Coffee consumers and preference
Who likes which coffee and why – how liking is measured, why consumers fall into preference groups rather than agreeing on one best cup, and how expertise, culture, biology and labels shift what people prefer and pay.
Across the chain: Science
In brief
- In consumer tests of brewed coffee, people split into preference segments. In a 2019 study, each of four segments liked three of four coffees and disliked a different fourth.7
- In a 2021 drip-coffee study, two consumer segments differed most in the brew strength they liked; the larger segment held 57% of consumers.5,6
- In a blind test, 205 naïve consumers split evenly between a high- and a low-graded coffee and could not pick the more expensive one better than chance.8
- Labels can change liking and willingness to pay: in a 2013 study, identical coffee called “eco-friendly” was preferred by consumers with strong sustainability attitudes.14
- The Specialty Coffee Association (SCA) lets a cupper score impression of quality by their own preference or by a known market preference.1
Contents
No single best coffee#
People do not agree on what good coffee tastes like. A 2018 review of consumer research on coffee grouped the reasons into five kinds of factor.3 They are personal preferences, economic attributes, product attributes, the context of drinking and socio-demographics.3
According to the SCA, some of these preferences are natural, such as an aversion to extreme bitterness or spoiled aromas.1 Others are cultural: learned ideas about what coffee should taste like.1 The SCA therefore expects judgements of quality to be somewhat diverse, even among professionals.1
How liking is measured#
Consumer scientists measure liking with affective tests. The standard tool is the 9-point hedonic scale.2 It was developed for the US armed forces by Peryam and Pilgrim and published in 1957.2 A 2023 study calls it the most widely used scale for consumer acceptance testing.2
Two further tools explain the scores. Just-about-right (JAR) questions ask whether a property, such as strength, is too weak, too strong or right.7 Preference mapping relates each consumer’s liking to the measured flavour of the samples.9,11 The methods themselves are described under sensory science methods.
The SCA uses the same 9-point format on its cupping form. Its labels, however, ask for an impression of quality, not for liking.1
Preference segments#
When consumers rate several coffees, their scores rarely line up. Statistical clustering sorts them into preference segments: groups whose members like similar coffees.5,7 A 1992 chapter on sensory segmentation described the same pattern for food in general. How liking changes with intensity differs from person to person.4
Coffee studies found such segments. In a 2019 study, 85 consumers rated four drip coffees and fell into four clusters.7 Each cluster liked three of the coffees and disliked the fourth, but each disliked a different one.7 In a 2021 study with 118 consumers, two segments emerged. They differed most in the brew strength they liked.5
Data table
| % | Source | |
|---|---|---|
| Basket study 2019, cluster 1 (n = 16) | 18.8% | 7 |
| Basket study 2019, cluster 2 (n = 25) | 29.4% | 7 |
| Basket study 2019, cluster 3 (n = 25) | 29.4% | 7 |
| Basket study 2019, cluster 4 (n = 19) | 22.4% | 7 |
| Strength study 2021, segment 1 (low strength liked best) | 57% | 6 |
| Strength study 2021, segment 2 (two favoured types) | 43% | 6 |
The authors of the 2021 study concluded that the “ideal” zone of the brewing control chart should be reconsidered to reflect these segments.5 How this changed the brewing control chart is explained in that entry.
Experts and consumers#
Expert grades and consumer choices do not always agree. In a 2016 blind test, 205 naïve consumers tasted two coffees that differed widely in expert grade and price.8 Their preferences split evenly, and they picked the expensive coffee no more often than chance.8 The authors suggest that expert ratings may not align with how consumers perceive quality.8
Professionals disagree among themselves as well. According to the SCA, its research for the Coffee Value Assessment (CVA) found distinct preference clusters among the participating cuppers.1 The SCA’s answer is to let a cupper score by their own preference or by a known market preference.1 A cupper may dislike earthy coffees, for example, while knowing that some consumer segments prize them.1
Culture, habit and biology#
What drives liking differs between groups of drinkers. Among 85 mostly young consumers in a University of California, Davis study, liking rose with floral, sweet and berry notes. It fell with bitter, burnt, rubbery and musty notes.7 Among 100 black-coffee drinkers in Thailand, roasted and ashy notes and bitterness went with higher liking. Sour, fruity and musty notes went with lower liking.10
The two studies used different coffees, so they do not prove a cultural difference. They do show that one list of desirable notes may not fit every market. In a Thai study of instant coffee mixes, caramel aroma and sweetness were drivers of liking.11 Words for flavour also differ between cultures; flavour language covers this.
Biology plays a part. In a 2018 genetic study, people predisposed to taste caffeine as more bitter drank more coffee.12 In a 2015 study, people less sensitive to the bitter compound PROP added more sugar to coffee.13 Bitterness in coffee treats these findings in detail.
Labels, price and setting#
What people know about a coffee changes how they rate it. In a 2013 study, participants were told that one of two cups held “eco-friendly” coffee. Both held identical coffee.14 Consumers with strong sustainability attitudes preferred the labelled cup and would pay more for it.14
Label effects can fade. In a 2024 study, information about bird-friendly production raised willingness to pay for the first coffee served.15 For the second coffee, its net effect was zero.15 A 2021 meta-analysis of 22 studies found that consumers were willing to pay more for organic, origin-labelled and Fairtrade coffee.16 Organic labels had the largest effect.16 What these labels require is described under certifications and labels.
The setting matters too. Cups, light, sound and the room can shift ratings of the same coffee; these effects are covered under crossmodal effects.
Reference values#
| Study | Consumers | Coffees | Method | Main result |
|---|---|---|---|---|
| Frost et al. 20197 | 85; 80.0% aged 18–24 | Four drip brews: light or dark roast (two different blends) × two basket shapes | 9-point hedonic, JAR, CATA; Ward’s clustering | Four clusters of 16, 25, 25 and 19 consumers; each disliked a different coffee |
| Cotter et al. 20215 | 118, mostly college-age | 27 drip brews of one medium-roast coffee varying in TDS, extraction and brew temperature6 | Hedonic, JAR, CATA; cluster analysis | Two segments differing most with TDS; flavour intensity and acidity drove liking in both |
| Guinard et al. 20236 | Segments of the 2021 study | Synthesis of the UC Davis drip studies | Response surfaces by segment | 57% and 43% of consumers in the two segments |
| Giacalone et al. 20168 | 205 naïve consumers | Two coffees differing widely in expert grade and price | Blind paired preference, then identification of the expensive cup | Preferences split evenly; identification at chance level |
| Giacalone et al. 20199 | 83 consumers, Copenhagen, Denmark | One washed Kenyan arabica in six roast profiles | 9-point hedonic, CATA; internal preference map | Mean liking: normal 6.0, baked 5.7, scorched 5.6, underdeveloped 5.4, dark 5.1, light 4.2; only baked not significantly below normal |
| Pinsuwan et al. 202210 | 100 black-coffee drinkers, Thailand | Eight commercial arabica coffees, light to dark | Hedonic, purchase intent, emotions; 8 trained panellists | Liking tended to rise with roast degree |
| Vittayaporn and Wirunthanakrit 201611 | Consumers in Thailand | Instant coffee mixes varying in coffee, sugar and creamer | Acceptance test; preference mapping | Caramel aroma and sweetness drove liking; bitter aftertaste drove disliking |
| Sörqvist et al. 201314 | Three experiments | Two identical coffees, one labelled “eco-friendly” | Preference and willingness to pay | Label raised preference and payment among consumers with strong sustainability attitudes |
| Grunert et al. 202415 | Tasting with or without bird-friendly information | Bird-friendly coffees; a first and a second coffee served | Becker-DeGroot-Marschak auction; questionnaires; facial expressions | Information raised willingness to pay for the first coffee served; net zero effect for the second |
The table lists the consumer studies consulted for this entry. Sample sizes are those reported in each source.
Segmentation in detail#
The 2019 basket study clustered 85 consumers by their overall liking of four coffees.7 Cluster 1 (16 consumers, 18.8%; shares are our calculation from the cluster sizes) disliked the dark roast from the semi-conical basket; most of its members found the roast flavour too strong.7 Cluster 2 (25, 29.4%) slightly disliked the light roast from the same basket.7 Cluster 3 (25, 29.4%) disliked the dark roast from the flat-bottomed basket.7 Cluster 4 (19, 22.4%) disliked the light roast from the flat-bottomed basket, finding its roast and coffee flavour too weak.7
The clusters also differed in habits. Cluster 4 had the largest share of consumers who add nothing to their coffee (52.6%).7 Cluster 3 had the largest share of drip and espresso drinkers.7
The 2021 study varied total dissolved solids (TDS), extraction yield and brew temperature.5 Liking formed a dome-shaped surface for one segment and a saddle-shaped surface for the other.5 In the 2023 synthesis, the majority segment (57%) liked coffees in the lower left of the brewing control chart best, at low strength.6 The other segment (43%) liked two different types equally. One had low extraction and medium to high TDS; the other had high extraction and medium TDS.6
The 1992 chapter described two forms of individual difference. People can differ in their overall level of liking, or in the intensity at which their liking peaks.4 It argued that such differences are not simply artefacts of the test room, pointing to supermarkets that sell one product in several flavours and intensities.4 In our reading, the coffee segments above fit the second form: segments peaked at different strengths.6,7
Segment results depend on the sample. The consumers in the UC Davis studies were young: 80.0% were aged 18–24 in the 2019 study, and most were college-age in the 2021 study.5,7 In the 2019 Danish roast study, liking of the dark and light roasts was bimodal, with a sizeable minority preferring them.9 The authors did not segment the 83 consumers because the sample was too small for a robust segmentation.9
Drivers of liking#
A driver of liking is a sensory attribute whose presence goes with higher or lower liking across consumers. In the basket study, liking rose with floral, sweetness, berry and dried fruit.7 It fell with rubber, burnt wood or ash, musty or dusty, and bitter.7 Bitter was ticked by 51.5% of consumers and burnt wood or ash by 37.1%; together they were the strongest negative drivers.7 Overall liking and flavour liking correlated closely (r = 0.90).7
In the 2021 study, overall flavour intensity and acidity drove liking in both segments.5 In the Thai study, the positive drivers included coffee identity, roasted, ashy, pipe tobacco, bitter taste, rubber, overall sweet, balanced, fullness and longevity.10 The negative drivers included sour aromatic, sour taste, fruity, woody, musty and molasses.10 Rubber and bitter thus worked in opposite directions in the two studies.7,10
The roast-defect study gives a third view. The normal roast of one Kenyan coffee was liked most, and consumers described it as the most balanced.9 All six roasts lay between 4.2 and 6.0 on the 9-point scale.9 The defects themselves are described under roast defects.
Data table
| points on the 9-point hedonic scale | Source | |
|---|---|---|
| Normal | 6.0 | 9 |
| Baked | 5.7 | 9 |
| Scorched | 5.6 | 9 |
| Underdeveloped | 5.4 | 9 |
| Dark | 5.1 | 9 |
| Light | 4.2 | 9 |
Expert and consumer judgement#
The 2016 test first asked 205 consumers which of two blind cups they preferred.8 They were then told that one cup was much more expensive and asked to identify it.8 Consumers who identified it correctly were more likely to prefer the high-quality sample; the authors call for further studies to confirm this link.8 They suggest that in the marketplace, extrinsic cues such as brand, label and price may count for more than intrinsic quality.8
The CVA defines impression of quality as a taster’s opinion of the distinctiveness and desirability of an attribute, reflecting their own or a known market preference.1 The SCA calls the ability to recognise what others value “inter-subjectivity”.1 It notes that the preference of a cupper whose scores do not conform to the group is not wrong.1 No study consulted for this entry tested whether CVA scores predict consumer liking.
Information and willingness to pay#
Willingness to pay (WTP) is the highest price a consumer would pay for a product. The 2024 bird-friendly study measured it with the Becker-DeGroot-Marschak procedure.15 For the first coffee served, information raised WTP directly and through better taste perception, which in turn depended on affect and moral satisfaction.15 For the second coffee, a negative taste inference from the same information left a net zero effect on WTP.15
The 2021 meta-analysis combined 97 WTP estimates from 22 primary studies.16 WTP for organic, country-of-origin and Fairtrade labels was positive and significant, with organic the largest.16 The authors suggest that choice experiments may reduce hypothetical bias, the gap between stated and real payment, more than other stated-preference methods.16 Estimates differed between regions.16
In the 2013 eco-label study, attitudes towards sustainable consumption predicted the bias in one experiment. In another, a general belief that eco-labelled products taste better appeared to predict it more strongly.14 According to the SCA, research on buyer behaviour and consumer willingness to pay shows that extrinsic attributes influence purchasing decisions.1 How information shifts perceived flavour is covered under crossmodal effects.
Individual differences in perception#
The 2015 taste-sensitivity study grouped participants by the density of their fungiform papillae and by PROP taster status.13 Participants with more papillae liked sweetened coffee more and added more sugar.13 PROP super-tasters rated sourness, bitterness and astringency as stronger than non-tasters, in coffee and in standard solutions.13 Coffee liking itself was unrelated to PROP status.13
The 2018 UK Biobank study used gene variants as natural experiments. Higher perceived intensity of caffeine went with a higher risk of drinking more than 4 cups a day (odds ratio 1.207).12 Higher perception of PROP and quinine went with lower coffee intake.12 The genetics are explained under how we taste coffee and bitterness in coffee.
Open questions#
- Segments beyond one population. The coffee segmentation studies consulted for this entry came from one research group at the University of California, Davis, and drew mostly young consumers.5,7 No study consulted tested the same segments in other countries.
- Cupping scores and consumer liking. No study consulted for this entry compared CVA scores with consumer liking of the same coffees.
- Cross-cultural tests. No study consulted for this entry served the same coffees to consumers in two or more countries.
- Durable label effects. In the one study consulted that served a second coffee, the label had no net effect on WTP for it.15
- Sweetened and instant coffee. Consumer research on sweetened and instant coffee mixes was scarce among the sources consulted; one Thai study was found.11
Key studies#
Frost, Ristenpart and Guinard, 2019 – Four preference clusters. Eighty-five consumers rated four drip coffees that differed in roast and brewer basket. Cluster analysis found four groups of similar size, each accepting three coffees and rejecting a different fourth; bitter and burnt notes were the strongest negative drivers overall.
Cotter et al., 2021 – Preferences across strength and extraction. Asked 118 black-coffee drinkers to rate 27 brews of one coffee. Two segments emerged whose liking depended mainly on TDS, and the authors argue that the “ideal” of the brewing control chart should be reconsidered to reflect them.
Giacalone et al., 2016 – Expert quality and consumer choice. In a blind test, 205 naïve consumers chose between two coffees that differed widely in expert grade and price. Preferences split evenly, and consumers did no better than chance at identifying the expensive one.
Pinsuwan et al., 2022 – Drivers of liking in Thailand. One hundred Thai black-coffee drinkers rated eight commercial coffees profiled by a trained panel. Roasted, ashy and bitter notes went with higher liking, sour and fruity notes with lower liking – a pattern different from studies with young US consumers.
Abdu and Mutuku, 2021 – Paying for labels. A meta-analysis of 97 willingness-to-pay estimates from 22 studies of coffee ecolabels. Willingness to pay was positive for organic, origin and Fairtrade labels, organic most; the authors suggest choice experiments may reduce hypothetical bias.
- Roast degree and consumer liking: in a 2022 study with 100 Thai consumers of black coffee, liking tended to rise from light to dark roasts10; in a 2019 Danish study with 83 consumers, the dark and light roasts of one Kenyan coffee were liked less than its normal roast.9 The studies used different coffees, roasts and consumers, so they do not test the same question.
Coffearium (2026). Coffee consumers and preference. Coffearium – The open coffee compendium. Retrieved 17 November 2026, from https://coffearium.org/entries/coffee-consumers-and-preference/
@misc{coffearium2026-coffee-consumers-and-preference,
author = {{Coffearium}},
title = {{Coffee consumers and preference}},
year = {2026},
howpublished = {Coffearium -- The open coffee compendium},
url = {https://coffearium.org/entries/coffee-consumers-and-preference/},
urldate = {2026-11-17},
note = {Retrieved 17 November 2026}
}Entries are updated when the evidence changes; the retrieval date tells your readers which version you used. Changes are listed in the Change log.
Change log
- – Softened three statements to the sources' own strength: expert ratings “may not align” with consumer perception (not “say little”); choice experiments “may reduce” hypothetical bias; the 2021 authors call for reconsidering the control chart's “ideal”, not a statement that no ideal fits all. Reason: The abstracts of Giacalone et al. 2016, Abdu and Mutuku 2021 and Cotter et al. 2021 say less than the earlier wording.
- – The bird-friendly study is described by first and second coffee served, not by first and second tasting; the second result is a net zero effect on willingness to pay. Reason: Wording of Grunert et al. 2024.
- – The SCA cupper-preference finding is attributed to the SCA without the year and the number of clusters; Thai instant-coffee study given without sample counts. Reason: Number rule; both rest on a single Tier B source.
- – Segment shares of the 2019 basket study marked as our calculation from the cluster sizes; all six roast-liking means added to the reference table; Danish consumers described as consumers in Denmark; cluster 1 reason narrowed to roast flavour. Reason: The paper gives counts only; chart values must appear in the text; the roast-defect paper names the University of Copenhagen, not the consumers' nationality.
Sources
- A/BSpecialty Coffee Association (writing team Fernández-Alduenda, M.; Ionescu, K. E.; Giuliano, P.; Rugolo, J.; Delrue, R.) (2024). A System to Assess Coffee Value: Understanding the Specialty Coffee Association’s Coffee Value Assessment (June 2024). Specialty Coffee Association.
- AAddo-Preko, E.; Amissah, J. G. N.; Adjei, M. Y. B. (2023). The relevance of the number of categories in the hedonic scale to the Ghanaian consumer in acceptance testing. Frontiers in Food Science and Technology 3, 1071216. doi:10.3389/frfst.2023.1071216 · other copies
- ASamoggia, A.; Riedel, B. (2018). Coffee consumption and purchasing behavior review: Insights for further research. Appetite 129, 70–81. doi:10.1016/j.appet.2018.07.002 · other copies
- BMoskowitz, H. R. (1992). Individual Differences in Preferences: Sensory Segmentation as an Organizing Principle. In: Doty, R. L.; Müller-Schwarze, D. (eds.), Chemical Signals in Vertebrates 6, Springer, Boston, pp. 571–574. doi:10.1007/978-1-4757-9655-1_87 · other copies
- ACotter, A. R.; Batali, M. E.; Ristenpart, W. D.; Guinard, J.-X. (2021). Consumer preferences for black coffee are spread over a wide range of brew strengths and extraction yields. Journal of Food Science 86(1), 194–205. doi:10.1111/1750-3841.15561 · other copies
- AGuinard, J.-X.; Frost, S.; Batali, M.; Cotter, A.; Lim, L. X.; Ristenpart, W. D. (2023). A new Coffee Brewing Control Chart relating sensory properties and consumer liking to brew strength, extraction yield, and brew ratio. Journal of Food Science 88(5), 2168–2177. doi:10.1111/1750-3841.16531 · other copies
- AFrost, S. C.; Ristenpart, W. D.; Guinard, J.-X. (2019). Effect of Basket Geometry on the Sensory Quality and Consumer Acceptance of Drip Brewed Coffee. Journal of Food Science 84(8), 2297–2312. doi:10.1111/1750-3841.14696 · other copies
- AGiacalone, D.; Fosgaard, T. R.; Steen, I.; Münchow, M. (2016). Quality does not sell itself: Divergence between “objective” product quality and preference for coffee in naïve consumers. British Food Journal 118(10), 2462–2474. doi:10.1108/BFJ-03-2016-0127 · other copies
- AGiacalone, D.; Degn, T. K.; Yang, N.; Liu, C.; Fisk, I.; Münchow, M. (2019). Common roasting defects in coffee: Aroma composition, sensory characterization and consumer perception. Food Quality and Preference 71, 463–474. doi:10.1016/j.foodqual.2018.03.009 · other copies
- APinsuwan, A.; Suwonsichon, S.; Chompreeda, P.; Prinyawiwatkul, W. (2022). Sensory Drivers of Consumer Acceptance, Purchase Intent and Emotions toward Brewed Black Coffee. Foods 11(2), 180. doi:10.3390/foods11020180 · other copies
- BVittayaporn, V.; Wirunthanakrit, W. (2016). Preference Mapping of Commercial Instant Coffee Mix in Thailand. Journal of Food Health and Bioenvironmental Science 9(2), 177–191.
- AOng, J.-S.; Hwang, D. L.-D.; Zhong, V. W.; An, J.; Gharahkhani, P.; Breslin, P. A. S.; Wright, M. J.; Lawlor, D. A.; Whitfield, J.; MacGregor, S.; Martin, N. G.; Cornelis, M. C. (2018). Understanding the role of bitter taste perception in coffee, tea and alcohol consumption through Mendelian randomization. Scientific Reports 8, 16414. doi:10.1038/s41598-018-34713-z · other copies
- AMasi, C.; Dinnella, C.; Monteleone, E.; Prescott, J. (2015). The impact of individual variations in taste sensitivity on coffee perceptions and preferences. Physiology & Behavior 138, 219–226. doi:10.1016/j.physbeh.2014.10.031 · other copies
- ASörqvist, P.; Hedblom, D.; Holmgren, M.; Haga, A.; Langeborg, L.; Nöstl, A.; Kågström, J. (2013). Who Needs Cream and Sugar When There Is Eco-Labeling? Taste and Willingness to Pay for “Eco-Friendly” Coffee. PLOS ONE 8(12), e80719. doi:10.1371/journal.pone.0080719 · other copies
- AGrunert, K. G.; Seo, H.-S.; Fang, D.; Hogan, V. J.; Nayga, R. M. (2024). Sustainability information, taste perception and willingness to pay: The case of bird-friendly coffee. Food Quality and Preference 115, 105124. doi:10.1016/j.foodqual.2024.105124 · other copies
- AAbdu, N.; Mutuku, J. (2021). Willingness to pay for socially responsible products: A meta-analysis of coffee ecolabelling. Heliyon 7(6), e07043. doi:10.1016/j.heliyon.2021.e07043 · other copies
Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.
The literature for this entry was last searched on 12 Nov 2026. A statement that no source consulted for this entry shows something refers to the sources above; newer or missing work may exist.
Discussion
A space for readers to share experience from the farm, the roastery or the bar, ask each other and offer other views. The Coffearium team does not reply here. Please keep to the community rules; to change the entry itself, use Suggest a change.