Pricing Single Malt Whisky

IN

DEGREE PROJECT TECHNOLOGY, FIRST CYCLE, 15 CREDITS

STOCKHOLM SWEDEN 2016,

Pricing Single Malt Whisky

A Regression Analysis

SANNE BJARTMAR HYLTA EMMA LUNDQUIST

KTH ROYAL INSTITUTE OF TECHNOLOGY

Pricing Single Malt Whisky

A Regression Analysis

S A N N E B J A R T M A R H Y L T A E M M A L U N D Q U I S T

Degree Project in Applied Mathematics and Industrial Economics (15 credits) Degree Progr. in Industrial Engineering and Management (300 credits)

Royal Institute of Technology year 2016 Supervisors at KTH: Thomas Önskog, Jonatan Freilich

Examiner: Henrik Hult

TRITA-MAT-K 2016:05 ISRN-KTH/MAT/K--16/05--SE

Royal Institute of Technology SCI School of Engineering Sciences KTH SCI SE-100 44 Stockholm, Sweden URL: www.kth.se/sci

Abstract

This thesis examines the factors that affect the price of whisky. Multiple regression analysis is used to model the relationship between the identified covariates that are believed to impact the price of whisky.

The optimal marketing strategy for whisky producers in the regions Islay and Camp- beltown are discussed. This analysis is based on the Marketing Mix. Furthermore, a Porter’s five forces analysis, focusing on the regions Campeltown and Islay, is exam- ined. Finally the findings are summarised in a marketing strategy recommendation for producers in the regions Campbeltown and Islay.

The result from the regression analysis shows that the covariates alcohol content and regions are affecting price the most. The small regions Islay and Campbeltown, with few distilleries, have a strong positive impact on price while whisky from unspecified regions in Scotland have a negative impact on price. The alcohol content has a positive, non-linear, impact on price.

The thesis concludes that the positive relationship between alcohol content and price not is due to the the alcohol taxes in Sweden, but that customers are ready to pay more for a whisky with higher alcohol content. In addition, it concludes that small regions with a few distilleries result in a higher price on whisky. The origin and tra- dition of whisky have a significant impact on price and should thus be emphasised in the marketing strategy for these companies.

Sammanfattning

Denna kandidatuppsats unders¨oker de faktorer som p˚averkar priset p˚a whisky. Mul- tipel regressionsanalys anv¨ands f¨or att modellera f¨orh˚allandet mellan de identifierade variablerna som tros p˚averka priset p˚a whisky.

Vidare diskuteras den optimala marknadsf¨oringsstrategi f¨or whiskyproducenter i re- gionerna Islay och Campbeltown. Analysen baseras p˚a en Marknadsmix-analys f¨or whisky i Skottland. Detta f¨oljs av Porters femkraftsmodell med fokus p˚a regionerna Islay och Campeltown. Slutligen sammanfattas resultaten i en rekommendation av marknadsf¨oringsstrategi f¨or producenter i regionerna Islay och Campbeltown.

Resultatet fr˚an regressionsanalysen visar att kovariaterna alkoholhalt och regioner har st¨orst p˚averkan p˚a priset. De sm˚a regionerna Islay och Campbeltown, med f˚a destil- lerier, har en stark positiv inverkan p˚a priset. Whisky fr˚an ospecificerade regioner i Skottland har d¨aremot en negativ inverkan. Alkoholhalten har en positiv, icke-linj¨ar inverkan p˚a priset.

I kandidatuppsatsen dras slutsatsen att det positiva sambandet mellan alkohol och pris ej kan f¨orklaras av Sveriges alkoholskatt, utan att kunder ¨ar redo att betala mer f¨or en whisky med h¨ogre alkoholhalt. Vidare konstateras att sm˚a regioner med f˚a des- tillerier resulterar i ett h¨ogre pris p˚a whisky. Whiskyns ursprung och tradition har en stor inverkan p˚a pris och b¨or d¨arf¨or betonas i marknadsf¨oringen.

Contents

1 Introduction 6

1.1 Background . . . 6

1.1.1 Thesis Background . . . 6

1.1.2 Whisky in Scotland . . . 6

1.1.3 Single malt whisky production . . . 7

1.2 Problem definition . . . 8

1.3 Purpose and aim . . . 8

2 Mathematical Background 9 2.1 Multiple regression analysis . . . 9

2.1.1 Description . . . 9

2.1.2 Ordinary Least Squares . . . 10

2.2 Assumptions in Ordinary Least Squares method . . . 10

2.2.1 Homoscedasticity and no multicollinearity . . . 11

2.2.2 Normally distributed residuals . . . 11

2.2.3 Strict exogeneity . . . 11

2.3 Errors . . . 11

2.3.1 Multicollinearity . . . 11

2.3.2 Heteroscedasticity . . . 12

2.3.3 Endogeneity . . . 13

2.4 Model valuation . . . 14

2.4.1 Hypothesis testing . . . 14

2.4.2 t-test and hypothesis testing . . . 14

2.4.3 R² . . . 15

2.4.4 F-test . . . 16

2.4.5 BIC - Bayesian information criterion . . . 16

2.4.6 AIC - Akaike Information criterion . . . 16

2.4.7 Combining AIC and BIC . . . 17

3 Method 18 3.1 Data Collection . . . 18

3.2 Variables . . . 18

3.2.1 Response Variable . . . 18

3.2.2 Covariates . . . 18

3.3 Initial model . . . 19

4 Results 21 4.1 Initial model . . . 21

4.2 Initial model validation . . . 22

4.2.1 Residual diagnostics . . . 22

4.2.2 F-statistic and p-value . . . 24

4.2.3 R² and R² adjusted . . . 25

4.2.4 VIF-test . . . 25

4.3 Reducing the model . . . 25

4.4 Final model . . . 26

4.5 Final model validation . . . 27

4.5.1 Residual diagnostics . . . 27

4.5.2 F-statistic and p-value . . . 29

4.5.3 R² and R² adjusted . . . 30

4.5.4 VIF-test . . . 30

5 Discussion and conclusion 31 5.1 Analysis of covariates in the Final model . . . 31

5.1.1 Alcohol content . . . 31

5.1.2 Islay . . . 31

5.1.3 Cambpeltown . . . 31

5.1.4 Other . . . 31

5.2 Discussion and conclusion of mathematical model . . . 32

6 Whisky in Scotland from a marketing perspective 34 6.1 Introduction . . . 34

6.2 Marketing Mix . . . 34

6.2.1 Product & Consumer wants and needs . . . 34

6.2.2 Price & Cost . . . 35

6.2.3 Promotion & Communication . . . 35

6.2.4 Distribution & Convenience . . . 36

6.3 Porter’s Five Forces Analysis . . . 36

6.3.1 Theory . . . 36

6.3.2 Analysis of Whisky from Islay and Campbeltown . . . 37

6.4 Recommendation of marketing strategy for producers in Islay and Camp- beltown . . . 39

7 References 40

4

List of Figures

1 Map over Scotland and its whisky regions . . . 7

2 Scale Location plot for the Initial Model . . . 22

3 Residuals versus fitted values for the Initial Model . . . 23

4 Normal QQ-plot for the Standardized residuals, Initial Model . . . 24

5 Scale Location plot for the Final Model . . . 27

6 Residuals versus fitted values for the Final Model . . . 28

7 Normal QQ-plot for the Standardized residuals, Final Model . . . 29

8 Porter’s Five Forces [14] . . . 37

List of Tables

2 Regression results for the initial model . . . 21

3 Continued, Regression results for the initial model . . . 21

4 Table of VIF-test values for the covariates in the initial model . . . 25

5 Covariate data and statistics . . . 25

6 Regression results for the final model . . . 26

7 Continued, Regression results for the final model . . . 26

8 Table of VIF-test values for the covariates in the final model . . . 30

1 Introduction

1.1 Background

1.1.1 Thesis Background

Malt whisky has historically been an important spirit with its many classes and types, especially in Scotland. Since the variation in grain, processing, maturation, region of origin and alcohol content is vast, the price of whisky differs significantly. Generally the price of a product reflects its quality and content, but when there are additional factors affecting price, the price model becomes more complex. This is the case for whisky, where it is difficult to derive a whisky bottle’s certain price. As whisky is considered to be such a traditional spirit, especially in Scotland, the brands and the distilleries’ regions are important for certain consumers. Thus, a whisky’s origin has a large impact on price.

1.1.2 Whisky in Scotland

Whisky is Scotland’s national drink and is made of fermented grain mash. The fermenta- tion together with the distillation and the aging in wooden barrels are the most significant characteristics for different classes of whisky. Various grains such as barley, corn, wheat, buckwheat and rye can be used and they all have a unique, characteristic taste. For single malt whisky, the grain is malted barley, but there are also a number of different barley varieties, resulting in different tastes. [8]

In this thesis we will only analyse single malt whisky from Scotland, i.e. malt whisky from one distillery only. Since the thesis aims to analyse the orgin’s impact, blended whiskys from several distilleries can not be included.

Scotland is divided into following six whisky regions:

1. Lowlands: The most southern region with a light and neutral whisky.

2. Highlands: The most northern and biggest region where the whisky is elegant and tasty with a little sweetness.

3. Speyside: Located in the northeastern corner of Scotland, sometimes part of the Highlands, with a large number of distilleries. The taste of the whisky is sweet, fruity and complex.

4. Islands: Covering the islands in the north west excluding the island Islay. Sometimes considered to be part of the Highlands. The whisky is extremely varied with few similarities, but generally smoky with peaty undertones and marked salinity.

5. Islay: Islay is one of the southernmost islands and has nine active distilleries. The whisky is powerful with a smoky, peaty character.

6. Campbeltown: The area around Campbeltown is a historical whisky region with only three distilleries remaining. The characteristics of the whisky include a defined dryness with a pungency, smoke and a solid salinity. [8].

Figure 1: Map over Scotland and its whisky regions

1.1.3 Single malt whisky production

The process of making single malt whisky is long and complex. It includes the following processes:

1. Malting: The barley is soaked in water to undergo germination, to convert the starch into soluble sugars. The barley is then dried in a kiln, traditionally with peat used to power it which influences the taste.

2. Mashing: The malted grain is crushed and mixed with hot water into a mash tun.

The sugar in the malt dissolves and is drawn off. This liquid is called wort.

7

3. Fermentation: The fermentation process starts by adding yeast to the wort. The sugars are turned into alcohol and at this stage the liquid is called wash.

4. Distillation: The wash is traditionally distilled twice in Scotland. The shape and material of the stills influence the taste and usually stills are made of copper to remove sulfur-based compounds from the alcohol. First, the wash distillation produces a liquid called low wines, with a low level of alcohol. It is then re-distilled in a spirit still.

5. Maturation: The whisky is matured in oak casks, giving it its characteristic taste.

The age of a whisky is the time between distillation and bottling. A whisky bottled for many years has still the same age but may get a rarity value. The type of cask has a great impact on the taste since the whisky undergoes a number of processes.

[8].

1.2 Problem definition

The research question in this thesis is: ”What factors impact the price of a bottle of whisky and which ones have the highest significance?”. There are many possible fac- tors, such as origin, storage time, processing and alcohol percentage that are believed to influence the price. From a business point of view, to maximize profit the pricing is extremely important for whisky producers. Given the price drivers of single malt whisky the additional research question is: ”What marketing strategy should whisky producers use?”.

1.3 Purpose and aim

This thesis aims to identify the most important factors affecting price for single malt whisky in Scotland by creating a pricing model. It is aimed at producers, to give them a clear and scientific pricing model on which parameters determine the price of single malt whisky in Scotland. This could be a tool for producers to improve their pricing strategy and when introducing new products. For example there might be differences in the optimal strategy for different areas in Scotland, depending on the traditions around them and how well known they are. In addition the pricing model can be used by consumers, who have a great interest in whisky as the model can provide them with additional information when comparing different bottles of whiskey.

2 Mathematical Background

2.1 Multiple regression analysis

2.1.1 Description

Regression analysis is a common technique in mathematical statistics that examines and models the relationship between certain independent variables, called covariates, and how they affect the dependent variable, called the response variable. The purpose is to find the function that best fits the observed data. Frequently used models in regression analysis is simple linear regression, multiple linear regression, polynomial regression, logistic regression and nonlinear regression. Regression analysis can be applied to many different fields including engineering, economics, management, so- cial sciences and biotechnology. [1]. In this paper we examine a relationship between several variables, and thus the multiple linear regression will be used.

The response variable, y_i, in multiple regression analysis is defined as a set of obser- vations that depend on the covariates, xij which are regarded as deterministic. βj

are estimated to get an estimation of the line. The residuals, _i, are random variables that are independent between observations. The definition of the linear regression model yields [3]:

y_i=

k

X

j=0

x_ijβ_j + _i, i = 1, 2, .., n (1) This can also be expressed in matrix form:

Y = Xβ +  (2)

where E[]=0 and E[^t]= Iσ²,

and Y =











 y₁ y₂ . . y_n













, X =













1 x₁₁ . . x_1k 1 x₂₁ . . x_2k

. . . . .

. . . . .

1 x_n1 . . x_nk











 , β =











 β₀ β₁ . . β_k











 ,  =













₀

₁ . .

_n













• The β are the unknown coefficients corresponding to the covariates X. The regression will estimate these values based on the data. The estimates of β explains what effect a certain covariate has on the response variable.

9

• E[]=0 means that covariates are not biased and affected by the dependent variable and E[^t]= Iσ² states that the variance for the different covariates is the same.

•  = (₀, ..., _n)^t is the error term containing the residuals for the covariates

In our work and analysis a structural interpretation will be used, and not predictions.

The covariates should hence influence the dependent variable, but never the other way around. This assumption make hypothesis testing possible.

2.1.2 Ordinary Least Squares

The Ordinary Least Squares (OLS) is a method used to estimate the values of β, which is required when performing a regression analysis. The method calculates the optimal estimate of the parameter β, provided that all the assumptions for the regression and for the Ordinary Least Squares method are met. [3]. The assumption is further discussed in the following section. The optimal estimates of the β_i and the residuals, _i are denoted as βˆi and ˆi. The Least Squares method minimizes the sum of squared residuals and therefore minimizes the following expression: [4]:

n

X

i=1

ˆ

²_i =

n

X

i=1

(yi− ˆyi)² = (Y − X ˆβ)^T(Y − X ˆβ) (3)

The estimate ˆβ is obtained by solving the following normal equation for ˆβ:

X^Tˆ = 0 (4)

Hence, the least square estimate ˆβ of β yields:

β = ( Xˆ ^TX)⁻¹X^T Y (5)

Furthermore the covariance matrix for ˆβ can be derived as:

Cov( ˆβ) = (X^TX)⁻¹σ² (6)

An unbiased estimation of σ² yields:

σ²= 1

n − k − 1|ˆ|² (7)

where n defines the number of observations and the k number of covariates.

2.2 Assumptions in Ordinary Least Squares method

The linear regression model is based on a number of basic assumptions that have to be met.

These assumption holds, and plays an important role, for the OLS method. [3]. The major assumptions are stated below. These are also the assumptions that will be investigated in our models in this thesis:

2.2.1 Homoscedasticity and no multicollinearity

All random variables have the same finite variance and they are not correlated with each other:

E[ij] = 0, i 6= j. (8)

E[ij] = σ², i = j (9)

2.2.2 Normally distributed residuals

It is assumed that the residuals have normal distribution conditional on the regressors:

|X ∼ N (0, σ²I_n) (10)

2.2.3 Strict exogeneity

Strict exogeneity: The covariates are not correlated to the residuals. Hence the residueals in the regression should have conditional mean zero:

E[|X] = 0 (11)

2.3 Errors

Three common problems that violates the assumptions in OLS are presented below.

2.3.1 Multicollinearity

Multicollinearity occurs when two or more covariates are moderately or highly correlated.

This means that they can be written as linear combinations of each other and the effect is increased variance.

There are several indicators and tests to find multicollinearity. Multicollinearity can be de- tected using the Variance Inflation Factor (VIF). This provides an index that measures how much the variance of an estimated regression coefficient is increased because of collinearity.

[5]

Consider the following linear model with k independent variables:

Y = β0+ β1X1+ ... + βkX_k+  (12)

We can calculate k different VIFs (one for each X_i) in three steps [5]:

11

1. Run an ordinary least square regression that has Xi as a function of all the other explanatory variables in the first equation.

For i = 1 this yields:

X1 = α2X2+ ... + α_kX_k+ c0+  (13) Where c0 is a constant and e is the error term.

2. Calculate the VIF factor for ˆβ_i with the following formula:

V IF = 1

1 − R²_i (14)

Where R²_i is the coefficient of determination of the regression equation in step one, with X_i on the left hand side of the equation and all other predictor variables on the right hand side.

3. Analyse the magnitude of multicollinearity by looking at V IF ( ˆβi) or the tolerance, T = _{V IF}¹ . A tolerance of less than 0.20 or 0.10 and/or a VIF of 5 or 10 and above indicates a multicollinearity problem.

To get rid of multicollinearity simply choose one of the variables to omit the other [4].

2.3.2 Heteroscedasticity

In linear regression the error terms i are assumed to be homoscedastic, which means that they all have the same variance σ. Heteroscedasticity occurs when this assumption does not hold, i.e. the error terms do not have constant variance. If a heteroscedastic equation is misspecified as homoscedastic the standard deviations of the residuals vary and cause inconsistent standard deviation of the β-values. The t-test explained later in this chapter are then no longer valid. The model specification is: [4]

y_i =

k

X

j=0

x_ijβ_j+ _i, i = 1, 2, .., n (15)

where E[i] = 0 och E[²_i] = σ²_i, ei are independent between observations

Heteroscedasticity can be caused by several factors, for example poor data collecting tech- niques making σ² fairly large or skewness in the data distribution. [4]

Heteroscedasticity can be detected graphically by study of a residual plot, where the resid- uals are plotted against the covariates. Large variation in the size of deviations from the x-asis or unevenly distributed point indicates heteroscedasticity. [3]

Heteroscedasticity can also be detected through a quantile-quantile plot (QQ-plot). This compares two probability distributions by plotting their quantiles against each other. In cases where the residuals would be tested against a normal distribution the residuals are plotted against the theoretical normal distribution. A straight line of 45 degrees in the graph indicates that the residuals are approximately normal distributed. Points that de- viates much from the straight line indicates that they are not normally distributed and consequently there might be a problem with heteroscedasticity. [6]

The White’s Consistent Variance Estimator can be implemented to reduce heteroscedastic- ity. White’s Consistent Variance Estimator is a covariance matrix that can be used instead of the usual covariance matrix. The matrix scales down with _n−k−1ⁿ and is defined as: [4]

Cov( ˆβ) = (X^TX)⁻¹X^TD(ˆ²)X(X^TX)⁻¹ n

n − k − 1 (16)

Where D(ˆ²) is a nxn diagonal matrix:

D(ˆ²) =











 ˆ

12 0 · · · 0 0 . .. 0 ... ... . .. ... 0 0 · · · 0 ˆn2













(17)

2.3.3 Endogeneity

Endogeneity is violation against exogeneity, which is assumed in OLS. This occurs when the expected value of ei depends on one or more of the covariates and the assumption that E[i] = 0 is violated. Endogeneity can arise as a result of measurement errors and autoregression with autocorrelated residuals. [4]

To detect endogeneity one can analyze the covariates to see if there are any covariates that has been omitted but should be in the model or if the residual is correlated with one or several of the covariates. [4]

If the models covariates are endogenous another method such as the one of instrumen- tal variables (IV) and Two Stage Least Squares (2SLS) is recommended as this will bring consistent estimations. [4]

13

2.4 Model valuation 2.4.1 Hypothesis testing

Hypothesis testing is a statistical method to examine if a hypothesis can be rejected at a ceratin significance level. When using hypothesis testing further assumptions arises, namely that the residuals should be normally distributed., _i ∼ N (0, σ²). [1]

The algorithm for the hypothesis is divided in 4 steps [7]:

1. Determine a significance value α based on preferences. Formulate the null hypothesis H0, and an alternative hypothesis Hα.

2. Identify the test statistic, T, that can be used to determine whether H0 should be rejected or not.

3. Compute the p-value. The p-value gives the probability that a test statistic is at least as significant as the one observed when H0 is true. The smaller the p-value, the stronger the evidence against H₀.

4. If the p-value ≤ α the null hypothesis is rejected and the alternative hypothesis is valid.

2.4.2 t-test and hypothesis testing

The t-test is a common test to test and identify the parameters significance. When perform- ing the t-test, only one linear constraint can be tested at a time. Examples of parameters that can be used during the test and in the linear constrains are one of the β-values and the correlation coefficient r.

When performing the t-test for β value , βi , the linear conditions is reformulated and set equal to zero. This forms the null hypothesis [1]:

H0: βi= 0

Hα: βi6= 0 (18)

The degrees of freedom are set equal to the number of observations subtracted with the number of parameters estimated in the regression. The null hypothesis is that the linear condition is true, and that β is zero which means that the covariate x_i is not the response variable. The alternative hypothesis H_α is that the covariate x_i is a influencing factor on the response variable and therefore that βi is not zero.

The t-value is obtained by dividing the β-value with the standard error for the same β-value [9].

ti = βi

S.E(βi) (19)

A |t| close to zero indicates that the tested covariates is not explanatory of the response variable, and that these might can be excluded from the model. To determine if the parameter can be excluded from the model, calculate the p-value:

p − value = P (T ≥ t), T ∈ t(n − 2) (20)

Determine the p-value with the chosen significant level α. If p ≤ α the null hypothesis can be rejected and the alternative hypothesis is valid. [9]

2.4.3 R²

R² is a measure of goodness of fit and is defined as the square of the sample correlation coefficient between y and the least squares estimation ˆy. R² measures the proportions of variance in the response variable that can be explained by the covariates. In the linear case which is used in this thesis, the least squares estimation is used and the total variance in the responding variable can be divided into two parts: the explained variance and the unexplained variance. The explained variance is defined as the sum of the squared deviations for the estimated value from its mean value. The unexplained variance is defined as the sum of the squared residuals. R² is defined as the fraction of the explained variance and the unexplained variance [4]:

R²= V ar(x ˆβ)

V ar(y) = 1 − V ar(ˆ)

V ar(y) (21)

R² is always increasing when adding more covariates, even though this not nessisarely means that the goodness of fit is better. A high R² can be an indication of over-fitting which means having too many covariates in the model. The adjusted R² is a modified version of R² that has been adjusted for the number of predictors in the model. The adjusted R² increases only if the new term improves the model more than is expected by chance. The adjusted R² will hence peak when the optimal amount of covariates appears.

The adjusted R², ¯R², is defined as: [10]

R¯²= 1 − (1 − R²)n − 1

n − p (22)

where p is the total number of explanatory variables in the model (not including constant term), and n is the sample size.

15

2.4.4 F-test

F-test is a statistical test in which the test statistic has a F -distribution under the null hypothesis. F s assigned probability distribution is used to calculate P r(X > F ), where X is a random variable. The F-test is used to test the hypothesis that a number of r of β-values are equal to zero, β1 = β2 = .. = βr = 0. It can be used for testing simple and multiple regression models significance. The test variable is defined as:

F = n − k − 1 r

 |e∗|²

|ˆe|² − 1



(23) where e∗ is the residual from the restricted regression and ˆe is the residual from the unre- stricted regression.

The null hypothesis is rejected if the F-statistic is large. A rejected null hypothesis implies that the covariates corresponding to the β-values are significant. The F-test is significant under the assumption that the residuals are normal distributed. The test is asymptotic significant if the residuals are not normal distributed. [4]

2.4.5 BIC - Bayesian information criterion

The Bayesian information criterion can be used to compare different models. BIC is a func- tion that minimize the squares of the residuals and assumes normal distributed residuals.

The BIC-value can hence be used to determine which models has a better approximation of the true empirical data. A lower BIC-value indicates a more significant model.The formula for calculating BIC yields: [11]

BIC = ln |ˆe|² N



+K ∗ ln(N )

N (24)

Where N is the number of observations, K is number of covariates + the intercept and ˆe is the residuals from the unrestricted regression. A common approach is to analyse the BIC while removing covariates to find the model that minimizes the BIC.

2.4.6 AIC - Akaike Information criterion

The Akaike Information criterion is a measure of the relative quality of a model for a given set of data. Given a set of models for the data, AIC estimates the quality of each model relative to the other models. It also denotes the relative quantity of information lost when a given model, with estimated parameters, is compared to the true process that generated the data. An AIC-test can be used to test if one or more covariates would enter the equa- tion. Similar to the BIC, one should choose the model that minimizes: [4]

AIC = N ∗ ln(|ˆe|²) + 2K (25) Where N is the number of observations, K is number of covariates + the intercept and ˆ

e is the residuals from the unrestricted regression.

2.4.7 Combining AIC and BIC

In this thesis we will analyse the model from both an AIC and BIC perspective when determining our final model. We take this approach as BIC often gives a model with too few covariates and the AIC often choose a model with too many covariates. [11]

17

3 Method

3.1 Data Collection

To find a pricing model for whisky bottles, prices and information of a large amount of bottles was needed. The data was provided in the assortment file available at System- bolaget’s website [2]. Since Systembolaget is a non-profit organization, all bottles should be equivalently priced which is a requirement for the analysis of the price drivers. The data includes all whisky products available at Systembolaget, almost 2000 bottles of malt whisky. We limited the data to contain only 700 ml bottles from Scotland, which resulted in approximately 1000 bottles. These bottles form the basis of our data collection, with each bottle’s name, producer, price, volume, storage time and region.

In most cases the time of storage was included in the name, and an excel command was used to obtain this number. In other cases, where the name did not contain the storage time, it was calculated by subtracting the bottle’s vintage with the year of sell start. The accuracy of the method was confirmed by comparing bottles with calculated year with bottles with given storage time and it had a high grade of precision.

Furthermore, we manually looked up the whisky region for each distillery. The whisky regions used were; Lowlands, Highlands, Speyside, Islands, Islay and Campbeltown [8]. In some cases the regions were not specified and therefore a new region variable Other was introduced.

Outliers were removed in two ways. First all prices below 400 SEK or above 6500 SEK were eliminated, corresponding to the 5% quantiles and thus the remaining data contained 90% of all 700 ml bottles in Scotland. Secondly, all bottles without complete information were removed, i.e. if any covariate was missing.

3.2 Variables

3.2.1 Response Variable

The response variable for the regression analysis is the price of a 700 ml bottle of whisky, y. To eliminate heteroscedasticity the prices were logtransformed. Since the aim is to find what impact regions, alcohol and storage time have on the price the choice of response variable was obvious.

3.2.2 Covariates

The covariates in the regression are all factors believed to have an impact on the whisky price, including both qualitative and quantitative variables. To analyse qualitative factors the use of dummy variables is central. The dummy variable takes the value of one if the

statement is true and zero otherwise.

Storage time [years]

The storage time has a great impact on the taste of whisky and is therefore believed to affect the price as well. Whisky without known storage time have been excluded in this regression.

Alcohol content [%]

The alcohol content ranged from 40.00% to 65.10% in the set of whisky used in the re- gression and it is of importance to examine whether this has an effect on the price or not.

The alcohol covariate was set to be the difference from 40.00%, to create positive numbers ranging from 0.00 to 25.10. Since 40.00% is the minimum level of alcohol this should be used as reference, as it is only of interest to see how much a change in alcohol content from this level affects the price. The new alcohol content covariate gives a clear result of the impact of one unit’s increase.

Region [dummy variables]

The covariate Region was cathegorized into following six regions: Lowlands, Highlands, Speyside, Islands, Islay and Campbeltown. For each whisky region in Scotland a dummy variable was created, taking the value of one if a whisky was produced in the region and zero otherwise. The benchmark used was Highlands since it is the biggest region containing many data points.

3.3 Initial model

The initial model included the response variable, all covariates described in section 5.2.2 and their corresponding coefficients.

Table 1: Table of the response variable and covariates in the initial model

Variable Description Unit

Response Variable

y Price SEK

Covariates

x1,i Alcohol Content % above 40.00%

x2,i Storage Time Years

x3,i Storage Time² Years²

x4,i Lowlands Dummy; 0 or 1

x5,i Speyside Dummy; 0 or 1

x_6,i Islands Dummy; 0 or 1

x_7,i Islay Dummy; 0 or 1

x_8,i Campbeltown Dummy; 0 or 1

19

The initial model was thus:

log(y_i) = β₀+x_1,iβ₁+x_2,iβ₂+x_3,iβ₃+x_4,iβ₄+x_5,iβ₅+x_6,iβ₆+x_7,iβ₇+x_8,iβ₈+_i, i = 1, 2, .., n (26)

4 Results

4.1 Initial model

All covariates were included in the initial model except the benchmark ”Highlands”. High- land was assigned benchmark as this region had the highest quantity and most consistent data. This is a standard procedure when conducting a regression with dummy variables, in order to reduce the risk for linear dependency and multicollinearity.

Table 2: Regression results for the initial model

Model R² Adjusted R² F-statistics p-value

Initial Model 1.8663*10⁻¹ 1.7688*10⁻¹ 8.884 < 7.283 ∗ 10⁻¹³

Table 3: Continued, Regression results for the initial model

Coefficients Estimated β Standard Error t-value p-value

Intercept 6.873 4.937*10⁻² 139.211 <2*10⁻¹⁶

Alcohol Content 1.944*10⁻² 2.926*10⁻³ 6.645 5.420*10⁻¹¹ Storage Time -7.124*10⁻⁴ 3.873*10⁻³ -1.840*10⁻¹ 8.541*10⁻¹ Storage Time Squared 3.877*10⁻⁵ 9.755*10⁻⁵ 3.970*10⁻¹ 6.912*10⁻¹

Lowlands 1.055*10⁻¹ 8.593*10⁻² 1.228 2.190*10⁻¹

Speyside 2.677*10⁻² 3.906*10⁻² 6.850*10⁻¹ 4.933*10⁻¹

Islands −6.489 ∗ 10⁻² 6.797 ∗ 10⁻² −9.550 ∗ 10⁻¹ 33.990 ∗ 10⁻¹

Islay 1.338 ∗ 10⁻¹ 5.466 ∗ 10⁻² 2.448 1.456 ∗ 10⁻²

Campbeltown 2.617*10⁻¹ 1.026*10⁻¹ 2.550 1.094*10⁻²

Other −2.812 ∗ 10⁻¹ 7.889 ∗ 10⁻² −3.565 3.84 ∗ 10⁻⁴

21

4.2 Initial model validation 4.2.1 Residual diagnostics

Figure 2: Scale Location plot for the Initial Model

The model had logtransformed prices to reduce heteroscedasticity. The assumption of ho- moscedasticity is central when performing an OLS and can be investigated by analysing the Scale Location plot, see Figure 1. From the homoscedasticity it follows that the variability of the residuals should be stable and not change over the range of the dependent variable.

Hence there should not be any discernible patterns in the plot. The Scale Location shows a relative straight line without any major discrepancies. This indicates that the residuals indeed are homoscedastic.

Figure 3: Residuals versus fitted values for the Initial Model

The plot of residuals versus fitted values indicates a curved line and that the dependent variable might not be linearly related to all the covariates. It should therefore be investi- gated if a quadratic term for one or more of the covariates is suitable in the final model.

The plot also shows that the positive residuals seem to be larger than the negative residuals.

This indicates a somewhat skewed distribution where the assumption of normal distributed residuals might be violated. The negative deviations are smaller than the positive ones and therefore the model is a worse predictor of premium bottles.

23

Figure 4: Normal QQ-plot for the Standardized residuals, Initial Model

OLS assumes that the residuals are normal distributed. The QQ-plot forms approx- imately a 45 degrees line, with tails that implies that the distribution of the residuals is heavier tailed than the normal distribution. The plot is still fairly straight 45 degrees and therefore the assumption of standard normally distributed residuals holds.

4.2.2 F-statistic and p-value

The initial model have a small p-value of less 7.283 ∗ 10⁻¹³ and a F-statistic of 8.884. The significance level is set at 5% and the p-value is well below this, which implies that there are β-values that should not be set to 0. The individual parameters’ p-values show that some covariates have a p-value above 5%. They should consequently be reduced in the final model.

4.2.3 R² and R² adjusted

The R² value for the model is 18.7%, stating that the covariates explains approximately 19% of the response variable variance. Adjusted R² is 17.7%. This indicates that we have too many non-significant covariates in the model. The initial model consists of many variables that have a p-value larger than our significance level of 5%, which also confirms that there are too many covariates in the initial model. This partly explains the low R² as well.

4.2.4 VIF-test

Table 4: Table of VIF-test values for the covariates in the initial model

Covariate VIF

Alcohol Content 1.012609 Storage Time 5.563966 Storage Time ² 5.539375 Lowlands 1.113456 Speyside 1.525631

Islands 1.187003

Islay 1.296808

Campbeltown 1.072981

Other 1.127728

The variance inflation factor test indicates that there exists multicollinearity between Storage Time and Storage Time². This is natural due to their mathematical connection.

The other variables have a VIF values between 1 and 1.6 which does not indicate any multicollinearity. Multicollinearity in the model is hence neglected.

4.3 Reducing the model

To reduce the initial model the least significant covariates were eliminated by analysing the p-value, BIC, AIC and R². The model was reduced iteratively by eliminating the covariates one by one, starting with the highest p-value, until the AIC and BIC reached their minimum values.

Table 5: Covariate data and statistics

25

Covariate p-value AIC BIC Adjusted R² Storage Time 8.541 ∗ 10⁻¹ 1102.941 1155.176 7.689 ∗ 10⁻² Storage Time² 5.880 ∗ 10⁻¹ 1100.974 1148.462 7.795 ∗ 10⁻² Speyside 4.912 ∗ 10⁻¹ 1099.271 1142.010 7.871 ∗ 10⁻² Lowlands 2.925 ∗ 10⁻¹ 1097.751 1135.740 7.928 ∗ 10⁻² Islands 1.754 ∗ 10⁻¹ 1096.868 1130.109 7.916 ∗ 10⁻² Islay 1.391 ∗ 10⁻² 1096.718 1125.211 7.825 ∗ 10⁻² Campbeltown 2.081 ∗ 10⁻² 1100.805 1124.549 7.274 ∗ 10⁻² Other 2.310 ∗ 10⁻⁵ 1104.181 1125.186 6.798 ∗ 10⁻²

As seen in Table 5 the following covariates were eliminated: Storage Time, Storage Time², Speyside, Lowland and Islands. As explained in section 4.3.7 both AIC and BIC were taken into account when reducing the model. AIC reached the minimum value when Islands were eliminated but Islay was still in the model. BIC on the other hand, reached its minimum value after another iteration, when Islay was eliminated but Campbeltown was included. Since the p-value for Islay was below the chosen significance level of 5%, this covariate was kept in the final model and hence AIC was followed. By analysing the adjusted R² it was suggested to stop the reduction after eliminating Speyside, when R² was maximized, but since the p-value of Lowlands and Islands were above the significance level the iteration was continued until the minimum value of AIC was reached.

As described in section 6.2.1 all covariates were not linear related to the response vari- able and a quadratic term for one or several variables was needed. Therefore the new covariate, Alcohol Content² was added in the final model.

4.4 Final model

Table 6: Regression results for the final model

Model R² Adjusted R² F-statistics p-value

Final Model 2.061 ∗ 10⁻¹ 2.008 ∗ 10⁻¹ 20.110 < 2 ∗ 10⁻¹⁶

Table 7: Continued, Regression results for the final model Coefficients Estimated β Standard Error t-value p-value Intercept 6.680 5.497 ∗ 10⁻² 121.510 < 2 ∗ 10⁻¹⁶ Alcohol Content 7.111 ∗ 10⁻² 1.127 ∗ 10⁻² 6.312 4.431 ∗ 10⁻¹⁰ Alcohol Content² −2.379 ∗ 10⁻³ 5.037 ∗ 10⁻⁴ -4.723 2.721 ∗ 10⁻⁶ Islay 1.052 ∗ 10⁻¹ 4.772 ∗ 10⁻² 2.206 2.7684 ∗ 10⁻² Campbeltown 2.256 ∗ 10⁻¹ 9.817 ∗ 10⁻² 2.298 2.182 ∗ 10⁻² Other −2.580 ∗ 10⁻¹ 7.410 ∗ 10⁻² -3.481 5.261 ∗ 10⁻⁴

The final pricing model for single malt whisky was thus:

log(P rice) = 6.6795555 + AlcoholContent ∗ 0.0711102 + AlcoholContent²∗ (−0.0023791) + Islay ∗ 0.1052422 + Campbeltown ∗ 0.2255714 + Other ∗ −0.2578951

(27) 4.5 Final model validation

4.5.1 Residual diagnostics

Figure 5: Scale Location plot for the Final Model

The Scale Location plot is similar to the one for the first model. The residuals versus fitted values forms a relative straight line without any major discrepancies. This indicates that the residuals indeed are homoscedastic.

27

Figure 6: Residuals versus fitted values for the Final Model

The plot of residuals versus fitted values now shows a more straight line which indicates that the price was not linear dependent on the alcohol content and that the variable, squared alcohol content, improved the accuracy of the model.

Figure 7: Normal QQ-plot for the Standardized residuals, Final Model

The Normal QQ-plot for the standardized residuals has changed some from the one corresponding to the initial model. The left tail is now almost 45 degrees but the right tail deviates more from the normal distribution. This implies a somewhat skewed distribution.

An even higher logarithm of price was tested but did not improve the model. As the plot is still fairly straight the assumption of standard normally distributed residuals is kept.

4.5.2 F-statistic and p-value

The final model have a small p-value of less than 2 ∗ 10⁻¹⁶and a F-statistic of 20.11. The p-value is smaller than the initial model, and the F-statistic larger. This indicates that the new model is more significant than the initial model. The individual p-values for our covariates are all below our significance level of 5% which indicates that they should all be included in the model.

29

4.5.3 R² and R² adjusted

The R² value for the model is 20.6%, stating that the covariates explains approximately 21% of the response variable variance. Adjusted R² is 20.1%. Both R² and adjusted R² has imporoved from our initial model, which again indicates that the final model is an improvment compared to the initial model.

4.5.4 VIF-test

Table 8: Table of VIF-test values for the covariates in the final model Coefficients VIF

Alcohol Content 15.415147 Alcohol Content² 15.416791

Islay 1.014715

Campbeltown 1.008009

Other 1.021471

The variance inflation factor test indicates that there exists an multicollinearity be- tween Alcohol Content and Alcohol Content². This is natural due to their mathematical connection. The other variables have a VIF values around 1 which does not indicate any multicollinearity. Multicollinearity in the final model is hence neglected.

5 Discussion and conclusion

5.1 Analysis of covariates in the Final model

The final model consists of the covariates alcohol content, and the dummies Islay, Campel- town and Other, which were also included in the initial model. Furthermore, we added the covariate alcohol content² in the final model.

5.1.1 Alcohol content

The alcohol content has an estimate of 0.07 and a high significance in the model. One explanation of the alcohol contents positive impact on the price could be the alcohol taxes in Sweden. This is calculated as 511.48 ∗ V olume ∗ %of Alcohol. To check if this was the explanation of the Alchol content significance we investigated the Final model when taking (P rice − 511.48 ∗ V olume ∗ %of Alcohol) as response variable. When doing this there was no major change in the significance level for alcohol compared to the Final model, on the contrary the p-value decreased slightly to 4.21 ∗ 10⁻¹⁰ and the estimate increased slightly to 7.310⁻² We can hence conclude that people pay more for a whisky with a higher alcohol content and that its impact in our regression model not is due to alcohol taxes in Sweden.

The squared covariate is negative which means that the impact is declining.

5.1.2 Islay

Among the regions it ware Islay, Campbeltown and Other that made it to the final model.

Islay has an estimate of 0.11. As mentioned in 3.1.2 Islay only have nine active distilleries and is a very small whisky region in Scotland. The region has a characteristic smokey taste. It it thus not surprising that bottles from Islay are more expensive, as it comes from nine exclusive distilleries. Furthermore, whisky drinkers might fancy the taste, which makes them willing to pay more for the whisky from Islay.

5.1.3 Cambpeltown

Campbeltown has an estimtae of 0.23 and as mentioned in 3.1.2 the region has even fewer distilleries than Islay. Only 3 are remaining today. By size the region is about the same as Islay. The smaller range of distilleries is also reflected in the covariate’s impact on the price. Campbeltown has a positive impact on the prices. Whisky from Campbeltown also have significant characteristics, including a defined dryness with pungency, smoke and a solid salinity.

5.1.4 Other

As explained in 5.1 whisky from the region Other is whisky from unspecified regions. The estimate of Other is -0.26 and the significance is the highest among the regions.

31

5.2 Discussion and conclusion of mathematical model

The final model of the regression analysis shows the most important factors affecting the price of a single malt whisky, which was the purpose of this thesis. In the initial hypothesis all regions were included but as the variation within the larger regions was large those were removed. Both Islay and Campbeltown have positive impact on price, but Campbeltown has twice as much impact as Islay. Whisky from both Islay and Campbeltown have specific characteristics and taste. Furthermore, these are the two smallest regions with only a few distilleries, resulting in an equivalent pricing within the region.

It is concluded that the origin have a positive impact on price and that whisky from smaller regions with fewer distilleries are more expensive. This is probably due to these regions are considered to be more exclusive as they have fewer distilleries than the other regions.

Furthermore, whisky from unspecified regions, Other, has a negative impact on price, with the highest significance level among the regions. This strengthens the hypothesis that re- gions with specific characteristics, limited number of distilleries and a long history have higher prices. Whisky from unspecified regions, Other, does not have any specific taste and is thus not considered to be as exclusive.

The regions not included in our model, Highlands, Lowlands, Speyside and Islands are big regions with many distilleries. Even though whiskys from Highland, Speyside and Lowland have characteristic taste the production differ within each region and the distil- leries may have different pricing strategies. Hence the variation in price for each region is large and therefore these regions are not significant for the model. Islands have whisky with a great variation in taste as each island has its own characteristics, which implies a great variation in price as well. As the big regions are excluded in the final model it does not give a true picture of the price drivers, which is seen in the low R².

Storage time was believed to have a significant impact on the price, but instead this co- variate had the highest p-value and was the first to be eliminated. The maturation gives a whisky its specific taste. Since different casks give different tastes this might have a larger impact on the price than the maturation time itself. Some distilleries have also introduced different woods into the maturation process which results in a discernible taste.

This special maturation is not explained by the covariate Storage Time but might have a big impact on the price.

An idea for future studies would be to introduce a covariate that explains the type of cask used for the maturation. It would also be interesting to include all types of grains in the model to see if the price of single malt whisky differs from other grain whiskys. These additional variables require excessive research though, since the casks used are rarely stated and the file from Systembolaget only included a few non-malt whiskys. The problem with

blended malt whisky still remains as this whisky comes from several distilleries and re- gions. To analyse the regions and brands even further a regression on the brands within each region could result in useful observations on the impact of a specific brand, but this was not included in the scope of the project. An additional factor that could improve the model is one explaining the number of bottles produced since rarity will increase the price and this can harm the model.

33

6 Whisky in Scotland from a marketing perspective

6.1 Introduction

Our mathematical model shows a relation between the origin of the whisky and the price.

Whisky from Islay and Campbeltown have a strong positive relation to the price. It is hence interesting to examine whisky from these regions from a marketing perspective, using our results as a case study. We will analyse whisky from Scotland by using the Marketing Mix and the Four P and Cs. After this we will immerse the marketing study by performing a Porter’s Five Forces Analysis on whisky from Islay and Campbeltown. Lastly a summaize of the findings will result in a marketing strategy recommendation for companies producing whisky in Islay and Campbeltown.

6.2 Marketing Mix

The marketing mix refers to the set of actions, or tactics, that a company uses to promote its brand on product in the market. The 4Ps make up a typical marketing mix - Price, Product, Promotion and Place. Marketing mix decisions should be made to influence the trade channels as well as the final consumers. Winning companies are those that meet customer needs economically with effective communication. The four Ps were introduced by E. Jerome Mc Carthy. Robert Lauterborn later suggested that the sellers’ four Ps cor- respond to the customers’ four Cs; Consumer wants and needs, Cost, Communication and Convenience respectively. [12]

The 4Ps and Cs are explained and applied on the whisky industry in Scotland.

6.2.1 Product & Consumer wants and needs

Product refers to the item actually being sold. The product must deliver a minimum level of performance and satisfies what a consumer demands. Marketers should consider how to position the product and exploit the brand and how to exploit the company’s resources and how to configure the product mix so that each product complements the other. A marketer must also consider product development strategies. Consumer wants and needs refer to that the company will only sell what the consumer specifically wants to buy. Hence customers should be studied to create a product that should be sold. [13]

The product in our thesis is the non-essential product whisky and it is positioned to the middle class. Depending on the consumer (a whisky expert and collector or main stream drinker) it is hard to substitute whisky and a specific brand. Many of the whisky brands are old and have a long history, and as shown in the result for our pricing model the origin plays a big part when pricing the products. The history and the brand plays a big part as this reflects the quality and taste. It is not a product you face and buy everyday, and a

bottle lasts for a long time depending on drinking behaviour. If the whisky brand is strong enough it is possible to create products accessories around the whisky such as glasses for example. Other examples of product development is to release Limited Editions, which also can be viewed as an opportunity to further strengthen the characteristics and market the brand. In our model we found that whisky from Islays and Campbeltown are more expensive than from other regions. This should be emphasised and taken into account when releasing further products. From our model, where we found that the origin plays a big role, we believe that it is important to keep the history and the exclusiveness of the brand. The companies should therefor be careful if they introduce new products.

6.2.2 Price & Cost

Changes in price will effect the demand and the sales depending on the price elasticity of the product. The marketer should set a price that complements the other elements of marketing mix. It is important that the marketer is aware of the price elasticity and the customer perceived value for the product. From a consumer perspective price is not only nominated in monetary terms but also in cost of time in acquiring the product or alternative cost for not buying another product. [13]

Whisky is a non-essential product, and depending on the luxury and history of the prod- uct it has different price elasticity. In our model we found that whisky from distilleries in Islays and Campbeltown are significant higher than from other regions. This indicates that whisky from these regions have a larger price elasticity. Many whisky consumers are experts and have whisky as one of their interests. Hence we believe that whisky with char- acteristic taste and strong historical background and brand have a relatively high price elasticity. As whisky is a non-essential product it will be be dismissed due to cost of quilt if the alternative was to buy essentials, such as food for the family.

6.2.3 Promotion & Communication

This refers to all the activities undertaken to make the product or service better known to the user and trade. Promotion is often divided into advertisement, public relations, personal selling, sales promotion and direct marketing. Promotion is viewed as a one way communication from the seller to the buyer, a form of manipulation, as communication aims to create a dialogue with the potential customers based on their needs and lifestyles.

[13]

As the history and origin is important we believe that promotion should be done in a careful and sophisticated way to remain the exclusive feeling associated with the brand.

Promotion is mostly done by advertisement and personal selling today, to make sure that the best retailers sell the whisky. Other promotion could be through having showings

35

of distilleries. Communication is important for the whisky brands as it not only sell the whisky but the whole brand, the heritage and authenticity and the lifestyle around it. For other some blended whisky brands such as Jack Daniel’s and Grant’s we have seen big sucessful marketing campagins focusing on this.

6.2.4 Distribution & Convenience

This refers to providing the product at a place which is convenient for consumers to access.

Distribution has become more and more replaces by the term convenience, as it today is important not only to be available but easy and convenient to find, buy, deliver amd several other factors. [13]

It is important for the whisky producers to make sure that their whisky is available on the best retail online sites. Many whisky drinkers collect whisky via Internet and hence availability here is important. Furthermore it is important to be in the assortment of the biggest whisky stores, or in cases such as Sweden where there exists a alcohol monopoly, to be available at Systembolaget.

6.3 Porter’s Five Forces Analysis 6.3.1 Theory

The business environment is in general rapidly changing these days and to investigate the implications of the changes Porter’s five forces analysis is a useful framework. The external environment that an organization is working in is of big importance and therefore the competitors’ activities are taken into account in Porter’s five forces analysis. Porter’s model determines the attractiveness of the market the business is operating in by analysing the competitive intensity.The following five forces are use to identify potential opportunities and risks:

1. Competitive Rivalry: The competition increases with number of market players, slow market growth, high fixed cost, high storage costs and high exit barriers. It decreases if switching costs and levels of product differentiation are low.

2. Threat of New Entrants: The number of potential new entrants depends on the bar- riers to entry, including internal economies of scale, governments, asset specificity, patents and proprietary knowledge.

3. Threat of Substitutes: The threat of substitutes increases when the substitute product is cheaper than industry product and when consumer switching costs are low. If the substitute product’s quality and performance is equal or superior the industry product this also implies an increased threat.

4. Bargaining Power of Suppliers: The bargaining power of suppliers increases with few suppliers, if switching costs are high, if the buyer is not price sensitive, if the supplier’s

product is highly differentiated or if substitute products are unavailable.

5. Bargaining Power of Customers: The bargaining power of customers increases when concentration of customers is high, switching costs are low, customers are price sensi- tive and well informed about the product, product differentiation is small and when the threat of backward integration is high. [?].

The first three forces are operating in the same direction within the market and are therefore considered ”horizontal” factors. The two last forces are instead classified as ”ver- tical” as they operate within the supply chain.

Figure 8: Porter’s Five Forces [14]

6.3.2 Analysis of Whisky from Islay and Campbeltown

Porter’s five forces is considered to be used for a set of highly related products for a spe- cific business need and is hence adequate for the analysis of whisky from both Islay and Campbeltown, to investigate their strategic positions. As the final regression model shows that the price of whisky is significantly higher for whiskys from Islay or Campbeltown than from other regions, the following analysis aims to discuss the underlying factors causing this price premium. This will be used to determine the marketing strategy for whisky producers from Islay and Campbeltown.

Competitive Rivalry

37

The competitive rivalry is high in the whisky industry as there are many players in the mar- ket and the product differentiation is low. The market is mature and has a relatively slow growth. For whisky distilleries on Islay and in Campbeltown there are a limited number of players though and hence the rivalry is lower there. Storage costs are very high for whisky production with its long maturation on casks, which also implies a great competition.

Threat of New Entrants

The threat of new entrants is low due to the high barriers to entry and the importance of tradition and history. Whisky production requires long production time, high initial costs of stills, casks etc. and knowledge of whisky. In addition the regulations are very strict in the whisky industry as well as the taxes.

Threat of Substitutes

The threat of substitute is medium since there are several substitutes such as other spir- its but none that exactly substitutes whisky with its special taste, history and traditions.

For whisky enthusiasts there are barely any substitute but for other consumers whisky is competing against gin, liquor, cognac and rum, which sometimes are cheaper. Even Champpagne is seen as a substitute in terms of celebration and sharing.

Bargaining Power of Suppliers

The bargaining power of suppliers is low since the supply of barley and other grains is large and the product is not very differentiated. The number of suppliers is relatively big and the switching cost to another supplier are low, resulting in a low bargaining power.

Bargaining Power of Customers

The bargaining power of customers is high for customers without much knowledge about whisky since they are price sensitive and the switching cost for them is low. For whisky enthusiasts on the other hand, the bargaining power is lower as their price sensitivity is lower and their loyalty high. These customers more frequently buy single malt whisky from Islay and Campbeltown. Overall, the buyers bargaining power is relatively low as the number of buyer is big and very few customers can impact the price.

To sum up, the attractiveness of the overall whisky market is relatively high but the competition is big. Unlike other more fast moving markets the whisky industry is stable, sustainable and slowly moving. Hence the above analysis is assumed to hold for several years and the market is not considered to change a lot.

6.4 Recommendation of marketing strategy for producers in Islay and Campbeltown

The mathematical model and our analysis above both conclude that the history, origin and brand is important for whisky producers in Islay and Campbeltown. It is thus impor- tant to emphasize this in the marketing. Factors such as price is not as important for the consumers as the single malt whisky has high price elasticity. Price should thus not be included in the marketing of single malt whisky from Islay and Campbeltown.

Single malt whisky from these areas differs from other whisky such as blended. Blended whisky brands have had many intense marketing campaigns such as Grant’s and Jack Daniel’s, focusing on extending its reach beyond that of its traditional drinkers to capture potential new consumers. As the global whisky market is big and many whiskys from the U.S. for example (which are not included in this thesis), already have a lot of commercial marketing this is not an attractive strategy for Scotch whisky.

For single malt whisky from Islay and Campbeltown, with their characteristic taste and traditions, the whisky is preferred by people who have a strong interest in whisky. Thus we do not believe intense marketing campaigns on television is suitable. This could hurt the brand image of being an exclusive whisky. Furthermore, it is important for the whisky to restrict the marketing to the guidelines set by The Scotch Whisky Association.

The Scotch Whisky Association are stating that the marketing not should encourage ex- cessive drinking. Furthermore, it states that the companies should encourage responsi- ble consumptions and that events such as tastings should broaden such customers edu- cation around whisky and the brand, and target customers who wish to taste different whiskies.[15]. We believe that tasting under these circumstances, with an educational starting point, is a good way for companies to market their brand.

Given the limited number of distilleries and the high barrier of entry for competitors to enter the Islay and Campbeltown market of whisky we believe this is a sustainable market- ing strategy in the long term. The companies should emphasise their history and discover new tastes and products by for example releasing Limited Editions.

39