Apple Preference Mapping
A central location trial was conducted for 10 dessert apple varieties using participants from Bristol and Hull. The following presentation will analyse the statistical results gained from the data received and suggest areas for future product launches within the UK apple industry.

Task 1

Recruitment of Panellists
The Market Research Company was asked to recruit 100 Panellists in each city for the trial with 50% of the panellists from each age group (1 & 2). The data shows that the number of people recruited in Bristol exceeded the required amount with a total of 169 Panellists. Further to this, 71% of the panellists recruited were from the age bracket 2. Recruitment in Hull matched closely to the total requirement of panellists, however again the recruitment favoured towards the age bracket 2 with 73% coming from this. The company therefore exceeded the total requirement of panellists by 71 people.
Sum of Age City Bristol Bristol Total Hull Hull Total Grand Total Age 1 2 1 2 Total 49 120 169 28 74 102 271

Task 1

Overall Mean Liking

Data Average of Top Red Average of Gibson's Green Average of Royal Gala Average of Braeburn Average of Fuji Average of Pink Lady Average of Granny Smith Average of Sun Gold Average of Johnson's Red Average of Golden Delicious Total 4.290229885 4.5 4.790229885 5.324712644 5.483908046 5.701149425 5.713218391 5.876436782 6.094252874 6.962643678

Overall Golden Delicious had the highest mean liking with a score of 6.96 overall. Top Red was liked the least with a score of 4.29 overall. Granny Smith and Pink Lady scored similar results with scores of 5.71 and 5.70 overall.

Task 2

Cluster Means
• Participants were divided into three clusters dependant on their scoring. The bar chart outlined below indicates the average cluster scores per individual apple.

Cluster membership 1 2 3 Grand Total

Average of Gibson's Green Average of Johnson's Red Average of Golden Delicious Average of Granny Smith Average of Pink Lady Average of Fuji Average of Top Red Average of Braeburn Average of Royal Gala Average of Sun Gold 4.724166667 4.83 6.7175 6.694166667 6.28 4.786666667 3.228333333 5.5225 3.9975 6.933333333 3.803921569 8.966666667 7.615686275 3.417647059 4.425490196 7.17254902 6.552941176 4.807843137 6.549019608 3.233333333 7.366666667 7.833333333 5.666666667 5.5 4.233333333 4.666666667 8.3 6.2 6.6 8.533333333 4.5 6.094252874 6.962643678 5.713218391 5.701149425 5.483908046 4.290229885 5.324712644 4.790229885 5.876436782

Task 2

Traits
Count of Consumer Cluster membership 1 2 3 Grand Total Age 1 50 25 2 77 2 Grand Total 70 120 26 51 1 3 97 174

Age: 1 = Less than 35 years old 2= >35 years old

The highest proportion of participants belonged to cluster 1 with 58% of these aged over 35 years old (2). Participants from Cluster 1 ranked their scores similarly for all the apple samples in comparison to clusters 2 and 3. Cluster 2 retained a valuable set of participants with equal counts from both age groups. They ranked highly on individual samples (such as Johnson’s Red) and lower on others. Cluster 3 only contributed 3 participants to the scoring. Cluster 3 scored highly on the majority of the varieties as an average (over point 5) and could be labelled as ‘slightly extreme’ scorers. It would be more appropriate to focus on clusters 1 and 2 rather than cluster 3.

Task 2

Traits
Count of Consumer Children Age 1 1 39 2 59 Grand Total 98 2 Grand Total 38 77 38 97 76 174

Children: 1 = Yes 2 = No

Participants Under 35 years of Age (1) were equally likely to have a presence of children in the family as to not. However, participants over the age of 35 (2) were more likely to have a presence of children (61%). It would therefore be interesting to analyse the trends that may exist between age/children and overall liking. One may assume that if children exist within the family then the participant may be less likely to over/under score on the samples as they may be more familiar with the product (Parents tend to encourage children to eat fruit and therefore may eat it more regularly themselves).

Task 3

Internal Preference Mapping

The Scree Plot suggests that we should look mainly at the first two dimensions as these correspond well, however the third dimension could be analysed in addition to the first two. The ‘Elbow point’ (F4-F8) where the plot begins to level off indicates the number of dimensions we should investigate. Dimensions that are higher than the elbow point spread the scores further apart are therefore more appropriate to investigate.

Task 3

Internal Preference Mapping

1

2

The plot indicates a segmented population with three clusters within the group. F1 Explains 29.53% of the variation between the samples in the 1st dimension. F2 explains 16.67% of the variation in the second dimension. There is more variation across the plot than up/down.

The segments are split into participants who like; red apples, green apples or mixed breed apples (red/green). New Breeds P1 should give a higher liking score for the segment that prefers Red Apples. P2 should help raise the overall liking score for both clusters that tend to like mixed and green apples.

Task 3 – Part 2

Internal Preference Mapping – Repeat Analysis
The graph displays the sensory attributes that correlate well for clusters 1 and 2. Cluster 3 can be ignored during the repeat analysis as we only want to consider the cluster means which have a class membership >20 (considered as more reputable) within the supplementary variables. Cluster 1 Sensory Drivers – The sensory attributes F_Acidic/Sour, T_Fibrous, A_Astringent, T_Fibrous bits, T_Crispness and A_Peel correspond well to the sensory liking for the cluster. Participants disliked an apple that provided characteristics including; T_Floury, T_Granular and A_Drying aswell as an apple that was sweeter in flavour and contained characteristics which related to pears. It may be considered that participants falling within Cluster 1 prefer an apple that would contain characteristics closer to varieties similar to that of Granny Smith or Golden Delicious that provide a sharper/crisper taste. New Breeds P1 suggests where we should be aiming to breed a new apple for Cluster 1. We chose this area as we take the liking from members of Cluster 1 and drive their liking towards other attributes that score closely. We hope it can bring two clusters together and provide and apple which cluster 1 like but also contains bitter attributes which other consumers may be attracted to.

Task 3 – Part 2

Internal Preference Mapping – Repeat Measures
Cluster 2 Sensory Drivers – The attributes T_Floury, F_Pearlike, F_Sweet, A_drying and A_Granular correspond well to cluster 2. The group also disliked the sensory characteristics that repond higher to green apples such as F_Bitter, A_Bitter and F_Acidic. Data suggests that cluster 2 prefer an apple that correlates well to fruity characteristics and is slightly more ripe with a sweeter flavour. Well-ripened apples begin to break down more within their internal structure leading to a granulated and floury texture which the cluster preferred. New Breeds P2 suggests where we should breed a new apple for cluster 2. We believe that the sensory characteristics of F_Watery and T_Pulpy could work well in conjunction with the sensory drivers for the cluster. We therefore hope that this new breed would help bring consumers that prefer a juicier apple towards cluster two and help open up a new market where consumers like the products on both dimensions.

2

1

Task 4

Extending the Portfolio to 4 Products
3 3 2

2

1 1 4

4

Positions 3 and 4 indicate the areas we would suggest for expanding the portfolio to four products. Position 3 Would allow for a green apple with flavour characteristics similar to grapes with slight bitter notes but with a bold juicy texture. We could boost the overall liking on the second dimension by including characteristics F_Watery and T_pulpy which also scored high to help contribute to the juiciness we will be looking for. Similar to Granny smith, which would be the products main competitor, we would expect the apple to score low on the first dimension but would be looking to boost the second dimension score above granny smith to offer a niche product which can compete and take control of the green apple market through distinct flavour offering of Grapes.

Task 4

Extending the Portfolio to 4 Products
3 3 2

2

1 1 4

4

Position 4 Would focus again on flavour characteristics of other fruits but this time Plums, Strawberries and Cherry’s would be the primary focus. The sensory drivers scored highly and the fruits are generally well-liked within the UK. Sweet, Floral notes could also be incorporated to boost the over all liking on the 1st dimension, however we would expect this product to score low on the second but feel that it has a niche offering to a specific target market which consumers may like. A dark red skin would be ideal in order to aesthetically replicate the fruit flavours we are aiming for. Johnson’s Red and Royal Gala would be the main competitors to this apple but again we feel the sweetness, distinct flavour notes and dark skin would help distinguish our new product away from these competitors.

Task 4

Extending the Portfolio to 4 Products
3 3 2

2

1 1 4

4

Marketing Recommendations We would recommend launching products 1 and 2 before extending the portfolio to successfully assess their success within the marketplace. We aim for product 1 to compete with product 3’s main competitor (Granny Smith) on the second dimension and for product 2 to compete with Product 4’s main competitors (Johnson’s Red and Royal Gala) on the first dimension. We can therefore analyse the positions these products are in and their overall liking in comparison to their competitors before launching additional future products. If the initial two products perform well and compete with similar sensory drivers and overall liking to these competitors then we aim to release the additional two products into the portfolio range. We feel four products is a large enough range for the time being and once assessing the consumer liking to the initial two products we feel we can drive additional sales volume into the retailers dependant on their likability and hope to repeat this forecast model for the additional products ranges.