Measuring palatability of pet food products: Sensory components, evaluations, challenges, and opportunities

Abstract

The pet food industry is a growing business launching a variety of new products in the market. The acceptability or preference of pet food samples has traditionally been measured using either one‐bowl or two‐bowl tests. Academic researchers and professionals in the pet food industry have explored other methods, including the cognitive palatability assessment protocols and the ranking test, to evaluate more than two samples. A variety of approaches and perspectives were also utilized to predict palatability and key sensory attributes of pet foods, including descriptive sensory analysis by human‐trained panelists and pet food caregivers’ perceptions of pet food. This review article examined a range of testing methods for evaluating the palatability of pet foods, specifically targeting products for dogs and/or cats. It outlined the advantages and disadvantages of each method. Additionally, the review provided in‐depth insights into the key sensory attributes of pet foods and the methodologies for assessing palatability. It also explored pets’ behavioral responses and facial expressions in relation to different pet foods. Furthermore, this review discussed current challenges and future opportunities in pet food development, including the use of instrumental analyses and artificial intelligence–based approaches.

1. INTRODUCTION

Pets are a staple part of family life for many people in the United States. It has been reported that 66% of the US households (equivalent to 86.9 million) own a pet (American Pet Product Association, ^{2023a, 2023b}). Dogs and cats are the most prevalent pets in the United States, with 65.1 million households owning dogs and 46.5 million households owning cats (American Pet Product Association, ^{2023a, 2023b}). During the COVID‐19 pandemic lockdown, there was a significant increase in pet ownership globally. According to Health for Animals Organization (2022), it is estimated that over half of the global population now owns a pet. In addition, the pet food market has experienced growth due to the trend of pet humanization (Verdon, 2022). This trend reflects pet caregivers’ increasing demand for healthier options for their pets (Bulochova & Evans, 2021). Consequently, the pet food industry is experiencing notable growth. According to the American Pet Product Association, the pet industry expenditures in 2022 reached approximately $136.8 billion, with pet food and treats constituting 42% of this figure. The upward trend continued into 2023, with a forecasted increase of 4.97% compared to 2022, reaching a total sale of $143.6 billion, of which $62.7 billion represents pet food and treats (American Pet Product Association, ^{2023a, 2023b}). Such a significant expansion of the pet food market has encouraged industry professionals to innovate their pet food products, providing pet caregivers with a broader array of choices in the market (Watson et al., 2023). The industry has been researching pet food product development, such as quality and processing methods to enhance sensory properties, including odors, tastes, and textural properties (Aldrich & Koppel, 2015; Araujo & Milgram, ²⁰⁰⁴).

One of the crucial considerations for purchasing pet foods is whether a product provides complete and balanced nutrition (Schleicher et al., 2019). Purchasers (i.e., pet caregivers) look for options that provide all the macro‐ and micronutrients their pets need (Bos et al., 2023). They also consider other factors, such as color and packaging (Schleicher et al., 2019). Although pet caregivers take into consideration all the factors mentioned before, they choose the best option in the market according to their perspective and feeding philosophy. Beyond nutrition, the palatability of a product is another critical consideration for pet owners. Since dogs and cats are the primary “consumers” of these products, their acceptance is vital. If pets do not find the food product appealing, caregivers are unlikely to make repeat purchases. Consequently, palatability becomes a key factor driving purchasing decisions in the pet food market (Aldrich & Koppel, 2015; Tobie et al., ²⁰¹⁵; Watson et al., ²⁰²³).

Evaluating sensory appeal or palatability of pet food products is challenging because pets cannot verbalize their perceptions and thoughts (Koppel, 2014; Pickering, ^2009a). In previous studies, researchers have evaluated pet foods with human descriptive sensory analysis to understand these products better by developing extended lexicons (Di Donfrancesco et al., 2012; Koppel & Koppel, ²⁰¹⁸; Pickering, ^{2009a, 2009b}). Specific aroma and flavor attributes described in the lexicon for dog foods were “grain,” “oxidized oil,” “brothy,” and “vegetable complex” (Di Donfrancesco et al., 2012). For cat foods, aroma and flavor attributes of “burnt,” “spice,” “methionine,” “soy,” and “vegetables” were included in the lexicon (Pickering, ^{2009a, 2009b}). Researchers have also used analytical instruments, such as gas chromatography and electronic noses (e‐noses), to evaluate the aroma compounds of pet foods and their impact on the intake (Koppel et al., 2013). Hall et al. (2017) also designed a canine olfactometer to determine odor preferences in pet foods. For example, Yin et al. (2020) studied the volatile compounds in different pet foods to determine their impact on intake ratios. Cheli et al. (2017) proposed using e‐noses and e‐tongues to analyze off‐odors/flavors in pet foods associated with food safety concerns. Even though human panelists may evaluate sensory aspects (esp., appearance and aroma) of the pet food products and provide proportionate information about them, the results from pets and human panelists may differ mainly because pets (cats and dogs) and pet caregivers (humans) perceive aromas and flavors differently. For example, in contrast to humans, cats cannot perceive sweet taste because they lack a receptor for recognizing this taste (Baldwin et al., 2014; Li et al., ²⁰⁰⁶), while dogs can perceive the five taste qualities (Spence, 2022). As a result of these differences, researchers have developed methods to determine the acceptability and preferences of pet foods by employing pets as their panelists (Di Donfrancesco et al., 2018, 2012; K.Koppel, ²⁰¹⁴; Tobie et al., ²⁰¹⁵). In these studies, researchers and pet caregivers observed the pets’ behaviors before and after trying the test samples, and they kept a record of the test food they consumed during a specific amount of time to determine acceptability and preference between two or more samples (Aldrich & Koppel, 2015). Also, researchers have developed methods for facial action coding systems to understand pets’ facial expressions (cats and dogs) while they are exposed to food products (Bennet et al., 2017; Bremhorst et al., ²⁰²²).

There are limited publications providing a systematic review of sensory components and palatability tests for pet foods. Koppel (2014) conducted a comprehensive review of existing methods for assessing pet food palatability, including the one‐ and two‐bowl methods. The review also covered descriptive sensory analyses conducted by human panelists to develop specific lexicons for cat and dog foods. In addition, Koppel (2014) examined consumer studies aimed at understanding the behaviors of pet caregivers during their purchasing decisions and highlighted instrumental flavor analyses using equipment such as e‐noses. The researcher also discussed future opportunities in pet food studies. Aldrich and Koppel (2015) reviewed the two traditional palatability testing methods, incorporating case studies to illustrate how to interpret test results, and explored potential directions for future research. Tobie et al. (2015) examined classical methods, such as the one or two‐pan tests, as well as complementary techniques, such as the liking test, the cognitive palatability assessment protocol (CPAP), and exploratory behavior analysis, which measures indicators such as enjoyment by observing pets’ interactions with food. Building on these previous studies, this review aimed to provide an overview of sensory evaluations of pet foods, focusing on key sensory aspects and evaluation methods of foods for cats and dogs. It highlighted the strengths and weaknesses of these methods while also addressing current challenges and future opportunities. In particular, this review explored how emerging technologies, such as artificial intelligence, can enhance the measurement of pet food palatability.

For this concise review, the Scopus database (https://scopus.com) was used to identify articles relevant to the topics covered by employing the keywords of “pet food,” “palatability test,” “one bowl test,” “two bowl test,” or “descriptive sensory analysis.” English‐language articles were retrieved for reviewing the palatability tests of pet food, excluding duplicates and unrelated articles. Additional articles related to the current topics were also sourced through the Google Scholar website (https://scholar.google.com).

2. SENSORY ASPECTS OF PET FOODS AND TREATS

2.1. Pet foods and treats

Pet food products are characterized as products designed for animals, usually domesticated cats and dogs, to provide the daily requirements of carbohydrates, protein, fat, and micronutrients (Gibson & Alavi, 2013). Currently, three types of pet food products are available in the market: dry, moist, and semi‐moist. These are classified depending on moisture content (Crane et al., 2000). While dry pet food has between 10% and 12% moisture (Choi et al., 2023), the semi‐wet presentation has a total of 25%–35% moisture. Wet pet food contains around 60%–87% moisture (Montegove et al., 2022). The most popular type of pet food products is dry kibble due to its greater convenience, longer shelf life, and cheaper cost (Crane et al., 2000). In addition, treats are another product found in the market that pet caregivers use to reward their pets when they accomplish a requested task in training sessions. Being used as a complement to the main intake of pet foods, pet treats should not constitute more than 10% of the daily nutritional requirements for pets (Kepinska‐Pacelik et al., 2023). Different kinds of these products exist, including soft and chewy training treats, cookies/biscuit treats, and dental treats (Nielson et al., 2023).

Pet foods are formulated using a variety of ingredients, offering pet caregivers and their pets a range of options in terms of quality and flavor. The Association of American Feed Control Officials (AAFCO) provides regulatory guidance in the United States on pet food labeling and laboratory standards to ensure the safety of both animals and humans (Association of American Feed Control Officials, 2024). AAFCO requires that all ingredients in pet food be listed on the product label in order of predominance by weight, on an “as formulated” basis, indicating the ingredient that constitutes the highest percentage of the total weight at the time of formulation is listed first (Association of American Feed Control Officials, 2024). The most common protein ingredients in pet food products include beef, chicken and its byproducts, duck, fish, lamb, salmon, meat and its byproducts, soy flour, soybeans, and turkey. Carbohydrate sources typically used in pet food products include barley, brown rice, corn, potatoes, sorghum, flax seeds, oatmeal, various forms of wheat, and vegetables (e.g., carrots and peas). Common fat sources include poultry fat, vegetable oils, corn oil, fish oil, flax seed, and safflower oil. Additionally, dietary fiber, vitamins, and minerals are often added to meet recommended daily nutritional values (Association of American Feed Control Officials, 2024). Additives and preservatives are also included in pet food to maintain freshness and prevent microbiological growth (Case et al., 2011). Some formulations incorporate functional ingredients that offer specific health benefits to pets. For example, peas are used to lower the glycemic index (Di Cerbo et al., 2017). Chondroprotective agents, commonly found in pet supplements for bones and joints, are also used in pet food products (Crane et al., 2000). In addition, palatants and flavor enhancers are employed to improve the taste and overall acceptability of pet food. These can include ingredients such as animal digest, Maillard reaction precursors and products, amino acids (e.g., lysine, arginine, and tryptophan), yeast ingredients, and fats (Samant et al., 2021; Watson et al., ²⁰²³).

Another notable category of pet food products is raw and fresh‐baked (cooked homemade‐style) foods. These products have gained popularity because of the trend of pet humanization, with pet caregivers seeking healthier alternatives such as raw diets and fresh‐baked foods (Bulochova & Evans, 2021). These products are considered healthier than traditional pet foods because they use more natural ingredients and contain fewer additives (Davies et al., 2019). However, despite their growing popularity, there are significant concerns. Raw and fresh‐baked foods often lack the complete nutrients pets typically require, which can lead to nutritional deficiencies, especially in minerals (e.g., selenium, copper, zinc, potassium, calcium, iron, and magnesium) (Choi et al., 2023). In addition, there are microbiological risks, such as contamination with Salmonella or Escherichia coli, associated with raw animal products (Davies et al., 2019).

While the pet food industry often uses similar ingredients for both cats and dogs, their nutritional needs differ significantly (Case et al., 2011; Watson et al., ²⁰²³). Cats, being obligate carnivores, have more specialized dietary requirements than dogs. For example, cats need a higher percentage of protein on a dry matter basis, requiring at least 26% for maintenance, compared to the 18% needed by dogs (Watson et al., 2023). With respect to crude fat, cats require 9%, while dogs need only 5.5%. Cats also have a lower carbohydrate requirement than dogs (Watson et al., 2023). Two specific amino acids, taurine and arginine, are essential in a cat's diet because they cannot synthesize them on their own (Li & Wu, 2023; Watson et al., ²⁰²³; Zaghini & Biagi, ²⁰⁰⁵). Taurine is especially crucial in cats because it is the only amino acid capable of conjugating with bile acids and plays an important role in the execution of urea, which is essential for renal function (Watson et al., 2023; Zaghini & Biagi, ²⁰⁰⁵). In contrast, dogs can synthesize these amino acids, making them nonessential in their diets (Watson et al., 2023). In addition, micronutrients such as vitamin A and niacin are particularly important for cats. Cats lack the enzyme dioxygenase, which is necessary to convert carotenoids into vitamin A (Morris & Rogers, 1982). Additionally, cats require higher levels of niacin because their metabolic pathways cause it to deplete more rapidly than it can be synthesized (Morris & Rogers, 1982). These nutritional differences suggest that pet food formulations vary not only based on species but also on other factors such as pet type and lifespan (Zicker, 2008), influencing the sensory aspects and palatability of the pet food (Hall et al., 2018; Zaghini & Biagi, ²⁰⁰⁵).

2.2. Sensory components of pet foods

Commercially available pet food products offer various flavors and texture options that appeal to both pets and their caregivers. Dry pet food, the most common type on the market, comes in diverse shapes, including triangles, bones, cylinders, cubes, steaks, elongated X, round disks, and rectangles (Gomez Baquero et al., 2018; Koppel et al., ²⁰¹⁴). These foods also appear in different colors and shades, including light and dark brown, golden brown, yellow, green, red, and dark red (Gomez Baquero et al., 2018; Koppel et al., ²⁰¹⁸). While dogs are red‐green colorblind, research indicates that they prefer foods with lighter or brighter appearances (Spence, 2022). Appearance cues in pet food products significantly influence caregivers’ acceptance of these products (Di Donfrancesco et al., 2014). In a consumer study by Di Donfrancesco et al. (2014), which analyzed pet food samples varying in kibble type, ingredients, price, brand, and sensory attributes such as appearance and aroma, the most favored samples were multicolored and multi‐shaped kibbles, while the least liked samples had the darkest brown color. The study also found a strong correlation between caregivers’ overall liking of a pet food and their ratings of its color. However, pet caregivers do not generally consider appearance the most important factor when choosing pet food products when non‐sensory factors are also considered (Schleicher et al., 2019). In a survey of 2181 pet caregivers, the top four factors influencing purchasing decisions were “health and nutrition,” “quality,” “ingredients,” and “freshness” (Schleicher et al., 2019).

Because pet caregivers typically do not taste pet food products, sensory attributes related to ingestion, such as tastes, retronasal odors, and texture (mouthfeel), are less considered compared to appearance. For cats and dogs, however, these sensory attributes are closely related to their food preferences. Olfactory cues, in particular, play an important role in pet food preferences, especially for dogs, who possess a highly sensitive orthonasal olfactory system (Kokocińska‐Kusiak et al., 2021). Studies indicate that dogs prefer beef over chicken due to its odors (Houpt & Smith, 1981). While cats tend to prefer foods with flavors such as meat, fish, sour (or acidic), liver, brewers yeast, yeast extract, and dairy flavors, dogs generally favor flavors such as meat, liver, sweet, sulfur, and baked notes (Tombre, 2004). For pet caregivers, although aroma attributes are not considered as a strong driver of overall liking for pet food products compared to appearance attributes, personal preferences for specific aromas may play a crucial role (Di Donfrancesco et al., 2014). Pet caregivers often prefer foods with minimal off‐odors, such as oxidized oil or musty/dusty notes (Di Donfrancesco et al., 2014; Samant et al., ²⁰²¹), and tend to favor products with low aroma intensity (Di Donfrancesco et al., 2014).

Cats cannot perceive the sweet taste because they lack the gene repertoire associated with sweet taste (Tas1r2) (Baldwin et al., 2014). Like carnivores, cats typically do not consume carbohydrates (Lei et al., 2015). In contrast, domesticated dogs are considered omnivores, and they prefer meat over grains, especially beef and pork (Houpt & Smith, 1981). Also, dogs can perceive sweet taste (Spence, 2022). A study by Ferrell (1984) found that beagle dogs prefer lactose, fructose, and sucrose and are less accepting of maltose. Regarding the salty taste, cats and dogs can tolerate it in low concentrations because meat products have a determined sodium concentration (Bradshaw, 1991, 2006). According to McGrane et al. (2023), cats also prefer an umami taste, which is linked to meat compounds. While both cats and dogs avoid bitter and sour tastes, cats have fewer bitter taste receptors compared to other animals such as herbivores and omnivores (McGrane et al., 2023).

The textural properties (i.e., mouthfeel) of pet food are largely determined by its moisture content. Dry pet foods, with moisture levels between 8% and 9%, are typically crunchy (Samant et al., 2021), while wet foods are available in various forms, including loaves, chunks with gravy, and chunks in loaves (Watson et al., 2023). Cats, which tend to nibble and take smaller bites than dogs, may be more sensitive to the texture and shape of their food, with properties such as tackiness and stickiness being more important (Le Guillas et al., 2024; Tombre, ²⁰⁰⁴). The serving temperature of wet food also affects cats’ preferences (Eyre et al., 2022). In a study with cats over seven years old, those given chunks in gravy at 6°C, 21°C, and 37°C preferred the test food as the serving temperature increased, with the warmest temperature (37°C) being the most preferred, probably due to enhanced flavor intensity (Eyre et al., 2022). Although further research is needed, this finding highlights the importance of serving temperature in enhancing pets’ preferences, consumption, overall nutritional balance, and well‐being, paralleling observations made in human studies (Pramudya & Seo, 2018).

3. TRADITIONAL AND CURRENT METHODS FOR MEASURING PET FOOD PALATABILITY BASED ON PETS’ RESPONSES

Palatability is defined as “a measure of subjective food preference and depends on taste, texture, and odor” (Araujo & Milgram, 2004). To assess the palatability of pet food, researchers have developed methods primarily focused on the pets themselves. Human input, whether from trained panelists conducting descriptive sensory analysis or untrained pet caregivers conducting consumer tests, has also been incorporated to characterize sensory attributes and predict palatability. In addition, instrumental analyses have been employed for objective measurements of sensory attributes and to indirectly predict palatability. This section first addresses traditional and current palatability testing methods that focus on pets’ sensory and behavioral responses. Section 4 will discuss approaches involving human and instrumental analyses.

To evaluate pets’ palatability, researchers typically use two types of tests: preference and acceptability. The preference test, also known as the two‐bowl test, presents two samples simultaneously, allowing pets to indicate a preference by either sniffing or consuming one option first. In contrast, the acceptance test, also known as the one‐bowl test, presents a single sample and measures intake to assess consumption levels (Kvamme, 2003; Tobie et al., ²⁰¹⁵). This test is also used in home environments, as pets can be either trained or untrained for the one‐bowl test. However, for the two‐bowl test, which is used to distinguish between two test samples, pets must receive training. Trained pets typically provide more accurate results compared to their untrained counterparts (Tobie et al., 2015). In addition, pets’ palatability is assessed through pets’ behavioral responses (Becques et al., 2014; Rogues et al., ²⁰²²) and facial expressions (Bremhorst et al., 2022; Hanson et al., ²⁰¹⁶) toward the test samples. Table 1 provides a summary of these methods, highlighting their respective strengths and weaknesses.

TABLE 1.

Summary of the strengths and weakness of each palatability test.

Category	Name	Strength	Weakness
Acceptability test	One‐bowl test	Test conditions are similar to those present at home, i.e., one bowl per trial. Untrained pets can participate.	Cannot provide detailed information for marketing claims and product development improvements.
	In‐home use test using one‐bowl test	Pets are fed on their regular schedule and place. Untrained pets can participate.	Uncontrolled parameters, such as noise and ambiance, can affect the results. Requires 100 untrained pet participants.
Preference test	Two‐bowl test	Two different products can be compared.	Pets need to be trained to participate. Extensive training and time are required. It can evaluate just two samples per time.
	Two‐bowl test using canine olfactometer	Complement preference test results with olfaction liking.	It needs to be run in a central location test. Pets need to be trained to participate.
	Cognitive palatability assessment protocol	Different types of foods (obvious differences) can be tested. Fewer pets can be used compared to those in the traditional two‐bowl test.	Real‐life conditions are not adapted. Pets that have a higher functioning cognitive ability can participate. Pets that have a higher cognitive ability can participate.
	Preference ranking test	More than two samples can be evaluated. Preferences of multiple samples can be evaluated. Less amount of food is needed compared to that of other tests.	Pets can get fatigue after evaluating up to five samples. Pets need time and familiarization with the method. The method is time‐consuming.
Behavior test	Behavior analysis	Behavioral aspects, such as sniffing and licking, complement the results of the traditional methods.	Results obtained by observations can be subjective.
	Facial expression analysis	Pet facial expressions can complement the results of preference tests.	Run in central test locations. Extensive time is required. Technology capable of interpreting facial expressions is essential.

3.1. Acceptability test

While the standard nine‐point hedonic scale is commonly used to measure how much human consumers like or dislike a test sample, it cannot be adapted for pets. Therefore, the one‐bowl test is typically used to assess pet food acceptability in both controlled laboratory settings and home environments. The one‐bowl test notably offers insights into “acceptance” but does not provide information on “preference” or “degree of liking” (Aldrich & Koppel, 2015). However, this method closely resembles real‐life home environments, where pets are typically served a single type of food per meal.

3.1.1. One‐bowl test

The one‐bowl test, also known as the single‐bowl test, one‐pan test, or monadic test, involves presenting a pre‐weighed food sample as the pet's main meal for a specific period, typically around 15 or 30 min (Aldrich & Koppel, 2015; Samant et al., ²⁰²¹; Tobie et al., ²⁰¹⁵). The test samples must provide the animal's daily caloric intake plus an additional 10% to ensure the pet's satiety (Aldrich & Koppel, 2015). The one‐bowl test assesses acceptability by measuring food intake and the speed of consumption. Food intake is calculated by subtracting the remaining food in the bowl after the test period from the initial pre‐weighed amount (Aldrich & Koppel, 2015). This method also helps researchers identify aromas, off‐flavors, or textures that pets find unappealing (Aldrich & Koppel, 2015; Tobie et al., ²⁰¹⁵). Conducted under regular feeding conditions, preferably at the same feeding schedule, the test is usually repeated over 5 or more days (Aldrich & Koppel, 2015). After 5 days, the food can be switched to a different sample to compare intake (Aldrich & Koppel, 2015). Di Donfrancesco et al. (2018) conducted a one‐bowl test at home with four dog food samples containing sorghum fractions to measure acceptability. Each sample was fed for 5 days for a total of 20 consecutive days. The study included a control diet, whole sorghum diet, sorghum flour diet, and sorghum mill‐feed diet, finding that the sorghum‐containing samples were accepted similarly to the control. In industry setting, the one‐bowl test can be completed in 2 days by giving Sample A on Day 1 and Sample B on Day 2 (Lambrakis & Kersey, 2021). Although there is no standardized protocol for the number of pets required in an acceptability test (Pires et al., 2020), studies have reported that 8–30 pets are sufficient in a one‐bowl test to identify acceptability trends of a test sample (Aldrich & Koppel, 2015; Tobie et al., ²⁰¹⁵). To minimize the potential influence of an unfamiliar laboratory setting on pet food acceptability, multiple dogs or cats can be used, with a balanced design across test periods (Aldrich & Koppel, 2015).

3.1.2. In‐home use test using the one‐bowl test

The in‐home use test (IHUT) is an adaptation of the one‐bowl test run at home, where it maintains its normal feeding schedule and location (Aldrich & Koppel, 2015; Koppel, ²⁰¹⁴; Tobie et al., ²⁰¹⁵). The principal advantage of the IHUT is to offer a more practical setting compared to a central location test (controlled laboratory setting). Another critical consideration is that pet's food preferences may be influenced by its feeding environment, being familiar with the environment can give more accurate results compared to a central location test (Griffin et al., 1984; Samant et al., ²⁰²¹). The main disadvantage of the IHUT is to require a larger number of pet participants due to the uncontrolled variables present. For example, the IHUT requires 100 untrained pets compared to 30 trained pets in a central location test (Samant et al., 2021; Tobie et al., ²⁰¹⁵). Additionally, since only one bowl (i.e., one sample) is tested, the information gathered is limited to consumption ratios and pet caregivers’ observations (Aldrich & Koppel, 2015).

Another variation of the one‐bowl test is the liking test (Rogues et al., 2022). Becques & Niceron (2014) proposed a protocol for measuring liking using a one‐bowl test with trained pets than untrained ones. This approach employs trained pet panelists to reduce bias and improve the accuracy of the one‐bowl test (Rogues et al., 2022). This test considers several criteria: the consumption ratio, calculated by dividing the percentage of food consumed by the total food offered; the percentage of finished bowls, which refers to the amount of food consumed; a comparison of the test consumption with the reference consumption level for each pet, based on its intake history; and the refusal percentage, defined as a case where the pet panelist consumes none of the offered food (i.e., consumption = 0).

3.2. Preference test

3.2.1. Two‐bowl test

The two‐bowl test, also known as the two‐pan test, assesses a pet's food preference by offering two different food samples simultaneously for approximately 15–30 min (Aldrich & Koppel, 2015; see also Pires et al., ²⁰²⁰). In this test, Samples A and B are weighed separately in two separate bowls, with each sample covering the pet's entire daily caloric requirement (Aldrich & Koppel, 2015; Hall et al., ²⁰¹⁷; Samant et al., ²⁰²¹). Researchers observe which sample is chosen for the first bite, often associated with olfactory attractiveness (Aldrich & Koppel, 2015). After the test duration, the remaining food is weighed to calculate the Intake Ratio for each sample. Some designs use multiple test days with randomized bowl placement to reduce bias (Samant et al., 2021).

One important feature of the two‐bowl test is that participating pets must be trained (Tobie et al., 2015). These pets are trained to conduct various controlled procedures and are exposed to different types of food, such as wet, semi‐wet, and dry, while also living in groups (Rogues et al., 2022). The sample size typically consists of 30 trained pets (Tobie et al., 2015), which is smaller than in the one‐bowl test, as trained pets reduce the risk of bias (Samant et al., 2021). Pires et al. (2020) proposed that a minimum number of 23 cats is required for the two‐bowl test, assuming a standard deviation of 0.20.

The intake ratio is calculated using Equation (1), where A represents the amounts (grams) of Sample A consumed, and B represents the amounts (grams) of Sample B consumed. A ratio of 0.5 indicates no preference between the Samples A and B. Ratios below 0.5 indicate a preference for Sample B, while ratios above 0.5 suggest a preference for Sample A (Aldrich & Koppel, 2015).

$$\frac{A}{A+B}\mathrm{Intake}\mathrm{Ratio}.$$ (1)

Another indicator measured in the two‐bowl test is the Consumption Ratio, calculated by dividing the amount of Sample A eaten by the total amount of food (Samples A and B) initially served, denoted as C (see Equation 2) (Samant et al., 2021). This metric is especially useful in studies involving dogs of varying sizes, because it accounts for differing levels of food consumption (Rogues et al., 2022). The interpretation of this Consumption Ratio follows a similar pattern to the Intake Ratio: a value below 0.5 indicates a preference for Sample B (Aldrich & Koppel, 2015).

$$\frac{A}{C}\mathrm{Consumption}\mathrm{Ratio}.$$ (2)

Pet food preference has been studied for over 60 years. Hesgeted et al. (1956) and Waterhouse and Fritsch (1967) were pioneers in developing the general methodology for the two‐bowl tests and addressing sources of bias, such as sample order, by randomizing bowl placement. Rashotte et al. (1984) assessed preference between dry foods with varying fat content, while Griffin et al. (1984) measured preferences for canned, semi‐moist, and dry foods in both home and kennel settings. Their study found that dogs showed consistent preferences for canned food across environments, but preferences for dry or semi‐moist foods were influenced by the dog's prior exposure (Griffin et al., 1984). Early life feeding experiences can also affect food preferences (Hepper et al., 2012), potentially leading to neophobia, an aversion to trying new foods (Bradshaw, 1986).

3.2.2. Two‐bowl test using canine olfactometer

Olfaction is a key element in shaping food preferences for both humans (Boesveldt & Parma, 2021; Seo & Hummel, ²⁰⁰⁹) and pets (Bradshaw, 1991; Houpt et al., ¹⁹⁷⁸) because smell and taste are fundamental for perceiving flavor (Hall et al., 2017). Traditional methods for assessing pet food palatability, such as the one‐bowl and the two‐bowl tests, rely on measuring food consumption to calculate the Intake Ratio (Koppel, 2014; Tobie et al., ²⁰¹⁵). However, these tests typically overlook the impact of olfactory preference (Basque et al., 2018). To address this gap, previous researchers have developed a canine olfactometer to supplement preference tests (Basque et al., 2018; Hall et al., ²⁰¹⁷).

Hall et al. (2017) investigated the preferences of dogs for four commercial chicken‐flavored dog foods using a two‐bowl test. Their apparatus included scales to immediately weigh the consumed food and an olfactometer to assess odor preference. The device featured two stainless‐steel plates for the food, each with a corresponding port through which dogs could detect the odors. The dogs were pre‐trained to use the olfactometer by receiving a food reward when they placed their noses near the odor port, helping them become familiar with the equipment and protocol. During the test, the dogs were initially given 15 s to observe and smell the food samples through a removable fiberglass screen before being allowed to eat. The results showed that the dogs made their food choices during this smelling period. Dogs consistently consumed more of the sample they selected by sniffing, suggesting that dogs may not need to taste food to make a decision, with olfactory cues playing a critical role in their food selection (Hall et al., 2017; Houpt et al., ¹⁹⁷⁸). Additionally, the study found that dogs preferred the sample with a food odor over the odorless one (Hall et al., 2017).

Basque et al. (2018) used the same olfactometer developed by Hall et al. (2016, 2017) to evaluate four products varying in palatability enhancers, including fat‐coated kibbles with and without premium or super‐premium meat palatability enhancers. The samples were also tested using the two‐bowl method. The results showed a clear preference for the super‐premium samples in both tests. They recommended further research on the odor preference test, suggesting the use of smaller sample sizes (2 g instead of 100 g) to reduce the gaseous phase and assess its impact on the outcomes (Basque et al., 2018).

Petel et al. (2018) described an adaptation of the two‐bowl test to evaluate odor detection, known as false‐bottom bowls. This method presents two stainless‐steel bowls with a perforated bottom, allowing the release of different odors (Petel et al., 2018; Samant et al., ²⁰²¹).

3.2.3. Cognitive palatability assessment protocol

Over the years, researchers have developed various methods to assess pet food preferences, including the CPAP, developed by Araujo and Milgram (2004). This method offers an alternative to traditional palatability tests, such as the two‐bowl test (Araujo et al., 2004), by incorporating an object discrimination task based on associative learning (Araujo & Milgram, 2004). In this method, pets are trained to associate specific objects, such as a margarine container, a soft‐drink can, and a diskette box, with particular food items (Araujo et al., 2004). During the test, three objects are presented simultaneously: two linked to food rewards and one serving as a control with no food reward. If the pet chooses the control object (e.g., a diskette box), no food is provided. However, when the pet selects one of the two objects associated with specific food items (e.g., a margarine container for buttermilk and a soft‐drink can for soda) based on its preference, it receives the corresponding food as a reward. The test ultimately determines which food the pet prefers by analyzing its choices (Araujo et al., 2004). CPAP provides several advantages. It allows researchers to evaluate preferences between distinctly different food types, such as dry versus wet food, and requires fewer pet panelists than traditional methods such as the two‐bowl test (Aldrich & Koppel, 2015; Tobie et al., ²⁰¹⁵). Additionally, the preferences expressed by the pet panelists are stable across time and repetitions, suggesting that CPAP is less influenced by prior feeding or satiety compared to other methods such as the two‐bowl test (Araujo et al., 2004; Tobie et al., ²⁰¹⁵). However, CPAP has some drawbacks. It is time‐intensive and requires significant effort and training to conduct the tests (Aldrich & Koppel, 2015; Tobie et al., ²⁰¹⁵). In addition, CPAP may not fully replicate real‐world feeding conditions (Aldrich & Koppel, 2015). Adapting CPAP for use with cats presents further challenges because it requires sustained attention from the pet panelists, making it more difficult to implement (Tobie et al., 2015). These limitations could restrict the broader applicability of the protocol.

3.2.4. Preference ranking test

While the two‐bowl test evaluates a pet's preference between two samples, there is a growing need for methods that can assess a pet's preference among more than two options. One such method is the preference ranking test, which ranks a pet's choices through multiple comparisons, allowing for the assessment of preferences across several samples (e.g., more than two). Li et al. (2018) proposed a preference ranking test to determine 12 Beagle dogs’ preferences among 5 edible treat samples inserted into rubber toys (“Kongs”). In this method, each dog was expected to extract its preferred treat based on aroma, with the most favored treat chosen first and the least preferred last. The test was conducted in five phases, each lasting 5 consecutive days. Phase 1 served to familiarize the dogs with the procedure, such as retrieving treats from the toys, and to assess which dogs were suitable for the study. During this phase, the dogs were observed and evaluated across five segments: (a) whether the dog smelled all the toys, (b) whether the dog showed initial interest in the test, (c) whether the dog maintained interest, (d) whether the dog successfully retrieved the treats, and (e) whether the dogs required assistance. These behaviors were scored on a scale of 1 (“Did not do the task at all”) to 5 (“Did the task perfectly”) in each segment, with a maximum possible total score of 25. Only dogs scoring at least 15 points in total were considered “qualified” for the study (Li et al., 2018). While dogs that did not meet this requirement were allowed to complete a study, their data from later phases were excluded from the analysis. As Phase 1 was primarily an orientation and practice session, data from this phase were not analyzed. In the subsequent Phases 2–4, the dogs participated in the preference ranking test for five treat samples that varied by ingredient, that is, fats, starches, and proteins in Phases 2–4, respectively. In Phase 5, five commercially available treats were tested. The results showed that the 12 dogs were able to consistently select 1 or 2 favorite treats over others, indicating that the preference ranking test is effective for comparing more than 2 samples simultaneously (Li et al., 2018). In a follow‐up study conducted 1 year later using the same dogs and methodology, Li et al. (2020) obtained similar results to their previous study (Li et al., 2018), indicating that the preference ranking test method is reliable. However, the primary drawback of this method is the significant time and effort required to complete the five phases compared to more conventional methods (Li et al., 2018). Since this method is new and in its early stages, further studies are needed to optimize testing conditions, such as the number of test days, sample sizes, and the type of food or treats tested. These additional studies will help generalize and validate this method for assessing pet food palatability.

3.3. Behavior analysis

Another key element in assessing the palatability of pet food is observing a pet's behavior before, during, and after consuming the test sample (Rogues et al., 2022; Watson et al., ²⁰²³). Van den Bos et al. (2000) identified taste reactivity patterns in cats, highlighting behavioral responses that indicate either preference or aversion to a particular food. For example, behaviors such as lip licking, face grooming, and licking the feeding bowl are associated with liking (Van den Bos et al., 2000). In contrast, nose licking and a combination of licking and sniffing the food are related to aversion (Van den Bos et al., 2000). To complement traditional methods, Becques et al. (2014) studied cats’ eating behavior by examining factors such as food consumption, speed of eating, latency to eat, posture while eating, as well as sniffing and licking. They found that cats spent less time sniffing highly palatable kibbles during the first encounter, although this pattern reversed during their second exposure in the study (Becques et al., 2014). They also observed that neither latency nor speed of consumption was a reliable indicator of kibble palatability.

It is important to note that behavioral responses to food may vary due to numerous factors, including species types, individual characteristics, hunger or satiety, feeding environment, prior feeding experiences, and even pet caregivers’ characteristics (Becques et al., 2014; Bradshaw, ²⁰⁰⁶; Bradshaw & Cook, ¹⁹⁹⁶). For example, Bradshaw and Cook (1996) found no uniform behavioral patterns among the 36 domestic house cats observed when presented with food. Therefore, further research is necessary to identify behavioral parameters that consistently reflect food preferences across these varying factors.

3.4. Facial expression analysis

Human facial expressions have long been linked to underlying emotions through the Facial Action Coding System (FACS) and its action units (Ekman & Friesen, 1978; Ekman et al., ¹⁹⁸³). Increasingly, researchers are exploring how these expressions can also be connected to food‐evoked emotions and food acceptance (de Wijk et al., ²⁰¹²; Samant & Seo, ²⁰¹⁹). Similarly, efforts have been made to analyze facial expressions in pets to understand their emotions and responses to food (Bennet et al., 2017). While Leyhausen (1979) developed a facial ethogram to capture cats’ offensive and defensive moods, his work did not record the movements of the eyes, lips, or tongue. To address this gap, Bennet et al. (2017) applied a facial action coding system for cats (“CatFACS”) to characterize their facial expressions and behaviors. They analyzed video recordings of 29 cats under two conditions: one with no human interaction (cats alone) and another with human interaction. Using cluster analysis on 275 video clips, the study identified three distinct groups of facial expressions and behaviors corresponding to fear, frustration, and relaxed engagement. Fear was associated with blinking and half‐blinking, while frustration was linked to expressions such as hissing, nose‐licking, jaw‐dropping, upper‐lip‐raising, nose‐wrinkling, lower‐lip‐depression, lip‐parting, mouth‐stretching, and tongue‐showing. Relaxed engagement was related to right‐gaze and head‐turn bias (Bennet et al., 2017). In another study, Hanson et al. (2016) examined cats’ reactions, particularly facial expressions, to pleasant (l‐Proline) and unpleasant (quinine monohydrochloride dihydrate) tastes, as well as a mixture of both at five different concentrations. The 13 cats involved displayed half‐closed eyes for significantly longer periods in response to pleasant tastes compared to water. Additionally, pleasant tastes more frequently induced facial expressions such as tongue protrusions, mouth smacks, and nose licks.

Facial expression analysis has also been applied to dogs using the dog facial action coding system (“DogFACS”) (Waller et al., 2013). Bremhorst et al. (2019) examined the facial expressions of 29 Labrador Retrievers under positive and negative conditions, with food rewards used to induce positive anticipation and the absence of rewards to induce frustration. The dogs exhibited distinct facial expressions in response to these conditions. In the positive condition (food reward), the ear adductor movement occurred more frequently, whereas in the negative condition (no food reward), expressions such as blinking, lip‐parting, jaw‐dropping, nose‐licking, and ear‐flattening were more common. A follow‐up study (Bremhorst et al., 2022) found similar results, with dogs showing ear adduction when a food reward was presented, and a range of negative expressions, including blinking, ears‐flattening, lip‐parting, jaw‐dropping, nose‐licking, downward ear movement, tongue‐showing, lip‐corner‐pulling, and upper‐lip‐raising when presented with a toy instead of food reward.

The analysis of pets’ facial expressions offers valuable insights into their emotional states and perceptions of food, particularly since they cannot verbalize their responses. This approach could supplement or even replace traditional methods of assessing pet food palatability, such as consumption and forced‐choice tests. However, there are challenges to address. For example, no facial expressions were exclusively linked to either positive or negative conditions (Bremhorst et al., 2022, 2019). While ear‐flattening and downward ear movements were observed in about 89% of negative conditions, these expressions also occurred in 55% and 44% of positive conditions, respectively (Bremhorst et al., 2022). Additionally, neutral facial expressions and those evoked by food can vary depending on species, breed, and morphological differences (Bennet et al., 2017; Bremhorst et al., ²⁰¹⁹). Further research is needed to overcome these challenges and enhance the accuracy of facial expression analysis in assessing pet food palatability.

4. SENSORY CHARACTERISTICS OF PET FOODS: HUMAN AND INSTRUMENTAL ANALYSIS

4.1. Descriptive sensory analysis

Descriptive sensory analysis is a method used to characterize sensory attributes of test samples. This method is extensively used in industry to evaluate sensory properties in shelf‐life studies, market, quality control, and product development. In this type of test, the attributes or descriptors for a specific product can be developed, and the intensity of each one is determined by using scales. The descriptors are labels created by trained panelists who have received extensive training to name the characteristics of test samples (Murray et al., 2001). The most common descriptive sensory analysis techniques are the Flavor Profile, Texture Profile, Quantitative Descriptive Analysis, and Spectrum Descriptive Analysis (Meilgaard et al., 2015; Murray et al., ²⁰⁰¹).

In the research and development of pet food products, processors, researchers, and marketers in the pet food industry must characterize the sensory attributes of test samples to gain a better understanding of the products. However, since pets cannot verbalize their perceptions of these sensory attributes or their intensities, descriptive sensory analysis using trained human panelists can be invaluable because it can characterize the test samples’ sensory attributes (Koppel, 2014). Although humans are not the end users, this analysis provides detailed descriptions of the sensory attributes and their intensities, which can assist industry professionals in product development, quality control, and marketing. Additionally, since pet caregivers are the purchasers of pet food, the descriptive sensory analysis results can help them better understand the sensory attributes of the products they choose for their pets.

Researchers have taken steps to create lexicons specific to different categories of pet foods by employing human descriptive panelists (Koppel, 2014; Samant et al., ²⁰²¹). Lin et al. (1998) were the pioneer in using descriptive sensory analysis for pet foods. They studied the effects of the sensory appearance and aroma attributes when applying different lipids and processing conditions, concluding that the fat percentage and the source effectively modified the intensity of aroma attributes such as “fatty,” “cardboard,” and “painty.” Pickering developed two lexicons: one for canned cat foods (Pickering, ^2009a) and the other for dry cat foods (Pickering, ^2009b). Additionally, Koppel and Koppel (2018) developed a lexicon for aroma attributes of retorted cat food. For dog foods, Di Donfrancesco et al. (2012) developed an extensive lexicon for dry dog foods with 72 attributes related to appearance, aroma, flavor, and texture.

Koppel et al. (2013) determined the volatile compounds of 14 different commercial dry dog foods using an analytical method to correlate which of these compounds can be associated with characteristic aromas evaluated by descriptive sensory human panelists using attributes from Di Donfrancesco et al. (2012) lexicon. The results showed that some aldehydes were related to rancidity (Koppel et al., 2013). Koppel et al. (2015) also studied the effect of fiber and kibble coatings on the sensory characteristics of dry dog food, using a human sensory descriptive panel based on the lexicon guidelines established by Di Donfrancesco et al. (2012). Tables 2, 3, 4, 5, 6 present the descriptors for appearance, aroma, flavor, texture, and aftertaste used in the lexicons available for pet foods.

TABLE 2.

Appearance‐related attributes used in previous studies of the descriptive sensory analysis of foods for cats, dogs, and pets.

Attributes	References
	Cat foods	Dog foods	Pet foods
Brown color		[1–3]
Fibrous		[1–3]
Flecks		[1,3]
Grainy	[4]	[1–3]
Green color		[1]
Lightness, overall color		[1,5]
Oil		[1]	[5]
Porous		[1–3]
Red‐brown color		[1,2]
Shape		[1]
Shape uniformity		[1]
Size		[1]
Size uniformity		[1]	[5]
Starchy		[1]
Surface roughness/smoothness		[1]	[5]
Wet moist		[1]

[1] Di Donfrancesco et al. (2012); [2] Koppel et al. (2014); [3] Koppel et al. (2015); [4] Koppel & Koppel (2018); [5] Lin et al. (1998).

TABLE 3.

Aroma‐related attributes used in previous studies of the descriptive sensory analysis of foods for cats, dogs, and pets.

Attributes	References
	Cat foods	Dog foods	Pet foods
Ashy		[1]
Bacon	[8]
Barnyard	[5]	[1–4]
Brothy‐beef	[5]
Brothy‐poultry	[5]
Brothy‐seafood	[5]
Brothy/brothy‐overall		[1–4]
Brown		[1–3]
Burnt	[7,8]	[1,2]
Butyric	[5]
Caramel	[7,8]
Cardboard		[1–4]	[6]
Celery		[1]
Clove		[1]
Cooked	[5]	[1,2]
Dusty/earthy		[1,2,4]
Earthy	[5]	[1]
Egg		[4]
Fatty			[6]
Fermented		[1,2]
Fish/tuna	[7,8]	[1–4]
Garlic		[1]
Grain/cereal	[5]	[1–4]	[6]
Hay‐like	[5]	[1,4]
Heated oil	[5]
Herb	[5,7,8]
Iron		[4]
Liver	[5]	[1–4]
Meaty‐beef	[5]
Meaty‐poultry	[5]
Meaty‐seafood	[5]
Meaty/meaty overall	[5]	[1–3]
Metallic		[2,3]
Methionine	[7,8]
Musty		[1–3]
Musty/dusty	[5]	[1,2]
Oil		[2]
Overall sweet	[5]
Oxidized oil/painty		[1–4]	[6]
Pepper, black/spice pepper	[5]	[1,2]
Plastic		[1,2]
Processed	[5]
Pungent	[5]	[1]
Rancid	[7,8]
Smoky		[1,2]
Sour		[4]
Soy	[7,8]	[1,2]
Spice brown	[5]	[1,2]
Spice complex	[7,8]	[1,2]
Stale		[1–4]
Starchy		[1]
Straw‐like		[1,2]
Sulfur	[5]
Toasted		[1–4]
Vegemite	[8]
Vegetable complex/vegetable	[5,7]	[1]
Vitamin	[5]	[1,3,4]
Yeast	[5]

[1] Di Donfrancesco et al. (2012); [2] Koppel et al. (2013); [3] Koppel et al. (2014); [4] Koppel et al. (2015); [5] Koppel & Koppel (2018); [6] Lin et al. (1998); [7] Pickering (2009a); [8] Pickering (2009b).

TABLE 4.

Flavor and taste‐related attributes used in previous studies of the descriptive sensory analysis of foods for cats and dogs.

Attributes	References
	Cat foods	Dog foods
Barnyard		[1–3]
Bitter	[4,5]	[1–3]
Bread crust	[5]
Brothy		[1–3]
Brown		[1,2]
Burnt		[1]
Cardboard		[2,3]
Carrot		[1]
Chicken	[5]
Dusty/earthy		[3]
Egg		[1,3]
Fish		[1–3]
Garlic		[1]
Grain/cereal	[4,5]	[1–3]
Hay‐like		[1,3]
Iron		[3]
Liver		[1–3]
Meaty	[4,5]	[1,2]
Metallic		[1,3]
Musty		[1,2]
Oil		[1]
Onion		[1]
Oxidized oil		[2,3]
Pepper, black		[1]
Prawn	[4,5]
Salt	[4,5]	[1–3]
Smoky		[1]
Sour	[4,5]	[1–3]
Soy		[1]
Spice brown		[1]
Spice complex		[1]
Stale		[1–3]
Sweet	[4,5]	[1–3]
Toasted		[1–3]
Vegetable complex		[1]
Vitamin		[1–3]

[1] Di Donfrancesco et al. (2012); [2] Koppel et al. (2014); [3] Koppel et al. (2015); [4] Pickering (2009a); [5] Pickering (2009b).

TABLE 5.

Texture‐related attributes used in previous studies of the descriptive sensory analysis of foods for cats and dogs.

Attributes	References
	Cat foods	Dog foods
Amplitude		[1]
Cohesiveness of mass		[1–3]
Fibrous		[1,3]
Firmness		[1]
Fracturability/brittleness	[4]	[1–3]
Graininess		[1]
Gritty	[4]	[1,3]
Hardness	[4]	[1–3]
Initial crispiness/crispness		[1–3]
Mouthcoat		[1,2]
Oily mouthfeel		[1]
Overall impact		[1]
Powdery		[1,2]
Roughness	[4]
Springiness		[1]

[1] Di Donfrancesco et al. (2012); [2] Koppel et al. (2014); [3] Koppel et al. (2015); [4] Pickering (2009b).

TABLE 6.

Aftertaste‐related attributes used in previous studies of the descriptive sensory analysis of foods for cats and dogs.

Attributes	References
	Cat foods	Dog foods
Barnyard		[1,2]
Bitter		[1,2]
Brothy		[1]
Brown		[1]
Cardboard		[1,2]
Fish		[1,2]
Grain		[1]
Liver		[1,2]
Meaty		[1]
Metallic		[1,2]
Musty		[1]
Oxidized oil		[1,2]
Salt		[1,2]
Sour		[1,2]
Stale		[1,2]
Sweet		[1,2]
Toasted		[1]
Vitamin		[1,2]

[1] Koppel et al. (2014); [2] Koppel et al. (2015).

4.2. Consumer study by pet caregivers

A consumer study using pet caregivers provides valuable insights into their purchasing behaviors for pet food products. Such studies often utilize questionnaires to gather information, including demographic data about pet caregivers and questions about their pets’ characteristics (Koppel, 2014). For example, as addressed in Section 2.2. sensory components of pet food, an online survey, completed by 2181 pet caregivers, found that the four factors, such as “health and nutrition,” “quality,” “ingredients,” and “freshness,” were considered very important or extremely important when purchasing pet food products (Schleicher et al., 2019). In contrast, less important factors, rated between 1 point (not at all important) and 2 points (slightly important) on a five‐point scale, included “gluten free,” “packaging,” “color,” and “on sale” (Schleicher et al., 2019). The study also found that most pet caregivers rely on commercial products, with cat owners being more likely to feed their pets wet food compared to dog owners (Schleicher et al., 2019). Another study by Prata (2022) examined emerging trends in pet food in Portugal through an online survey of 74 pet caregivers. The findings revealed that 67.5% of participants fed their pets commercial diets, while 19.0% chose alternative options, such as natural and organic foods, due to the perception of superior quality. The top factors influencing pet food purchases were suitability for the pet's metabolism and age (39.7% of participants), followed by the ingredient list (31.5%). Interestingly, while flavor (1.4%) and laboratory validation (1.4%) were deemed less important in purchasing decisions, flavor (32.7% of participants) was the key motivator for pet caregivers to switch diets (Prata, 2022).

Previous studies have assessed pet caregivers’ acceptance and preference for the appearance of kibble products. For example, Koppel et al. (2018) asked 120 consumers in Thailand to evaluate 30 different kibbles using a 9‐point hedonic scale. The study also incorporated Check‐All‐That‐Apply (CATA) questions, covering attributes such as natural ingredients and overall health benefits. The findings showed that Thai consumers favored kibbles with yellow colors and bone shapes, while dark brown and small‐sized kibbles were the least preferred. These results indicate the importance of appearance factors, such as color and shape, in influencing repurchase decisions (Koppel et al., 2018). Similarly, in Poland, 120 consumers assessed 30 different dry dog foods using an online survey (Gomez Baquero et al., 2018). They rated the product samples on a 9‐point hedonic scale and completed CATA questions about functional product attributes. Polish consumers showed a preference for medium‐sized kibbles in brown or golden‐brown colors and disliked kibbles that were too small or large. Shape preferences leaned toward triangle‐prism kibbles, while stick shapes were the least preferred (Gomez Baquero et al., 2018).

Human studies reported that consumer purchasing behavior is influenced not only by sensory preferences but also by the emotions‐evoked byproduct samples (de Wijk et al., ²⁰¹²; Samant & Seo, ²⁰²⁰). While food samples may not differ significantly in hedonic ratings, they can be distinguished by the emotions they evoke (de Wijk et al., ²⁰¹²; Pramudya et al., ²⁰²¹; Samant & Seo, ²⁰²⁰). Similarly, consumer research has examined the emotional connections between pet caregivers and pet food, extending beyond hedonic evaluations of sensory attributes such as appearance and aroma. For example, Delime et al. (2020) studied the relationship between kibble odors and the emotional responses of pet caregivers in France, the United States, and La Réunion Island. A total of 583 pet caregivers (294 cat owners and 289 dog owners) evaluated the odors of 9 kibbles with respect to odor liking rated on a 9‐point hedonic scale and emotions measured using EsSense25 (Nestrud et al., 2016) with the CATA methodology. Additionally, the flavored kibble samples were evaluated by a descriptive sensory analysis using Petscript^®, a universal sensory lexicon designed to characterize the odor of pet food (Delime et al., 2018). Findings showed that while pet food samples were discriminated by odor likings, they differed in terms of evoked emotions. Odor liking ratings of pet food samples were positively associated with pleasure and activation. Higher overall odor intensity was associated with activation. Specific odor attributes were also connected to emotional attributes. For example, odor attributes of “spicy,” “aromatic herbs like,” “yeast—bouillon like,” and “roasted chicken like” were associated with arousal, such as enthusiasm and activity, while “fatty—rancid,” “viscera like,” and “cereal like” were associated with de‐activation such as calmness and boredom. Notably, cross‐cultural differences were observed in odor liking and emotions. While the American pet caregivers were more sensitive to negative emotions and rated odor liking lower than French and La Réunion caregivers, French pet caregivers used emotions related to pleasure and activation (e.g., “enthusiastic,” “active,” “adventurous,” and “joyful”) to differentiate pet food samples more often than American pet caregivers.

In a recent study, Tsai et al. (2020) developed a comprehensive list of emotion terms to characterize the emotional experiences of pets and their caregivers. Through a series of four 90‐min focus group sessions, the researchers identified 39 conceptual emotion terms for dogs and 53 for cats. Additionally, 33 emotion terms were ascribed to dog caregivers and 60 to cat caregivers. Several emotion terms, such as “content/satisfied,” “energetic,” “excited,” “happy,” “anxious,” “embarrassed,” “frustrated,” and “fearful,” were commonly used to describe both pets and their caregivers (Tsai et al., 2020). This overlap may be attributed to anthropomorphism (Arahori et al., 2017; Wynne, ²⁰⁰⁷), where pet caregivers project their own emotions onto their pets (Tsai et al., 2020), leading to shared emotion terms.

4.3. Volatile organic compounds analysis

Pet food products are complex, containing a variety of aromas and flavor attributes (Di Donfrancesco et al., 2012). The aromas are characterized by volatile organic compounds (VOCs), typically present in low concentrations. These VOCs are analyzed using analytical instruments such as gas chromatography combined with mass spectrometry (Feng et al., 2019; Koppel et al., ²⁰¹³).

Koppel et al. (2013) identified VOCs that influence the sensory profiles of dry dog food samples. They discovered 54 compounds, including alcohols, aldehydes, pyrazines, furans, alkanes, benzene derivates, and terpenes, with aldehydes and ketones being the most abundant VOCs across six grain‐free and eight grain‐added dry dog food samples. Total VOC levels were higher in grain‐added samples compared to grain‐free ones. Furthermore, a combination of instrumental and descriptive sensory analyses revealed associations between VOCs and sensory attributes, such as benzene derivatives with plastic‐like aromatics and aldehydes with rancid aromatics (Koppel et al., 2013). Chen et al. (2017) conducted research to identify key aroma compounds in dog food attractants, focusing on those associated with dogs’ food preferences. They utilized headspace‐solid phase micro‐extraction and GC‐MS techniques to identify VOCs in seven dog food attractants. The study assessed the dogs’ preferences using both acceptance (one‐bowl test) and preference (two‐bowl test) tests with eight beagle dogs. A partial least squares regression (PLSR) analysis determined that 23 out of the 53 identified VOCs were significantly associated with the dogs’ preferences for the attractant samples. Subsequently, three compounds, benzaldehyde, vanillin, and 2,5‐dimethylhydrazine, were added to dry dog foods to validate the PLSR findings. The incorporation of these three VOCs significantly enhanced both the preference and consumption of dog foods. Notably, the consumption ratios of samples treated with 2,5‐dimethylhydrazine and vanillin exceeded 75% (Chen et al., 2017), highlighting the significant role of odors in enhancing the palatability and consumption of dog food.

Yin et al. (2020) also employed GC‐MS techniques to identify key odor compounds and correlated these findings with preference and acceptance test results using PLSR analysis. They found that (E)‐2‐hexenal, 2‐furfurylthiol, and 4‐methyl‐5‐thiazoleethanol increased the palatability of pet food samples for beagles, whereas (E)‐2‐octenal reduced it. Additionally, (E)‐2‐decenal, 2‐furfurylthiol, and 4‐methyl‐5‐thiazoleethanol, characterized as having “roasted,” “meaty,” “nutty,” and “sweet” notes, contributed to an increased intake ratio of the pet food samples.

4.4. Electronic nose or electronic tongue analysis

Monitoring volatile and non‐VOCs related to food quality is crucial in the food industry (Cheli et al., 2017). While gas chromatography (GC) and high‐performance liquid chromatography (HPLC) are powerful analytical tools that provide detailed chemical compositions, they come with significant drawbacks. These techniques require destructive sample preparation, skilled personnel, substantial initial investment and operating costs, and considerable time to deliver results. To overcome these challenges, industries have increasingly adopted sensor‐based technologies such as electronic nose and electronic tongue (e‐tongue) for assessing various aspects of food quality, including shelf life, safety detection, quality control, food adulteration, origin discrimination, and off‐flavor detection (Lu et al., 2022). Unlike GC and HPLC, e‐nose and e‐tongue technologies generally require nondestructive or minimally invasive sample preparation and less technical expertise, while offering rapid or real‐time results through pattern recognition (Wardencki et al., 2013). This makes them valuable tools for continuous monitoring by industry professionals.

While descriptive sensory analysis performed by trained human panelists is highly effective in characterizing all sensory attributes and quantifying their intensities, it has limitations. Descriptive sensory analysis is often time‐consuming, costly, and not feasible for certain samples, such as pet food, that are unsuitable for human consumption. In these cases, e‐nose and e‐tongue, which mimic human olfactory and gustatory systems, have proven useful for providing a holistic sensory profile of test samples in a time‐ and cost‐efficient manner (Cho & Moazzem, 2022; Ross, ²⁰²¹). These instruments follow a sensing, interpreting, and discriminating process, allowing them to identify and differentiate odors and taste compounds (Jung et al., 2017; Tan & Xu, ²⁰²⁰), often yielding results comparable to those from descriptive sensory analysis (Cho & Moazzem, 2022; Ross, ²⁰²¹). Additionally, the e‐tongue can profile taste‐related attributes of pet food that human panelists cannot consume.

Given these advantages, the e‐nose and e‐tongue offer promising alternatives or complementary tools to descriptive sensory analysis, particularly in the pattern recognition and discrimination of pet food samples (Cheli et al., 2017; Jeong et al., ²⁰²³). Cheli et al. (2017) suggested that these technologies can be employed across a wide range of applications in the pet food industry, including assessing compositional patterns, detecting off‐odors indicative of contamination, oxidation, or degradation, evaluating product stability and shelf‐life, profiling flavors for new pet palatants, replacing traditional pet food preference tests, and analyzing the effects of packaging on sensory quality. However, since the e‐nose and e‐tongue are limited to detecting aroma, taste, and flavor compounds, supplementary assessments of appearance and texture‐related attributes are necessary for a comprehensive understanding of pet food samples.

5. CURRENT CHALLENGES AND FUTURE OPPORTUNITIES

Market trends are significantly affecting the innovation of pet food products (Watson et al., 2023). These trends are characterized by the inclusion of raw ingredients and homemade products in pets’ diets (Davies et al., 2019). Notably, there is a growing evolution toward pet foods that more closely resemble human foods (e.g., humanized foods for pets, fresh‐cooked pet foods) (White, 2022), but it is also important to approach these products by prioritizing the nutritional needs and palatability for animals to ensure their well‐being.

Another improvement that the industry has made is that it has developed different types of dry foods, including baked kibbles, air‐dried, dehydrated, and freeze‐dried foods (Watson et al., 2023). Palatability test studies have been mostly conducted with traditional types of pet foods, which include extruded dry food, wet food, and semi‐wet food (Griffin et al., 1984). However, further investigation is necessary regarding the sensory characteristics, acceptance, and preference of various other types of pet foods commercially available in the market (e.g., organic, byproducts of animal products, and raw vs. cooked). Pet food industries are trying to reduce the carbon footprint that these types of products generate. Industries are innovating and looking for alternatives in processing and ingredient selection (Acuff et al., 2021). For future research opportunities, it would be valuable to compare the acceptability of the actual products with the more sustainable ones (Fantechi et al., 2024).

Even though palatability tests for pet foods have been studied for many years, there are still numerous factors that have not been taken into consideration, similar to contextual and multisensory influences on preference for human foods (Seo, 2020; Spence, ²⁰¹⁵). These potential factors include feeding setting (Spence, 2022), the number of pets per household, pet profiles (e.g., breed, sex, and size), pet's familiarity with the test samples, sensory‐specific satiety (Rolls et al., 1981), single versus multiple consumptions (Cosson et al., 2020; Seo et al., ²⁰²⁰), temporal perception and consumption (Visalli et al., 2023), and multisensory flavor perception (Spence, 2022).

While considering various factors influencing palatability of pet foods, it is crucial to standardize a test protocol for measuring palatability of pet foods, which will be advantageous in comparing palatability across different products and test conditions, facilitating quality control and product development of foods. Consequently, the standardization of palatability test protocols can strengthen the reliability of pet food quality evaluations, aiding in ensuring compliance with industry regulations and trading pet food products between different countries.

As previously discussed, these instrumental analyses and rapid assessment tools, such as chromatography, e‐tongue, and e‐nose, are effective in identifying key organic compounds that are indicative of the palatability, quality, and freshness of pet foods (Cheli et al., 2017). In addition to odorous and tasting cues, it is essential to examine the impact of textural characteristics (e.g., hardness or chewiness) on pet food palatability when considering the fact that textural properties significantly affect sensory acceptance and preference for human foods (Choi & Seo, 2023; Luckett et al., ²⁰¹⁶). Instrumental analyses using texture analyzer or rheometer can also be beneficial to objectively measure the textural characteristics that are associated with palatability of pet foods.

Finally, in recent years, the pet food industry has increasingly adopted emerging technologies, such as artificial intelligence, to enhance its operations and product quality. Artificial intelligence is actively being utilized to facilitate interactions between pets and their caregivers, as well as to streamline the feeding process. For example, Nogueira et al. (2019) demonstrated Robot Chow capable of serving food and water at pre‐scheduled times and volumes for pets, thereby monitoring the pet's feeding status and unusual behaviors. Leveraging automated feeding and monitoring solutions enables researchers to track the palatability of target food samples more efficiently and with less time and effort than is required for traditional palatability tests. Related to this, the Internet of Things can be more extensively leveraged to comprehensively understand the palatability of pet foods. Specifically, machine learning models, leveraging datasets that encompass a wide range of parameters including body temperature, heart rate, eating habits, motions, sleep quality, urine pH, and consumption rate, are capable of real‐time estimation of a pet's preference and health status (Ravi & Choi, 2022; Zhang et al., ²⁰²⁴). These models enable researchers and caregivers to customize feeding conditions to meet the unique needs of individual pets. This holistic approach enables researchers to integrate real‐time data from connected sensors into data of pet's preference, consumption rate, and behavior toward specific pet food samples, offering deeper insights into pet's palatability for pet foods. Moreover, similar to advancements in human food products (De Clercq et al., 2016; Zhang et al., ²⁰¹⁹), machine‐learning and deep‐learning techniques can be employed to optimize a combination of multiple ingredients in pet foods, enhancing both sensory pleasure and nutritional value for individual pets. Finally, artificial intelligence technologies have been used to enable researchers and caregivers to identify emotions of pets based on their barking sounds (Patel et al., 2023) and facial images (Sinnott et al., 2021). However, further studies are warranted to enhance the precision in identifying emotions and to expand the range of emotions recognized by artificial intelligence technologies. This advancement will assist in predicting the palatability of pet foods by analyzing evoked emotions, mirroring results demonstrated in forecasting consumer preference for human foods (Pinsuwan et al., 2022; Samant & Seo, ²⁰²⁰).

6. CONCLUSIONS

When pet caregivers purchase food for their pets, their decision‐making process involves multiple factors. While health and nutrition are typically prioritized over sensory qualities, sensory palatability, particularly flavors, often plays an important role in motivating them to switch their pets’ diets. Traditionally, using the one‐bowl and two‐bowl tests, pets’ acceptance and preference for specific foods have been assessed. However, in recent years, new approaches have been developed to incorporate additional factors such as olfactory cues, chemical compositions, pet behavior, and pet's facial expressions, enhancing the insights provided by traditional methods. Researchers have also examined pet caregivers’ perceptions and emotions regarding pet food. Further research utilizing emerging technologies can deepen our understanding of the roles individual sensory components play in the palatability and intake of pet foods. In addition, for enhancing the palatability of pet foods and the overall well‐being of pets, it is important to consider a variety of factors and topics, including innovative processing techniques, the use of sustainable ingredients, complete and balanced nutrition, and holistic approaches to evaluate food palatability and ingestive behavior of pets. Both traditional and contemporary methods are valuable to the pet food industry, providing a better understanding of pets’ reactions and behaviors toward food stimuli, as well as insights into pet caregivers’ preferences. This knowledge ultimately helps make pet food products more appealing to both pets and their caregivers. In addition, establishing a standardized protocol of pet food palatability testing is essential to ensure consistent and reliable results.

AUTHOR CONTRIBUTIONS

Natalia Calderón: Conceptualization; methodology; visualization; writing—original draft. Brittany L. White: Writing—review and editing. Han‐Seok Seo: Writing—original draft; writing—review and editing; conceptualization; methodology; visualization; funding acquisition.

ACKNOLWEDGMENTS

This study was based upon work that is supported, in part, by the United States Department of Agriculture National Institute of Food and Agriculture Hatch Act funding (7001030) to H.‐S.S.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Calderón, N. , White, B. L. , & Seo, H.‐S. (2024). Measuring palatability of pet food products: Sensory components, evaluations, challenges, and opportunities. Journal of Food Science, 89, 8175–8196. 10.1111/1750-3841.17511