Corresponding author HELEN VAIKMA shares the results of a study market mapping the sensory attributes of five different categories of plant-based beverages.[1]

The commercial importance of coffee beverages which include dairy milk or plant-based alternatives is obvious for almost every person working in coffee. Equally obvious to any person working in coffee, no matter where in the world they are based, is the strong trend among consumers to replace dairy milk with plant-based beverages, even if it means paying a bit more. At the same time, coffee professionals are worried about their products’ attributes: they want their coffee beverage to be appealing and with great flavor, nutritious, and sustainable. Unfortunately, coffee professionals currently need to work through all the possible options themselves—if coffee is generally a little-studied product, when it comes to the interactions between coffee and dairy milk, there are very few studies to be found. And, to the best of our knowledge, the interactions between coffee and plant-based beverages have not been studied scientifically at all! This leaves trial and error as the only method available for coffee professionals who need to make decisions about plant-based beverages.

Happily, this situation is starting to change. When we were preparing the contents for Sensory Summit Europe 2022 last September, we wanted to bring a researcher to present about a product other than coffee for the very last session (tasting coffee and ingesting caffeine for two whole days may be a bit too much, even for Sensory Summit participants). Well aware of the increasing importance of plant-based beverages, we thought it would be ideal to include a presentation about them, but we thought it would be a very long shot to find a presenter with research about this topic. We were ecstatic when we found the paper referenced in this article, “Market mapping of plant-based milk alternatives by using sensory (RATA) and GC analysis,” by researchers from the Center of Food and Fermentation Technologies, and the Tallinn University of Technology, both in Tallinn, Estonia. I contacted Sirli Rosenvald, one of the paper’s authors, through a common friend, and invited her to present at Sensory Summit, which she graciously accepted.

The research presented at the Sensory Summit by Sirli, and described here by co-author Helen Vaikma, had almost everything we wanted to be included in that event’s presentation: a very strong sensory characterization, complemented by both chemical and nutritional analyses, of over 100 commercial, plant-based beverages of all kinds. Of particular interest to us was the wide range of sensory attributes presented by this product category, ranging from nutty to dairy to cereal flavor, or from oily to lumpy to watery mouthfeel, for example. Getting acquainted with the style of sensory “flavor map” presented in this paper can be valuable for coffee professionals— similar sensory profiles have been created for coffee in other studies, to characterize the resulting flavor profile of processing methods, regions of origin, varieties, or other treatments. What kind of attributes do you want your beverage to have? Using the map developed by the researchers of this study will help you narrow down the choices you might want to explore.

Reading this study naturally makes me curious about plant-based beverages outside Estonia and if they would behave like their analogous beverages sourced in the Baltic country. It also makes me wonder what I would expect every coffee person might be thinking when they read this study: how do these different plant-based beverages interact with coffee? We can hypothesize some will develop a good texture while others will result in great flavor. Might there be a product that works well on all fronts? That is surely the next assignment for plant-based beverage researchers!

Dr. Mario Fernández-Alduenda
Technical Officer, SCA

The market for plant-based beverages continues to grow rapidly[2] as people look for ways to make more sustainable choices. However, it is also increasingly difficult for consumers to orientate themselves because of the wide variety of raw materials used to make plant-based drinks.

Although the reasons for choosing a plant-based beverage are often linked to sustainability considerations (e.g., environment, health, ethics), taste is the most important aspect when choosing a food product—and both sustainability and taste have considerable variability. Not all products are as sustainable and tasty as people expect![3] Variability in taste can also make consumer choices difficult: each plant protein has its own specific flavor nuances, which can be both a product’s strength and its weakness. Thus, it is important for food developers to understand the properties of the raw materials used, as sensory properties play an important role in the consumption of plant-based beverages.

Collecting the Data

For this study, we applied both sensory and aroma analysis to all the plant-based beverages available in Estonia[4] in order to create an overview of the market in terms of their sensory characteristics. We included drinks that were based on only one raw material, excluding blends and flavored options, for a total of 90 beverages evaluated.

Sensory analyses were carried out in a standardized sensory room at the Centre for Food and Fermentation Technologies (Tallinn, Estonia) with the participation of 10 trained expert assessors, performed across four different sessions, each lasting 40–60 minutes. To avoid palate fatigue during such long sessions, we asked tasters to use a “Rate-All-That-Apply” (RATA) method across 32 terms for appearance, odor, taste, and texture. While evaluating each sample, the assessors were asked to select the most relevant terms and then to indicate the intensity of the attribute on a five-point scale.

To perform aroma analysis, we turned to gas chromatography-mass spectrometry-olfactometry (GC-MS-O). First, samples were put in a vial and heated by the gas chromatograph-mass spectrometer (GC MS), which then detected, measured, and recorded different volatile compounds that were released into the empty “headspace” of the vial once heated. Not all of these volatile compounds are perceivable by humans, however, so we additionally employed a gas chromatograph-olfactometer (GC-O), which similarly heats and releases volatile compounds to be assessed, but this time directly into the nose of a trained assessor, who sniffs and describes the volatile compounds they’re able to perceive. By cross-referencing the results of both the GC-MS and the GC-O, we were able to identify which specific volatile compounds were likely the key aroma compounds in each sample, while also identifying compounds that may induce “off” or unpleasant flavors.

Of the 90 beverages we evaluated, the majority fell into the cereal-/pseudocereal-based and nut-based categories, as shown in Table 1. Most samples in the cereal-/pseudocereal-based category were either oat drinks (21%) or rice drinks (10%), whereas nut-based beverages were mostly almond-based (23%). There was also a moderate group of soy drinks (14%) from the legume-based category.

**Table 1.** Distribution of plant-based beverages in categories according to the raw material.

While to consumers it may seem that there are quite a lot of products available, for the purposes of this study, this sample was rather limited: some categories only contained a few choices (quinoa, brazil nut, buckwheat, and cashew) and, although there were several soy options available, soy was the only legume- based drink available. There was also a limited choice of seed-based beverages, including only one hemp drink. Of course, this was specific to the Estonian market, and other markets may have more variety. However, other studies generally confirm that consumers would like to see more plant-based drink options on the shelves beyond what is already available.[5]

Plant-Based Beverages’ Sensory Attributes

Assessors most frequently referred to samples as having a dark appearance (“darkness” was selected in 81% of evaluations), followed most frequently by sweet taste (77%), and watery texture (64%). Across half of the evaluations, grayness (51%), astringent taste (51%), and sweet odor (50%) were also selected. Assessors used “dairy-like” to describe the taste of only 18% of the evaluations in total, suggesting that plant-based beverages may be at a disadvantage when trying to mimic or replace dairy milk. This “dairy- like” attribute was perceived more intensely in nut- based drinks such as coconut and hazelnut drinks (Figure 1), but it’s possible this may have been related to more than just taste, as these samples had a thicker texture, adding more body to the drink (possibly relating to a more dairy-like mouthfeel).

**Figure 1.** Partial least squares discriminant analysis (PLS- DA) for sensory results of different raw material categories. The different descriptors used by the assessors are the averaged “points” on the graph, with the abbreviated letter in front indicating whether it was related to appearance (A), odor (O), taste (T), or texture (X). The placement of the colored text (abbreviations of the different plant proteins used to create the different beverages) is based on which descriptors were most frequently used in relation to the category (beverage abbreviations, from left to right: HZLN is hazelnut, COCO is coconut, ALM is almond, CSHW is cashew, BRAZ is brazil nut, BCKW is buckwheat, QUIN is quinoa). Larger ovals or circles in the corresponding color indicate a larger number of samples that had some variation in descriptors. For example, the coconut-based samples (COCO) were most consistently recognized as having a nutty texture, nutty odor, and thicker texture. Very generally speaking, the closer the descriptor is to the raw material category, the more it relates to that category as well.

It was clear that the sensory differentiation between the various plant-based beverages was based on the nuances of the raw materials: cereal- or pseudocereal-based drinks were perceived as having a cereal flavor, nut-based drinks as “nutty,” and soy (a legume) as having a legume flavor (see Figure 1). Some specific aroma compounds commonly found in some plants were also evident in this study. However, many factors affect the overall sensory profile of each beverage. For example, GC-MS-O detected high levels of sulfur compounds in oat, soy, and hemp samples. Previous research indicates that these compounds may be related to the heat treatment processes. This is important to note because sulfur compounds may have caused off-flavors in the product—like bitterness, astringency, and a more intense aftertaste—as well as enhanced raw material nuances, such as graininess in oat drinks or earthiness in soy drinks. There were also other compounds found in the samples, which previous literature suggests are associated with increased heat. In some cases, these compounds may induce non-characteristic attributes: for example, leguminous flavor (Figure 1) in rice drinks may have been linked to the decanal (an aldehyde compound with waxy and fatty odor nuances), possibly related to prior heat treatment.

Soy drinks were one of the most interesting examples from this study. On one hand, the market situation shows that this is still a very popular category. After all, one of the strengths of this category is its long history and availability on the shelves, but also its protein- rich (Figure 2) composition. But on the other hand, soy drinks can often have a specific leguminous flavor and a distinctive off-flavor (i.e., metallic, astringent, earthy). For example, dimethyl sulfide previously associated with heat treatment may cause earthy nuances in soy. Furthermore, the aroma analysis showed that vanillin was the most abundant in soy samples compared to other beverages. This may suggest that manufacturers are aware of the sensory deficiencies of their product and try to mask them with vanillin, even in unflavored products. The average consumer may not notice these off-flavors, especially if they’ve been masked somehow, but vanillin may still affect the overall taste (including when used in coffee as a milk alternative!).

**Figure 2.** Nutritional information of the included samples according to label on the packaging. The size of the bubble illustrates the differences in the amount of calories (kcal/100g).

However, aroma compounds produced during processing are not always a bad thing. In hazelnut samples, for example, the assessors noted nutty and roasted notes. The aroma analysis showed that their flavor profile was associated with different pyrazines, which are also related to the roasting process and are sensorially recognized as nutty and roasty aromas. As coffee is also generally a heat-treated product, such changes in the profile of plant-based beverages may be critical. The desirability of such changes depends on whether the aroma profile is appealing to the consumer or not. This leads back to the well-known question of whether plant-based beverages need to be more neutral to mimic milk, or whether they can taste good on their own.

Impact of Nutrients on Texture

Although not included in the original paper, we also analyzed the nutritional information on the packaging (Figure 2). Soy-based drinks were the most like regular cows’ milk in terms of nutrients, due to their higher protein and lower carbohydrate content (and making it beneficial for use in coffee drinks, as proteins help to stabilize the foam!). The protein content of other plant-based beverages was much lower, and as you might expect, cereal- and pseudocereal-based samples had higher carbohydrate content.

It is also important to stress the importance of texture, which is related to fat content—at the very beginning, I mentioned that assessors selected “watery texture” in 64% of the evaluations! Texture influences the sensory experience in multiple ways. First, it affects the mouthfeel: a fattier and thicker drink helps to create a fuller and creamier sensation, which is a particularly important aspect for barista milks. Second, it affects the release of flavor compounds during consumption, so it also influences the final odor and taste characteristics of the product. Nevertheless, fat is another complex issue for plant-based beverages, as much depends on the composition of the raw material itself: for example, while nut- and seed-based drinks were between 2% and 3%, cereals and pseudocereals are more likely to be around 1% fat content. The impact of fat content on perceived sensory characteristics is one of the reasons why fats are often added to plant-based beverages during the production process, although it’s important to keep in mind that the addition of fat also influences other sensory characteristics (including possible off-notes).

Looking to the Future

In the two years since we first completed this extensive research into the sensory attributes of plant-based beverages, more dairy-milk alternatives have become available on the shelves of Estonian supermarkets. For instance, pea-based drinks, an addition to the legume category, are now available. However, diversity does not automatically mean a wider choice of products that consumers are likely to enjoy.

As this category continues to grow, producers of plant-based beverages should consider the sensory variability of different raw materials as well as the conditions in their production process—we were surprised to find so many options with “off-flavors”—and we hope this research helps articulate how some compounds can be linked to certain conditions (and what can be done to improve or avoid this!). For example, some specific compounds which induce earthy and hay-like nuances may be linked to fat oxidation, which can be addressed through ensuring good preservation conditions. We also hope coffee shops and consumers will find this information thought-provoking. Attempting to understand how the sensory profile of each plant protein changes with different treatments, as well as during the product’s shelf-life, is key to having the product you’re most likely to enjoy when combined with the coffee in your cup. ◇

HELEN VAIKMA is a sensory scientist researching at the Center of Food and Fermentation Technologies in Estonia. She is currently also a PhD student at Tallinn University of Technology, focusing on consumer perception of plant-based alternatives.

References

[1] For this study, "Market Mapping of Plant-Based Milk Alternatives by Using Sensory (RATA) and GC Analysis," the researchers used the categorization put forth by Sethi, Tyagi, and Anurag in 2016, which splits plant milks into five groups according to the raw materials used: cereal-based, pseudocereal-based (i.e., amaranth, quinoa, and buckwheat), legume- based, nut-based, and seed-based. (Swati Sethi, S. K. Tyagi, and Rahul K. Anurag, “Plant-Based Milk Alternatives an Emerging Segment of Functional Beverages: A Review,” Journal of Food Science and Technology 53, no. 9 (2016): 3408–3423, https://doi.org/10.1007/s13197-016-2328-3.

[2] Mordor Intelligence, “Dairy Alternative Products Market – Growth, Trends, COVID-19 Impacts, and Forecasts

(2023–2028),” (2021), https://www.mordorintelligence.com/ industry-reports/dairy-alternatives-market.

[3] Some of the raw materials used to produce plant-based milk alternatives are also associated with environmental problems, such as deforestation or bee mortality, while others can be linked to health disadvantages, such as lower nutritional value or capacity to cause an allergic reaction.

[4] All 90 beverages evaluated were sourced in Estonia in January 2020.

[5] ProVeg International, European Consumer Survey on Plant-Based Foods – Describing the Product Landscape and Uncovering Priorities for Product Development and Improvement (Berlin, 2020).

We hope you are as excited as we are about the release of 25, Issue 19. This issue of 25 is made possible with the contributions of specialty coffee businesses who support the activities of the Specialty Coffee Association through its underwriting and sponsorship programs. Learn more about our underwriters here.

Mapping the Market: An Overview of Plant-Based Beverages’ Sensory Attributes | 25, Issue 19