Dr. Carla L. Kuesten Fast Moving Consumer Goods, USA
Exploring Sensory Brand Identity for Retail Coffees Abstract
Attribute Scales
Sensory benchmarking was conducted using 40 unflavored, U.S. retail coffees representing ground, whole b ean, and instant coffees coffees from th e top 10 leading brands of each category.The coffees covered a range in degree-of-roast—light to dark and included both caffeinated and decaffeinated varieties. Principal Component Analysis was applied to the d ata as well as linear Discriminant Discriminant Analysis to determine whether the sensory attribute ratings could be u sed to correctly identify the specific brand and type of coffee and the presence or absence of unique brand profiles.
Attributes are defined below:
Introduction The connection between sensory characteristics and brands of retail coffees was explored. Brand identity, a company’s competitive advantage and strategic asset, encompasses many factors—brand name, image, associations, personality, logo, symbols, colors, jingles, etc.—that portray the perception of the brand and what the brand stands for to the consumer. The sensory profile has a strong influence on how the brand is perceived and is often used to communicate the brand identity and position. Successful companies establish a sensory brand identity.
Materials and Methods Sample Preparation. The coffees were pur chased chased at retail and prepared as follow s: • All ground and whole bean coffee samples were standardized using the same amount of coffee to water ratio in the same Bunn ® pour-over machine. • Whole bean bean coffees coffees were ground using a spinning blade type grinder for a specified specified length of time. • All instant coffee coffee samples were standardize standardized d using the same coffee to water ratio. • All coffees coffees were made using Poland Poland Spring ® water. • Coffees Coffees were presented presented in preheated preheated cups.
Color is Color is a measure of the visual gauge of a coffee’s color intensity ranging from light to dark as perceived by the panelist.
Sour Aromatics measures Aromatics measures sour aromatics such as vegetable-like, fermented, or fatty acid sour. The The scale ranges from non e to strong.
Cup Aroma is a measure of the intensity of the aromatics present in the headspace of the cup using a scale of none to strong. Panelists Panelists describe the type of aromatics in their comments.
Sweet intensity Sweet intensity is a measure of the level of sweet b asic taste.The reference standards are sucrose solutions. The scale ranges from none to strong.
Balance is Balance is a measure of the blend of the product. It takes into account the basic tastes, aromatics, and mouthfeels. The scale ranges from unblended to blended. Fullness measures Fullness measures the richness or body of flavor when tasting the product. The scale ranges from thin to full.
Mou thfeel is is a measure of perceived amount of trigeminal effects such such as astringency o r tannins usually associated with coffee. The scale ranges from none to strong.
Other Coffee Aromatics is Aromatics is a measure of the strength and type of other coffee-related characteristics such as grainy, nutty, brothy (hydrolyzed vegetable protein) and burnt or green bean aromatics. The scale ranges from none to strong.
Others is Others is a measure of basic tastes or arom atics (often off-flavors) off-flavors) that are not measured in any of the previous attributes. Panelists score this none to strong and describe their notes w hich can include Wet Ash Tray(WAT), Tray(WAT), woody, m usty, Dirty Dishrag sour (DDR), (DDR), and rubbery sulfide.
Degree of Roast is a measure of the perceived amount of roasting of the coffee character, from none to dark (espresso-like). Sweet A romati cs cs measures measures sweet aromatics such as vanilla/vanillin, burnt sweet, and caramelized sweet. The scale ranges from none to strong.
• Coffee Coffee samples samples were randomized and and coded for blind evaluation by the panelists. panelists. • Panelists tasted tasted the same amount amount of each sample sample and evaluated the specified attributes at their appropriate temperatures, aroma between 150º–160ºF, and flavo r bet ween 140º–150ºF.
Aftertaste measures the intensity Aftertaste measures of the basic tastes, aromatics or mouthfeels still present in the mouth one minute after the last swallow. Panelists score score this none to strong and describe the characteristics characteristics in their comments.
The flavor attribute data were summarized summarized into two flavor indices using Principal Component Analysis (PCA).
“Coffee Identity”
Low er er sc sc or or es es ar ar e. e. ....
H ig igh er er sc sc or or es es ar ar e. e. ....
Less degree of roast Less coffee aromatics Thinner Less bitter Less aftertaste Less cup aroma Less blended
More degree of roast More coffee aromatics Fuller More bitter More aftertaste More cup aroma More blended
More others/off-notes More other coffee aromatics (i.e. grain, nuts, burnt or green beans)
Index 2 “Other Characteristics”
• Brands Brands are identif identified ied by color with individual individual products represented by form and caffeine level. • The brand brand mean is represented represented by a diamond, color coded by degree of roast.
• Remaining emaining brands could could improve sensory sensory brand identity by delivering pr oducts that are more simi lar across different forms and caffeine levels.
• The ellipses ellipses are centere centered d at the brand mean and represent represent the average brand wi thin produ ct variation.
• Some brands brands were notably notably less consis consistent tent than other brands; some products could be confused confused with neighboring brands.
• Coffees Coffees demonstrating a strong strong sensory brand brand identity will fall w ithin the brand ellipse.
• Decaffeina Decaffeination tion may change change products enough to be noticeably different from the rest of the products within a brand.
Coffee Benchmarking Discriminant Analysis Map
I
2 t n a n i m i r c s i D r a e n i L
J
Type
Attributes
Descriptors Recorded
Appearance Color Aromatic
Flavor
Cup Arom a
Scale 1 2 3 4 5 6 7
Less others/off-notes Less other coffee aromatics (i.e. grain, nuts, burnt or green beans)
The “Coffee Identity” i ndex includes attributes that are closely associated with coffee, i.e., Degree of Roast, Coffee Aromatics, Fullness, Bitterness, and Aftertaste.
Ligh t
Dark
N one
Strong
Balance Unblended Fullness Thin Coffee Aromatics None Other Coffee Aromatics None Degree of Roast None SweetAromatics None SourAromatics None Sweet None Sour None Bitter None Mouthfeel None Others None Aftertaste None
Blended Full Strong Strong Dark Strong Strong Strong Strong Strong Strong Strong Strong
The “Other Characteristics” Ind ex includes off-notes and other coffee arom atics (such as grain, nuts and green or burnt beans)
F ig ig . 2 . P CA CA I n di di c es es
C
The 33 retail coffees selected for this study show a wide range of flavor characteristics.
15 Acorn Park
6 33
K H
29
10
21
0.84
Least Significant Difference
3.14 20
25
30
Lessdegree of roast More degree of roast Lesscoffeearomaticintensity More coffee aromaticintensity Thinner Fuller Lessbitter More bitter Flavor Index 1 lessaftertaste More aftertaste “Coffee Identity” Lesscup aroma More cup aroma Lessblended More blended KEY: Grd = Ground WB = Whole Bean Ins = Instant
Decaffeinated - empty Caffeinated - filled
P CA CA F l av av or or M a p
Contact: Dr. Carla Kuesten
Light Roast Light - Medium Roast Medium - Dark Roast Dark Roast
Ground Instant
• Manufactur Manufacturers ers that that produce more than than one brand may intentionally design their brands to span different regions of the map, allowing them to reach consumer taste segments with targeted, well-defined sensory brand identities.
Linear Discriminant 1
Fig. 4.
Linear Discriminant Analysis Bi-plot Bi-plot
Results of the discriminant analysis shows: • The sensory sensory attributes attributes found to be most important in predicting the coffee brand and type include cup aroma, coffee intensity and type, and sour aromatics.
• The white spac space e within the map suggests suggests opportunities between light and d ark roast roast coffees coffees to establish additional products w ith a unique sensory profile.
Conclusions Sensory benchmarking using a trained panel with principal component and discriminant analyses analyses can be used for guidance in selecting and maintaining a unique sensory brand identity in the marketplace.
• This confusion is most readily readily evident in the lighter roasted coffees which deliver less flavor intensity overall. Attribute plots are useful for illustrating the similarity or differences differences in brand (or prod uct) sensory profiles. BA
Color
AB
Sweet
AB BA BA AB BA
Bitter Others/Off-Notes
Fig. Fig. 5.
3.0
4.0 5.0 Rating
D DA A
e Other Coffee Aromatics t u Degree of Roast b i r Sweet Aromatics t t A Sour Aromatics
AB
Mouthfeel
A
Coffee Aromatic Intensity
A BA BA
2.0
A
Fullness
BA
Coffee Aromatic Intensity
D D
A
Balance
B
e Other Coffee Aromatics t u Degree of Roast b i r Sweet Aromatics t t A Sour Aromatics Sweet
Cup Aroma
BA
Balance Fullness
DA DA
Sour Bitter
DA
Others/Off-Notes
AD A D AD A D
Aftertaste 7.0
Coffee Recent Recent Developmen ts . 2001. R. J. Clarke and O. G. Vitzthum, editor s. Blackwell Blackwell Science Ltd.
D
DA
Mouthfeel
6.0
References Dawido wska, Katarzyna. April 2002. Caffeine Caffeine Overload. Ameri can Demograph ics, Vol. Vol. 24:4, 16. 16.
DA A D
Color
BA
Cup Aroma
1.0
2.0
3.0
• Brands that share share a common common sensory space, space, such such as “ light roast,” probably compete in the marketplace for the same consumer taste segments; the consumer’s perception of quality may be the final differentiator. • Brands in a different different sensory space, space, such such as “ dark roast,” focus on a very different consumer segment.
Caffeinated Decaffeinated
4.0 5.0 Rating
6.0
Sensory Sensory Attribute Attribute Line Plots Plots
• Locations Locations within the map represe represent nt different different sensory profiles and possibly different consumer taste segments.
22 5.0
Degree of Roast
Several key learnings from this study are summarized summarized below :
4
30
31 27
24 12 20 7 10 18 19 16 23 17
6.0
15
F ig ig . 3 .
Coffee Attribute Attribute Scales Scales
7.0
25
26 28 1
8
More others/off-notes More other coffee aromatics 4.0
Panelists record word descriptors of the perceived effect effect for the indicatedattributes.
Fig. Fig. 1. 1.
32
Caffeine: Form Whole Bean
F
1.0
” s c i t s 2 i x r e e t d c n a I r r a o h v C a l r F e h t O “
L
G E G E
Aftertaste
Lessothers/off-notes - notes 8.0 Lessothercoffee aromatics
D
B A
Sour
• Panelists used spring water and unsalted unsalted crackers crackers to rinse their m ouths between samples and waited approximately 5 min utes between samples. samples.
Brand A B C D E F G H I J K L
• Evidence Evidence supports supports highly accurate accurate brand classification based on the sensory data alone, though some m isclassific isclassification ation for a few less discriminated brands exists.
Results and D iscus iscussion sion
Index 1
• Five of the twelve brands brands showed showed similar sensory sensory brand id entity for their pr oducts, regardless of form or caffeine level.
Bitter intensity is a measure of Bitter intensity the level of bitter basic taste. The reference standards are caffeine solutions. The scale ranges from none to strong.
Coffee Aromatics Intensity is Intensity is a measure of the strength and type of coffee-like flavor aromatics. Panelists Panelists score this none to strong and describe their notes wh ich can include descriptions of coffee roast and mercaptans.
Protocol. The coffees were evaluated by a trained panel in a controlled environment as follows: • Profile Attribute Analysis Analysis (PAA), a quantitative quantitative descriptive analysis method, was used to evaluate the coffees using the scales defined below.
Sour measures Sour measures the intensity of sour b asic taste.The reference standards are citric acid solutions. The scale ranges from none to strong.
Linear discriminant analysis shows the coffees may be correctly classified classified by brand membership using the sensory attributes.
• Cup aroma, aroma, coffee coffee intensity and and type, sour aromatics, balance, and bitterness are the sensory sensory attributes most im portant in distinguishing and classifying classifying brands. • Consistenc Consistency y is key key to ensuring ensuring sensory sensory brand identity (illustrated by the relative size of ellipses).
7.0
Heath, Robert. 1999. Just popping down to the shops for a packet of image statements: A new theory of how consumers perceive perceive brands. Journal of Market Research Society, London. Vol. 41:2, 153–169. Building Strong Brands . 1996. David A. Aaker. New York: The Free Press. Marketing Aesthetics: The Strategic Management of Brands, Identity and Image . 1997. Bernd H. Schmitt and Al ex Simo nson. The Free Free Press. Press. Anonym ous. January 2000. Which brew for you? Consumer Reports, Vol. 65:1, 29–33. Watkins, Carolyn. Carolyn. May 2000. Coffee service: A whole latte opportunity. Food Management, Management, Vol. 35:5, 66–72.