelf life, what is the fresh reference, when are later samples pulled, and what result ends the commercial life of this SKU?
  • Who prepares and codes the samples, who sees the data, and does the plant run the test or send it to another lab?
  • What does the agreement call a pass on sensory results for the pilot? That pass belongs with the commercial terms in the co-packer note.

Get the answers for this SKU. A result on another formula, another container, or another storage time is a result for that other case.

Where this sits with NXTCPG

Choosing the plant sits in Source & produce. If the open question is which co-packer or formulation partner can run the product, and which sensory question the project needs answered, start a project. For equipment, minimums, who buys the ingredients, and who owns the formula, use how to choose a co-packer. For the manufacturing-practice rule under that choice, use what cGMP is.

FAQ

What is sensory analysis?

Sensory analysis is the evaluation of a product by the senses. ISO 5492:2008, Sensory analysis — Vocabulary, defines terms relating to sensory analysis and applies to industries concerned with the evaluation of products by the sense organs. The Institute of Food Technologists Sensory & Consumer Sciences Division describes its subject as the measurement, analysis, and interpretation of human beings' reactions to the characteristics of foods and related materials as perceived through the five senses. Ohio State's Sensory Evaluation Center quotes a 1975 formulation: a scientific method used to evoke, measure, analyze, and interpret responses to products as perceived through sight, smell, touch, taste, and hearing.

How do a triangle test, a duo-trio test, and a paired comparison differ?

Each is a discrimination method with its own ISO standard, and each asks about a difference or a similarity rather than about liking. ISO 4120:2021 specifies the triangle test, a forced-choice procedure that applies even when the attribute behind the difference is unknown, and the catalog states that it is statistically more efficient than the duo-trio test, with limited use on products that have strong carryover or lingering flavors. ISO 10399:2026 specifies the duo-trio test. Its catalog states that the method is easier for assessors and describes a constant-reference form, for when one product is familiar, and a balanced-reference form, for when neither product is more familiar. ISO 5495:2005 specifies the paired comparison test for the intensity of a named sensory attribute, including the direction of that difference. It states the direction of that difference, not the size. A match on that attribute leaves every other attribute untested.

What is descriptive analysis, and what is the Spectrum method?

Descriptive analysis uses trained assessors to identify attributes and their intensities. ISO 13299:2016 gives guidelines for establishing a sensory profile. The Spectrum method, also called the Sensory Spectrum method, is one descriptive approach. The University of Arkansas System Division of Agriculture describes it as an objective method for describing attribute intensity using references for the attributes. The University of Wisconsin–Madison Center for Dairy Research states that its descriptive panel uses a hybrid of QDA and Spectrum methodology. Ask a partner which method, which references, and which scale it will use. A panel size or a training length from one study is not a rule for every product.

What is the 9-point hedonic scale?

It is a category scale for how much a person likes a product. A history available from the University of Nebraska–Lincoln digital commons states that Peryam and Pilgrim (1957), in Food Technology, is the paper most often cited for the nine-point hedonic scale, and that it is not the first paper on the scale. A Cal Poly thesis citing that paper describes a balanced scale with a neutral center and these phrases: dislike extremely, dislike very much, dislike moderately, dislike slightly, neither like nor dislike, like slightly, like moderately, like very much, and like extremely. A Virginia Tech enology note says a hedonic scale may use five to nine categories from "dislike extremely" to "like extremely." Nine points is a common form in that literature. ISO 11136:2014 is general guidance for measuring, in a controlled area, how much consumers like or relatively like products.

What should a brand ask a co-packer about sensory testing?

Ask whether the test is meant to show a difference, describe attributes, or measure liking in a specified consumer population. Ask which written method will be followed, such as an ISO standard named in this note or a method from ASTM Committee E18, and whether the assessors are trained or are consumers. Ask what result would count as a difference, a match, or a change in liking for this SKU, including on a shelf-life pull. Ask who prepares the samples and who sees the data. Current good manufacturing practice is a separate question from the sensory result. This answer is not a product-specific test plan.