November 1, 1999

18 Min Read
Sensory Evaluation



Sensory Evaluation
in Quality Control
November 1999 -- QA/QC

By: Bruce M. Floyd
Contributing Editor

  Each one of us, whether making and selling food, or simply consuming it, has expectations of how a particular food should taste. On a hot afternoon, we can just imagine the taste of our favorite beverage, for example. The last thing we want is for the flavor of our favorite product to fall below expectations.

  Recently, there were some well-publicized incidents in Europe involving The Coca-Cola Company of Atlanta. One issue involved a fungicide that was sprayed onto empty-can pallets. The fungicide was transferred to the outside of empty cans while in transit to a plant in Dunkirk, France, causing the cola cans to have a strange odor. Another problem was that contaminated gas reportedly "created a foul smell in some products" made in Antwerp, Belgium. The problem in Belgium caused some illness, according to the July 13, 1999 Wall Street Journal. In this case, since the odor was obvious to consumers, it would seem that the quality problem would have been easy to detect by the sensory-evaluation procedures that are a part of many regular QC programs. Often, such procedures can minimize the number of disappointed customers. So, what's involved in the processes, and what considerations arise when setting up an in-plant program?

Preliminary considerations

  In the past, some companies conducted "cuttings" of the previous day's production. Cutting was the term used in the "old days" to denote sensory and visual evaluation of a product. It didn't matter what the analytical results were, if the product did not make it through the cutting, it did not ship. Today, there are even better ways to make sure that off-flavored product doesn't get made in the first place. Basically, sensory evaluation should begin with raw and intermediate materials. It's important to do some type of sensory testing as part of a raw-material acceptance and in-process QC program. However, be aware of the limitations of such a program, and don't become overconfident in a program that is, for example, designed solely to pick up gross flavor defects in major raw materials. Remember, although it can be difficult to design and implement a sensory program at the plant level, that's no excuse not to do it.

  Even many large companies do not have an ongoing sensory program at the plant level. As Harry Lawless, Ph.D., professor of food science, Cornell University, Ithaca, NY, puts it: "Sensory testing during production is the corporate equivalent of diet and exercise. Everyone thinks it's a good idea, but few are willing to invest the time and energy to make it happen." Often, a significant amount of time and money is spent on sensory evaluation during product development, but then that product is turned over to a production plant where it is never tasted again. This is equivalent to an automobile company shipping a car that has never been test-driven.

  Lawless believes that one factor in the scarcity of sensory programs in QC is the personal comfort level with instrumental analysis - few undergraduate programs stress the use of sensory evaluation in QC. Others point to the resistance to change and new methods often present in production management. Kathleen Rutledge, president, 21st Sensory, Bartlesville, OK, observes that "management support is the most important consideration. When there is a lack of management commitment, the programs just do not happen." And, as Mona Wolf, president of WolfSensory, Mason, OH, points out: "If we ignore sensory in favor of all analytical testing, we are not using all of the tools that are available to us."

  If a company has an ongoing sensory-evaluation process built into its QC program, a number of product problems can be detected quickly. These include problems due to rancidity, sanitizer interactions, off-flavors from improper processing, missing ingredients, too much of an ingredient, or improper storage. Sour milk is easily detected by taste; however, taste is the last thing one should use to determine biological contamination. Many pathogens and their toxins are tasteless. If at all possible, wait until microbiological analysis has been completed on all "at risk" raw materials before tasting them. The old adage "when in doubt, throw it out" is a good rule to go by. If microbial testing is not possible, smell, but do not taste, the ingredient.

Starting the program

  Often, the problem with starting a program is knowing where to begin. If a company includes tasting as part of its QC procedure, but has never really given the program much consideration, it's almost at the same level as those companies that don't have a program. "A little sensory testing is worse than nothing, since one makes the false assumption that they have a sensory program," Rutledge warns.

  Many companies cannot afford the instrumentation required to run sophisticated instrumental analyses such as an electronic nose or gas chromatograph (GC). There is also some doubt as to the efficacy of this new technology at the plant level. The nose of a trained sensory person can be very sensitive; however, it's necessary to screen potential participants to determine their ability to differentiate between odors. This can be done by presenting a series of different, known odors packaged in small glass vials, and determining who can distinguish between them. Use odors familiar to the personnel at the plant. If there's a critical odor that everyone must be able to distinguish, it should be included. A simple form on which the participants write their findings and their name will suffice. Don't use exotic odors; this is a test to determine who will be screening all of the ingredients and products used and made at your plant. All those who will be evaluating flavors must pass the test, including supervisors and QC personnel.

  When selecting panelists, not only do you want people who can distinguish between different flavors and odors, but those who are trainable as well. "It takes a long time to establish parameters for a product. Once they are established, the QC panel has to be trained to measure them," says Rutledge. "A lot of resources have been invested in this process. It's not important that the panelist likes the product. What is important is that the panelist can determine if the product is within the established parameters." Also, the program must be set up so that no one person can overrule the panel. This is where careful screening of panel members, definition of product standards and the support of upper management come into focus.

  Once the panelists are identified, the next question is the location of the evaluation. If a plant makes highly flavored products, the entire facility may smell like the finished product. This makes it difficult for personnel who have worked at the plant for some period of time to differentiate odors. Consider bringing in someone to evaluate the plant and find an area suitable for sensory testing. For facilities that store and use spices, one can smell the spices even in the parking lot, making it difficult to find a suitable area. It may be necessary to take the samples to an outside location for evaluation, or to build a special room with a filtered, deodorized air supply. In extreme situations, everyone may need to change clothes and bathe before performing an evaluation.

  When constantly exposed to odors, the human nose becomes desensitized to them. For example, vanilla is a fairly delicate odor. If a lab smells like vanilla, the flavor intensity of the product in question may be judged to be below specifications. In an effort to correct the "problem," additional vanilla flavor might be added. Processing can cause flavor loss. This will intensify the vanilla odor in the plant air, potentially leading to yet another adjustment.

  The real problem in this scenario is that the finished product may have a flavor intensity that sets the customer's teeth on edge. This illustrates a number of points that must be kept in mind. If the plant's background odor is the same as the product, sensory evaluation will have to be made somewhere else. Also, be aware of the limitations of judging flavor levels at the plant, and don't forget that strong chemical odors can cover up the very flavors sensory panelists are trying to taste. If a smelly chemical procedure is performed every afternoon at 2:00 p.m., finish flavor evaluation first. Sensory evaluation must be given the same status as chemical or microbiological procedures - a well-designed sensory program will yield quality results.

Basic training

  Now that the panel and a location have been found, it's time to train the panel. Michael O'Mahony, Ph.D., professor of food science at the University of California-Davis, states that "just like a gas chromatograph is no good to someone who does not know how to use it, the mouth, as a sensory instrument, is no good if the person is not taught how to use it. They must first train the instrument between their ears." He suggests that anyone interested in sensory testing take a course to get a basic level of understanding before beginning an in-plant program. The University of California-Davis, for one, offers a short course through its extension program that covers statistics and the fundamentals of sensory testing.

  Quite often, the assistance of a consultant knowledgeable in the company's product and in sensory evaluation techniques will be necessary. It may also be a good idea to join the Chicago-based Institute of Food Technologists sensory division, or participate in the activities of the American Society for Testing and Materials (ASTM), which is based in West Conshohocken, PA.

  Many books are available on sensory evaluation, most of which are directed toward research. Two books in particular, however, are specifically targeted to plant QC. The first, The Role of Sensory Analysis in Quality Control, edited by June E. Yantis, is part of the ASTM Manual Series MNL14, and provides a basic guide to in-plant sensory testing. The second, another industry standard, is Sensory Evaluation in Quality Control, by Alejandra M. Muñoz, Gail Vance Civille and B. Thomas Carr. Unfortunately, this book is out of print, but should be available in the food science library of a university near you.

  What other resources are available? For ingredient suppliers, customers with sensory programs may be able to assist. Usually they're happy to help, especially if it leads to better-quality ingredients. Another invaluable resource is the information from the plant floor. The plant personnel may not be able to distinguish subtle flavor differences, but if an employee points out that the tank smells funny, take it seriously. No matter how good the sampling program, it's still a sampling program, and does not check every container of raw material or every item off the end of the line. In non-automated plants, the employees do handle every container, and if properly motivated, will point out when something is different. Even in automated plants, if a flavor problem develops, employees can notify their supervisor if the room odor changes.

Getting the lingo

  One challenge to sensory testing, according to Wolf, is a lack of common vocabulary. "Vocabulary is key to good quality control in the plant. People need training in the vocabulary that is relevant to their products - good and bad," she says. "This is especially true if there are several plants or the corporate offices are located in a different place than the plant."

  A report stating that a product tastes "bad" doesn't help. If employees are trained in a descriptive vocabulary, and are shown examples of what that vocabulary means, the information conveyed will be more significant. The dairy industry has made flavor judging a part of its dairy science program for many decades. This competition has fostered a common vocabulary of descriptive terms that describe almost all attributes of dairy products, allowing two people from different companies to describe products with a reasonably clear understanding of the terms.

  To illustrate the difficulties that can arise when there's no common vocabulary, just consider the world "stale" - what does that mean, specifically? If a given industry or industry segment doesn't have a common vocabulary, they need to create one, and teach it to everyone involved in production and quality control. This includes those at headquarters who assess product quality. It's also good to include the marketing department, so that they're able to relate consumer feedback to other departments.

Carrying out the plan

  Once there's a common vocabulary, check each employee report of a quality discrepancy carefully, and give the findings to the employee as well as to management. If employees don't see results of an investigation, they will assume that no one cares, and may quit reporting line discrepancies to management.

  Never miss an opportunity to use the day-to-day experience of the line employees. Consider the following scenario. An employee staging ingredients notices a musty odor, and that some of the bags are discolored. Upon investigation, it turns out there was a severe thunderstorm at the product storage site, and that the warehouse roof leaked. The warehouse personnel thought they had found and segregated all the wet product, but obviously some got by. If the manufacturing plant had a "stop the line when something is wrong" policy, the staging employee could prevent the manufacture of bad product. Even if the ingredients were used, this information would allow product segregation prior to shipment, and further evaluation. At a minimum, the reporting employee should receive recognition for reporting this problem.

  In quality control, the object is to prevent the manufacture of bad product, not catch it before shipping. (But that's the second most important thing.) The more employees that are watching for problems and trying to prevent them, the better the quality. Never forget that quality is built into a product, not added on by inspection. All end-product testing can do is prevent shipping off-flavored product after generating all the costs. It's difficult to reject finished product that is only a "little" out of spec. Higher management sometimes forgets that the customer does not have the same profit incentives as they do.

  How do you know what's normal? This can be determined several ways. The best method is for the product developers to select the raw-material standards. These could include normal and acceptable quality ranges. But what is quality anyway? It's producing the same product day in and day out - the texture, color, flavor and mouthfeel should never significantly vary. Some might criticize McDonald's for being "mediocre," but their product is the same everywhere. Most importantly, the customer knows what to expect, and when they go to a McDonald's anywhere, they get it.

  All standards should reinforce the established quality level for a given product. It does not serve the customer to make the best product in the world six out of seven days each week. Management must determine the acceptable level of quality that the company can deliver every day, and define it with standards that can be taught to everyone who will be responsible for making the product.

Setting Up a Sensory Program

  • Sensory evaluation takes a commitment on the part of corporate and plant management. The resources must be committed in order for the program to have a chance to work.

  • A quality standard must be developed for each product and ingredient. These standards must be stable and be accepted by management. Variation from a standard, whether above or below, is not acceptable.

  • A standard vocabulary must be developed that describes all of the attributes - good and bad - of each product and ingredient. This vocabulary must be taught to everyone involved in making or approving the product.

  • Find a suitable area for sensory testing at or near the plant. Make whatever building modifications are necessary before starting the program.

  • Employee training in the proper technique for tasting is necessary, and management must have training in the proper application of sensory testing and interpretation of the results. This includes some method of quantification of the results.

  • There must be follow-up to keep the program going. This could include outside audits of the program and an annual review of all products and standards. This is the time to evaluate whether or not the market has changed. It is also a time to renew corporate management's commitment to the program.

Test time

  What should sensory programs designed for quality assurance encompass? Everything! This includes everything that enters the food. Start with the raw materials, and don't overlook the water. If the product is made at several different plants, is the water the same at each? Does the water change taste with the seasons? Systems that obtain their water from rivers are subject to seasonal fluctuations in flavor due to run-off and algae in the water, and ground water is subject to contamination over time. At one time, one section of Los Angeles had water with an ammonia flavor due to the area's concentration of dairy cattle.

  In an accessible location, list all the ingredients that the plant uses. Add to this list water, steam, compressed air and any gases that are used or injected into the product. Also, if equipment is sanitized prior to production, there must be some method of determining if the sanitizer has been rinsed from the equipment. Next, determine which items will require special preparation before tasting or smelling. These would include chemicals, vitamins, gums, steam, and other minor or difficult-to-taste ingredients. Write preparation procedures for all of these materials. Do not leave this to chance - sample preparation should not become a variable.

  To test hazardous ingredients, it is necessary to dilute them to use levels, or, in some cases, only to smell them. Also, many times distinct off-flavors are associated with chemicals and vitamins that will ruin the flavor of finished products even at very low levels. Techniques specific to your products must be developed. When tasting an ingredient using a method that is significantly different than that of a supplier, it's best to give them a copy of the procedure, along with the reasons for the change. They need to buy into this test method, or it may be difficult to get them to make product to specification, or to take it back when it does not meet specification.

  How should reference standards be stored so that they do not become non-standard with age? If possible, store the standards in a sealed glass container in a suitable environment such as a refrigerator. Avoid exposure to light. It may be necessary, according to Edgar Chambers, Ph.D., professor of sensory analysis, Kansas State University, Manhattan, to develop specific standards for the attribute being considered - browning, for example. Specific standards make it easier to prepare fresh standards, and also to train panelists to look for specific characteristics associated with the product. It is also possible to use a different product that's very consistent as the basis for a flavor characteristic such as brown-sweet, while appearance of the product can be addressed with product photographs.

  Once the standards have been developed, it's time to train the panelists in the proper tasting and in the acceptable ranges. This will require a lot of time, and is the heart of the program. Again, experts are invaluable at this stage of the process. Training must be specific to the ingredients and products being evaluated. To list particular methods here would be misleading, since these must reflect particular ingredients, products and the program's level of sophistication.

  Chambers states that "it's not whether or not the food is good or bad, it is simply an evaluation of the attribute within the established range. This is no different than chemical or microbiological analysis." It is very important that everyone buy into this for the program to work. According to Wolf, "after the training, the plant people have to take ownership of the program. The plant manager has to want the program and buy into it also or it will not get off of the ground."

  Lawless sums up a realistic view of sensory testing at the plant level: "It is like a garden that takes constant attention, or it will wither and die. Panelists change jobs, quit or lose interest, management loses the initial enthusiasm that existed at the start, and new problems divert resources away from sensory testing."

  Another issue to consider is technique. This is not an easy area. Many books have been written about converting subjective data to numbers, and an equal number of books exist on statistical methods for comparing this data once it is converted to numerical values. There are also issues of first taste, saturation and significance. If it comes down to seasonal variation or marginal differences, product developers or a professional sensory lab need to evaluate the sample.

  Finally, flavor problems are sometimes not easily detected before, or during, production. Such problems might be due to slow oxidation or other chemical changes caused by storage, shipment, processing or ingredient interactions that are initiated before or during processing. It's a wise policy to have a regular review of finished product over its shelf life. The time-frame for the review depends on the shelf life, but it's advisable to do it no less than at quartiles of the shelf life.

  Sensory evaluation as part of the quality-control process has many facets, but if implemented correctly, can be an invaluable part of a company's overall agenda for producing quality food products.

Back to top


Subscribe and receive the latest insights on the health and nutrition industry.
Join 37,000+ members. Yes, it's completely free.

You May Also Like