University of Groningen Faculty of Economics and Business
Persistence modeling with fast moving consumer goods
To what extent do advertising and price result in persistent and/or immediate sales effects for hedonic and utilitarian fast moving consumer goods?
By
Reuven Yang-Sam Ayal
Persistence modeling with fast moving consumer goods
To what extent do advertising and price result in persistent and/or immediate sales effects for hedonic and utilitarian fast moving consumer goods?
By
Reuven Yang-Sam Ayal
University of Groningen Faculty of Economics and Business
Management summary
This study examines to what extent advertising and price result in persistent and/or immediate sales effects for hedonic and utilitarian fast moving consumer goods. Furthermore, the moderating effect of brand size is also taken into account as previous studies had shown the difficulty of finding unit roots. There are four different types of advertising that can be distinguished in this study, namely advertisement by the manufacturer, feature only, display only, and the combination of display and feature advertising. The price variable used is a price index. The relationship of these variables is tested by the use of impulse response functions which have been executed on the respective VAR model. Positive relationships with sales volume is the result for the advertising variables, and a negative relationship with sales volume was found for price on the individual brand level. Price seems to be more effective for utilitarian products, as well as the combination of display and feature advertising. Hedonic products react on average more strong to advertisement by the manufacturer and display only advertising. It is important to note that the power of the effects may differ quite extensively among different brands and categories. Persistent effects were tested using the same type of VAR models; however, the sales variable should have a unit root. The results had the same valence as with the analysis of immediate effects. Also, advertising effects on sales seem more effective for small brands; however, further research would be required.
Table of Contents
1. Introduction ... 1
2. Literature review ... 2
2.1 The marketing mix ... 3
2.2 Hedonic and utilitarian products ... 6
2.3 Persistent and transient effects on sales ... 7
2.4 Conceptual model ... 9
3. Research design ... 9
4. Results ... 12
4.1 Unit root tests ... 12
4.2 VAR models ... 13
4.3 Impulse response functions ... 15
4.4 Persistent effects ... 18
5. Discussion ... 19
5.1 Immediate effects ... 20
5.2 Comparing hedonic and utilitarian products ... 21
5.3 Persistent effects ... 22
6. Conclusion ... 24
7. Managerial implications and limitations ... 26
7.1 Managerial implications ... 26
7.2 Limitations and avenues for future research ... 27
References ... 28
Appendices ... 32
Appendix 1 Product categories ... 32
Appendix 2 Unit root tests hedonic FMCG’s ... 32
Appendix 3 Unit root tests utilitarian FMCG’s ... 40
Appendix 4 Impulse response function coefficients tables hedonic brands ... 49
Appendix 5 Impulse response function coefficients tables utilitarian brands ... 51
1
1. Introduction
Accountability in marketing has increased over the last couple of years, as data insights help managers to take better decisions. Scanner data can provide this possibility to gain insights, making it possible to analyse marketing expenditures and sales output. As the focus of marketing is on long-term benefits for a company, gaining insight is important to monitor what is going on and how marketings efforts contribute to sales. Next to the type of effect, the streng and length of the effect can be analysed. Persistent effects are permanent, hence a change in base sales from the period before the marketing instrument has been used. Transient effects have a short-term influence on sales (Osinga, Leeflang and Wieringa, 2010). Insights on the effects of marketing stimuli such as advertisement and price are for these reasons interesting for managers and the companies they work for (Verhoef and Leeflang, 2009).
The retail environment knows a lot of competition and marketing activity in terms of advertisement and price. Gaining more insight on whether these marketing efforts impact buying behaviour of consumers for fast moving consumer goods (FMCG’s) would therefore be interesting. This environment of FMCG’s is mainly considered as the supermarkets. Supermarket chains promote several products over a certain period of time, but what is the effect of these promotions and do they also have a long-term effect on sales? Types of FMCG’s that are sold can be distinguished as either being hedonic or utilitarian (Price, Feick and Higie, 1989). These two product types are different in terms of the product feature levels and what consumers prefer. This current study focusses on the differences in terms of marketing effectiveness on sales volume, and tries to find any differences between these two types of product. This topic relevant for marketing managers as it will research the influence of certain marketing tools. Furthermore, more can be learned from how sales from hedonic and utilitarian products might react to different marketing stimuli. Results can assist marketing managers to make better decisions in terms of the marketing budget allocation.
2
sudden change in sales is permanent or not, followed by a VAR model to determine the strength of this effect and to what extent this can be accounted for by marketing efforts (1995). The research question that is posed in this current paper is as follows:
To what extent do advertising and price result in persistent and/or transient sales effects for hedonic and utilitarian fast moving consumer goods?
The dataset used consists of supermarket data of the Netherlands from 1994 to 1998. The economic environment in this time frame showed favorable aspects, with an average economic growth of 3 percent and a rising employment rate over time (Rijksbegroting, 2002).
This paper starts with a literature review where scientific articles are used to address what is already known about this topic, and upon these insights several hypotheses are formulated. The methodology section will follow with a research design explaining how the research question and hypotheses will be approached. Next, the results of the study will be presented followed by a discussion and conclusion. Last, some managerial implications will be presented and limitations and avenues for future research will be discussed.
2. Literature review
3 2.1 The marketing mix
The marketing mix instruments that are taken into account are price and advertisement. Price deals with the actual price of the product and advertisement is split up in several types of advertising. Advertising in general is used to stimulate sales, this can be direct or indirect, and most of the times it has a short term design for quicker and larger quantities of product purchases (Zia & Shahzad, 2015). Ailawadi and Neslin state that advertisement induces consumers to consume more and faster (1998). Sethuraman et al. performed a meta-analysis on advertising elasticities, and state that the elasticities are higher for durable goods compared to nondurable goods, and that they are higher for newer brands than for older brands (2011). These statements are a brief display of the research that has already been done on advertising. The different types of advertisement accounted for in this study are: Advertisement by the manufacturer, feature only, display only, and the combination of display and feature advertising.
Advertisement by the manufacturer
‘Advertising refers to any form of paid communication by an identified sponsor aimed to inform and/or persuade target audiences about an organization, product, service or idea’ (Fennis & Stroebe, 2010). The identified sponsor here is the manufacturer of the product itself. This type of advertising could for example be a commercial on television, radio, billboards, posters and so on. Advertising has the functions to inform and to persuade, and is powerful as it can affect consumers’ beliefs, attitudes and behaviour towards a brand or even a whole product category (Vakratsas & Ambler, 1999; Fennis & Stroebe, 2010). However, advertising cannot create needs, but is able to translate the need of a customer to a want, which a manufacturer can provide (Kotler, 1997). Advertising can establish habits, which are desirable because of the cognitive lock-in it creates; however, a habit to buy the product of a competitor is of course not desirable (Johnson, Bellman & Lohse, 2003). This means that advertising can help to make purchases of a certain product become habitual. Consumers will be more likely to keep buying those products resulting in loyal customers to that brand. As of these reasons, advertisement shows to be able to attract customers towards the advertised product, and therefore result in a positive relationship with sales.
4
Feature advertising
Cameron and Haley describe feature advertising as printed advertisements looking like a short feature article (1992). They can be communicated by for example newspaper inserts and store leaflets (Bodapati and Srinivasan, 2006), and are placed by retailers to inform consumers about their promotions (Pieters, Wedel & Zhang, 2007). Therefore, consumers are able to see the promotion before they go to the store. Feature advertising can also be seen as a sort of cooperation between the retailer and the manufacturer. Manufacturers pay retailers to get their products featured, and retailers can attract consumers to their stores and increase sales (Pieters et al., 2007). Feature advertising is presented in such a way that it makes the distinction between advertising and editorial material hard to see, and might therefore not directly be interpreted as an ad by the reader (Cameron & Haley, 1992). Furthermore, feature advertising is considered as a non-emotional or a rational message; it contains practical information for the consumer (Janssens & De Pelsmacker, 2005). The desired outcome would be that consumers acknowledge the promotion and add the promoted product to their consideration set (Zhang, 2006). This means that consumers will consider buying the brand when they are a point of purchase. Allenby and Ginter state that feature advertisement is important in the formation of a consumer’s consideration set (1995). Strong evidence was found that feature advertising affects sales in the short run (Zhang et al., 2009), as it reduces the risk consumers relate to a product (Ailawadi & Neslin, 1998).
Hypothesis 1b: Feature only advertising has a positive immediate effect on sales of fast moving consumer goods.
Display advertising
5
for the average item, which was on a display, could get up to four hundred and seventy-three percent of normal sales (1975).
Hypothesis 1c: Display only advertising has a positive immediate effect on sales of fast moving consumer goods.
Hypothesis 1d: The combination of display and feature advertising has a positive immediate effect on sales of fast moving consumer goods.
Price
Price is an important factor in a consumer’s buying decision. A reduction in price, or price promotions are also frequently used by Dutch supermarkets. In fact, the most frequently used type of promotion is a price promotion and twenty-four percent of all purchases take place under some promotional support (Dekimpe et al. 2005). When these price promotions are used frequently, consumers are less likely to purchase that product when it is not on promotion, and rather wait for that product to be on promotion again. This leads to less frequent purchases, but an increase in the amount purchased at one time (Van Heerde et al., 2013). Hence, price promotions do little for brand building (Ataman, Van Heerde & Mela, 2010). After a promotional sales peak, a post promotion dip is expected, although Van Heerde, Leeflang and Wittink rarely found any of those dips at either the category or brand level (2000). The effect of a price promotion only holds on for two to eight weeks, and sales immediately increase as consumers want to make use of the promotion (Pauwels, Hanssens & Siddarth, 2002; Sethuraman, Raj & Srinivasan, 2002). Hence, price discounts might help to attract new buyers to the category or brand. Furthermore it is stated that price discounts only have temporary benefits for established brands, as most consumers have bought it already and have had their first experience with it, and therefore the learning experience is limited (Pauwels, Hanssens & Siddarth., 2002).
6
increased (2010). Consumers also refer to previous purchases. Reference prices are the comparison consumers make between the price they have to pay now and what they have paid last time. This is also important to consumers, as when they notice a price increase it can have negative effects on the purchase probability for that moment (Kan et al., 2014).
Hypothesis 2: Decreasing price has a positive immediate effect on sales of fast moving consumer goods.
2.2 Hedonic and utilitarian products
7
Hypothesis 3a: Immediate effects caused by advertisement by the manufacturer are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
Hypothesis 3b: Immediate effects caused by feature only advertising are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
Hypothesis 3c: Immediate effects caused by display only advertising are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
Hypothesis 3d: Immediate effects caused by the combination of display and feature advertising are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
Hypothesis 4: Immediate effects caused by increasing the price are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
2.3 Persistent and transient effects on sales
8
three percent (2008). Therefore, it seems more likely to find persistent effects with smaller brands than large brands.
Ataman, Van Heerde and Mela say that advertising positively affects base sales, but price discounts have a negative effect (2010). In addition, as was mentioned earlier, price discounts only have temporary effects of about two months (Pauwels et al., 2002). The assumption that advertising could result in persistent effects on sales, is therefore more likely than effects of price. And as the economic environment of the data points can be considered as favourable (Rijksbegroting, 2002), advertising elasticities tend to increase and price sensitivity tends to decrease (Van Heerde et al. 2013); therefore, the effect of the advertising variables is more likely to be assumed than the effect of price. In combination with the greater likelihood of small brands to have persistent effects on sales, and the theory discussed for the marketing mix, the same relationships for advertising variables are assumed for persistent effects. Therefore the following hypotheses have been created.
Hypothesis 5a: There is a persistent positive effect on sales caused by advertisement by the manufacturer for fast moving consumer goods of small brands.
Hypothesis 5b: There is a persistent positive effect on sales caused by feature only advertisement for fast moving consumer goods of small brands.
Hypothesis 5c: There is a persistent positive effect on sales caused by display only advertisement for fast moving consumer goods of small brands.
Hypothesis 5d: There is a persistent positive effect on sales caused by the combination of display and feature advertisement for fast moving consumer goods of small brands.
9 2.4 Conceptual model
3. Research design
This chapter will outline the design of this study. To test the conceptual model, the research will consist out of two main parts. First a search for unit roots in the sales variables, followed by an analysis to compare immediate and persistent effects. The data that has been used is from Dutch supermarkets, which is aggregated on the level of the whole market. Data points include the period of week 29 in 1994 up to and including week 28 in 1998, adding up to two hundred and eight observations. The product distinction between hedonic and utilitarian products has been acquired by adding results from a UK survey to the dataset. Twenty FMCG product categories have been chosen to include in the research with an even split between hedonic and utilitarian products. To distinguish between small and large brands, the market share of each brand was taken at the beginning of the observation period. Four brands were selected from each product category; two large brands (>10% market share) and two small brands (<3% market share) (Slotegraaf and Pauwels, 2008). By this manner both moderators,
10
product type and brand size can be accounted for in the analysis. An overview of the initially selected product categories and brands can be found in appendix 1.
The first part of the analysis considers the search for persistent effects on the sales variable of each brand and is descriptive. The analysis tries to find out if sales are stable or not by using the unit root concept. This means that if the sales return to the overall average, no permanent effect can be addressed. The sales variable of every brand has been transformed into logarithms, followed by the Augmented Dickey Fuller test to search for unit roots. In order to accept the presence of a unit root, the outcome of the Augmented Dickey Fuller test should have a probability that exceeds 0.05 in the situation where the trend and intercept are included in the equation. After the unit root analysis for the sales variable, all other variables are transformed in logarithms and also checked on unit roots. The same Augmented Dicky Fuller test is performed on those variables. To give a clear overview of the combination of variables researched, see table 1 below.
Dependent variable
Independent variables Amount of analyses
Sales volume
Price index, Feature only, Display only, Display and feature, Advertisement
Hedonic Large brand 20 Hedonic Small brand 20 Utilitarian Large brand 20 Utilitarian Small brand 20
Table 1 to be tested relationships
11
A requirement was set by the researcher, that only VAR models would be created for brands that have sufficient data for all variables. The output of a VAR model provides multiple equations with each variable being a dependent variable in one of these questions; however, only the equation with sales volume as dependent variable will be considered to evaluate as this is the only relation that is researched in this study. An overview of the variables included is given below. [ 𝑆𝑖𝑡 𝑃𝑖𝑡 𝐹𝑖𝑡 𝐷𝑖𝑡 𝐷𝐹𝑖𝑡 𝐴𝑖𝑡 ] = ∑ [ 𝑆𝑖𝑡−𝑘 𝑃𝑖𝑡−𝑘 𝐹𝑖𝑡−𝑘 𝐷𝑖𝑡−𝑘 𝐷𝐹𝑖𝑡−𝑘 𝐴𝑖𝑡−𝑘] 𝒏 𝒌=𝟏 + [ 𝑄1 𝑄2 𝑄3 𝑇𝑟𝑒𝑛𝑑 ]
𝑆𝑖𝑡= Sales volume sold of the fast moving consumer good per brand type and size i at time t
Pit= Price index per brand type and size i at time t
Fit=Weighed distribution figure for a ‘feature-only’ promotion per brand type and size i at time t
Dit=Weighed distribution figure for a ‘display-only’ promotion per brand type and size i at time t
DFit=Weighed distribution figure for a ‘Display and Feature’ promotion per brand type and size i at time t
Ait=Weighed distribution figure for Advertisement by the manufacturer per brand type and size i at time t
Q1, Q2, Q3=Indicator for the first, second, or third quarter of a year with the fourth quarter
as benchmark
Trend= Ordering indicator that goes up by one every week
12
effects in specific weeks and the corresponding standard errors. From this information t-statistics can be calculated followed by a calculation to p-values. Then significant effects can be interpreted. Also, an added z test will be performed to be able say something about the significance of the total of all brands and a difference between hedonic and utilitarian products.
4. Results
This chapter provides the results from the analyses that have been described in the research design. First the results from the unit root tests will be given, which is about the stationarity of the sales data. Next the output from the VAR models and impulse response functions executed on the VAR models will be presented. The impulse response function output will be discussed for hedonic and utilitarian products separately, and finally persistent effects will be presented.
4.1 Unit root tests
From forty brands in ten hedonic product categories, nine had a unit root in their sales. The brands belong to the categories of Apple juice, Cornflakes, Crème rinse, Custard desserts and Popcorn. Almost all of these are small brands, except for Domo, which is considered a large brand. Three brands from the Crisps category (Bahlsenpicanterie, Croky Bubble chips and Croky Vakantiepak), and two brands from the Mustard category (Amora and Colmans) could not be used as they did not have sufficient sales figures, and therefore have been excluded from the research. This left thirty-five hedonic brands to continue with.
13
Product category Brand Brand size
Apple juice AK Small
Cornflakes No Brand Small
Old Farmer Cornflakes Small
Crème rinse Andrelon Cream Care Small
No Brand Small
Custard desserts Domo Large
Domo Dessertvla Small
Domo Lang Lekker Small
Popcorn Borden Crackerjack Small
Table 2 Unit roots in Sales of Hedonic FMCG's
Product category Brand Brand size
Cat food Belamy Large
Whiskas Small
Kazan Small
Coffee milk Menkomel Small
Ketchup Heinz Small
Liquid soups in tins & cartons Becas Large
Campbells Tastes Of The World Small
Razor blades Gillette Large
Gillette gii plus Small
Spaghetti Epifine Small
Tea Duikertjes Small
Yoghurt Coberco Large
Danone Small
Fage Small
Table 3 Unit roots in Sales of Utilitarian FMCG's
4.2 VAR models
14
from the research. Large brands did not have sufficient data for the advertisement by the manufacturer variable in fourteen cases, sufficient information for the feature only variable was lacking in one case, and the same problem occurred twice for the combination of display and feature variable. Small brands had insufficient data for the same type of variables: twenty-four times for the advertisement by the manufacturer variable, seven for feature only, and fourteen for the combination of display and feature.
The brands which could be used in the analysis can be found in table 5 and 6 below. The adjusted R² is given for the equation with Sales as the dependent variable. The stability condition should be met as the VAR models were created with stationary data; however, a double check was performed by consulting the modulus of all variables as this had to be < 1. In terms of lag structure for the VAR models, all brands required one lag according to the Bayesian Information Criterion (BIC).
Product category Brand Brand size Adjusted R²
Apple juice Euroshopper Large 0.300
Cornflakes Pablo pepe Large 0.344
Crème rinse Nivea Large 0.191
Euroshopper Large 0.418
Custard dessert Naarmann Large 0.320
Instant soup Gasa ayam special Large 0.232
Mamee Large 0.144
Rice Oryza Small 0.326
Shampoo Selsun Large 0.247
Alberto vo5 Large 0.421
Crinopex Small 0.399
Dicosmetics Small 0.596
Table 4 VAR models of Hedonic FMCG's
Product category Brand Brand size Adjusted R²
Cat food Belamy* Large 0.772
Kazan* Small 0.303
Coffee milk Campina Large 0.419
Instant coffee Avogel bambu Large 0.621
Grandos Small 0.792
Ketchup Woodmans Large 0.153
15 Liquid soups in tins & cartons Becas* Large 0.299
Menziterrineculinaire Small 0.365
Razor blades Bic Large 0.843
Spaghetti Antonio deniro Large 0.215
Tea Kanis Gunnink Large 0.423
Van Nelle Twee Thee Small 0.640
Toilet tissue Pura Ninfea Large 0.013
Eldou Small 0.251
Yoghurt Coberco* Large 0.172
Danone* Small 0.147
Table 5 VAR models of Utilitarian FMCG's (*unit root in the sales variable)
The utilitarian brands Belamy and Kazan from the Cat food category, Heinz from the Ketchup category, Becas from the Liquid soups category and Coberco and Danone from the utilitarian Yoghurt category are the six brands that displayed a unit root in their sales and had enough observations for all variables to create VAR models. All other brands for which VAR models were created had stationary sales data.
4.3 Impulse response functions
The results from the impulse response functions (IRF) will be presented next. First each variable of the hedonic product categories will be discussed separately, followed by the same structure for utilitarian products. To show the interpretation of the IRF output, an example is given; one standard deviation shock in the price index of brand A leads to a negative effect of 5.50% in the sales volume of brand A, while keeping all other variables constant. It is important to note that the effects in the sales volume are caused by a shock of one standard deviation in the independent variable, and this standard deviation is different for every brand and every variable. For this reason it is harder to say that a higher effect on the sales volume means that the independent variable has an absolute stronger effect compared to the other; however, in order to make comparisons and interpret these, this will be tolerated.
Hedonic
16
coming effects from the IRF and will therefore not be repeated every time as complete. The effect of Pablo Pepe was the smallest effect, and the largest effect was for Di Cosmetics -23.35%. The unweighted average effect of the price index on sales volume is -12.58% and a median of -12.66%. The effect of feature only advertising on sales volume is positive in all cases and results for all brands were significant (unweighted average: 9.62% and median: 7.15%). The lowest magnitude of one standard deviation shock is for Naarmann (11.10%) and the highest for Mamee (19.13%). Display only advertising showed ten brands with significant results, and all had positive effects (unweighted average: 7.22% and median: 5.69%). The lowest magnitude is for Euroshopper Apple Juice (2.33%) and the highest influence for Di Cosmetics Shampoo (12.76%). The combination of display and feature advertising had eleven significant positive results, (unweighted average: 12.70% and median: 10.22%) with the lowest magnitude for Euroshopper Crème Rinse (7.53%) and the highest influence for Mamee Instant Soup (20.63%). The effect of advertisement by the manufacturer on sales volume had an immediate effect, which was significant for just two brands of one category; Crinopex and Di Cosmetics from the Shampoo category. Crinopex had the lowest magnitude of effect (2.07%) and Di Cosmetics the highest (5.43%). In terms of long-term effects caused by the marketing effort done at week 0; at week fifty-two, none of the variables significantly affected the sales. Below, table 6 gives the overview of the range, unweighted average and median.
Variables Minimum Maximum Median Unweighted average
sig Positive Negative
Price index -0.0406 -0.2335 -0.1266 -0.1258 11/12 0 11 Feature only 0.0111 0.191336 0.0715 0.0962 12/12 12 0 Display only 0.0233 0.1276 0.0569 0.0722 10/12 10 0 Display and feature 0.0753 0.2063 0.1022 0.1270 11/12 11 0 Advertisement by
the manufacturer
0.0207 0.0543 0.0375 0.0375 2/12 2 0
Table 6 Coefficients of IRF from hedonic brands
Utilitarian
17
example will be given of the interpretation of the output. One standard deviation shock in the price index of Woodmans Ketchup leads to an immediate negative effect of 2.45% in the sales volume of Woodmans Ketchup while keeping all other variables constant. This effect was the smallest in magnitude among the utilitarian brands (unweighted average: -15.08% and median: 11.17%), and Menzis Terrine Culinaire Liquid Soups had the largest magnitude of -43.95%. The effect of feature only advertising on sales had all positive significant estimate results. Bic Razor Blades had the lowest effect in magnitude of 0.87% and Menzis Terrine Culinaire Liquid Soups the highest with 34.85% (unweighted average: 9.76% and median: 4.17%). The effect of display only advertising on sales volume was significant for seven out of eleven brands, and all had positive effects, ranging from 1.53% for Antonio Deniro Spaghetti to 13.31% for Menzis Terrine Culinaire Liquid Soups (unweighted average: 6.03% and median: 4.81%). The combination of display and feature advertising had only positive effects, and ten out of eleven estimates were significant (unweighted average: 14.04% and median: 6.69%). The lowest magnitude was for Bic Razor Blades 1.34% and the highest for Menzis Terrine Culinaire Liquid Soups 44.30%. The effect of advertisement by the manufacturer on sales volume only had a positive immediate effect for Kanis Gunnink Tea 1.66%, and all others were not significant. Just like with the hedonic product categories, none of the variables still affected the sales after fifty-two weeks of the initial marketing activity.
Variables Minimum Maximum Median Unweighted average
Sig Positive Negative
Price index -0.0245 -0.4395 -0.1117 -0.1508 9/11 0 9 Feature only 0.0087 0.3485 0.0417 0.0976 11/11 11 0 Display only 0.0153 0.1331 0.0481 0.0603 7/11 7 0 Display and feature 0.0134 0.4430 0.0669 0.1404 10/11 10 0 Advertisement by the manufacturer 0.01661 0.01661 0.01661 0.01661 1/11 1 0
Table 7 Coefficients of IRF from utilitarian brands
18 4.4 Persistent effects
The VAR models which were created with the sales variable in its first difference have not been taken into account in the representation of the results yet. This because all of the effects found can be interpreted as persistent effects compared to the previously reported immediate effects. These models will be discussed next. In total six utilitarian brands were modelled into VAR models; Belamy and Kazan from the Cat food category, Heinz from the Ketchup category, Becas from the Liquid soups category and Coberco and Danone from the Yoghurt category. To improve the reporting validity of effects from marketing on sales, a couple of extra analyses have been made. The selection of brands that was added had to fulfil the following requirements: the unit root in the sales variable should disappear when taking the first difference and only the advertisement by the manufacturer variable could have insufficient data to create a VAR model. The latter of the requirements was taken as this variable missed sufficient observations in most cases, which lead to excluding brands from the initial research, besides it had the most nonsignificant results for the brands that were used in the initial analyses. This lead to six extra VAR models that were created, of which table 8 gives a clear overview, the estimates for the independent variables have been added to the tables in appendices 4 and 5.
Product category Brand name Brand Size Adjusted R² Hedonic
Cornflakes No Brand Small 0.219
Custard desserts Domo Dessert Vla Small 0.053
Utilitarian
Liquid soups Campbells Tastes of the World Small 0.275
Razor Blades Gillette gii Plus Small 0.233
Spaghetti Epifine Small 0.197
Tea Duikertjes Small 0.027
Table 8 Overview of extra VAR models
19
average effect was 7.25% and a median of -4.07%. The effect of feature only advertising reported nine significant results (unweighted average: 6.67% and median: 5.68%), all of which are positive. The smallest effect was for Coberco (2.47%) and the largest for Heinz (19.09%). Display only advertising also had only positive effects for the significant estimates (unweighted average: 13.32% and median: 4.01%). Four were not significant, and therefore cannot be interpreted. The smallest effect occurred for Danone (1.41%) and Domo Dessert Vla had the largest effect (54.86%). The combination of display and feature advertising had significant positive estimates for eleven brands (unweighted average: 8.02% and median: 6.11%). Coberco had to smallest effect (1.81%) and Domo Dessert Vla had the largest effect (26.73%). As the extra models did not include the advertisement by the manufacturer variable, the effect could only be analysed for the six initial models. Three of those six initially created models reported significant results for this variable (unweighted average: 2.00% and median: 2.49%). The smallest effect for Coberco is a negative effect of -1.43% and the largest effect for Heinz is a positive effect of 4.92%.
Variables Minimum Maximum Median Unweighted average
Sig Positive Negative
Price index -0.2434 1.1072 -0.0407 0.0725 11/12 2 9
Feature only 0.0247 0.1909 0.0568 0.0667 9/12 9 0
Display only 0.0141 0.5486 0.0401 0.1332 8/12 8 0
Display and feature 0.0181 0.2673 0.0611 0.0802 11/12 11 0 Advertisement by the
manufacturer
-0.0143 0.0492 0.0249 0.0200 3/6 2 1
Table 8 Coefficients of IRF for persistent effects
5. Discussion
20
standard deviation shocks are the same. In this manner, the effects on the sales volume can be compared to a certain extent.
5.1 Immediate effects
The first set of hypotheses states that there is a positive immediate effect of the marketing stimuli on sales. To answer this, the results from the impulse response functions need to be consulted. The assumed positive immediate effect of advertisement by the manufacturer on sales can be confirmed, although only five significant estimates were found, they all reported positive effects in week one. As can be seen from the output, one standard deviation shock in advertisement by the manufacturer leads to an immediate positive effect in the sales volume ranging from 1.66% to 5.43% while keeping all other variables constant. The unweighted average is 3.27%. Hypothesis H1a is accepted. All other effects have the same interpretation, but will be presented more briefly. Feature only advertising reports estimates with the same positive valence, and the estimates for all brands are significant. All effects are positive in week one, ranging from 0.87% to 34.85% which is why hypothesis H1b is accepted. The unweighted average is 9.99%. Seventy-six percent of the estimates for display only advertising were significant, ranging from 1.53% to 24.51%. As these are all positive H1c is accepted too. The unweighted average is 7.94%. The last sub-hypothesis assumed an immediate positive effect for the combination of display and feature advertising. Ninety-two percent of the brands resulted in a significant positive effect. Therefore, hypothesis H1d can be accepted. The range of the effects goes from 1.34% to 44.30%. The unweighted average is 12.76%, which also is the largest among the types of advertising that have just been discussed. These results are all in line with what was expected and show that it is beneficial to promote FMCG’s by the use of advertisement by the manufacturer, feature only, display only, display and feature. In null cases was a significant effect left from the advertisement effort in week fifty-two.
21 5.2 Comparing hedonic and utilitarian products
The next set of hypotheses is about the difference in strength of the effects for all the advertising variables between hedonic and utilitarian brands. An added z test was performed to make sure that the variables are overall significant in each of these two groups. This allows for comparing the effects from one group with the other.
As was assumed from the literature (Price et. Al 1989; Fennis & Stroebe, 2010) the influence of marketing on sales for hedonic products is stronger compared to the influence on utilitarian products. These comparisons are also done ignoring the difference in terms of the standard deviation size of the shocks that were added with the IRF. When comparing the difference in the effects, it seems that there can be quite large differences between brands. To be able to answer the hypotheses, forming the separate brands into one hedonic and one utilitarian group can help. Therefore, the unweighted average value of all significant brand estimates has been taken to compare hedonic versus utilitarian brands.
From the estimates it can be seen that advertisement by the manufacturer has a stronger positive effect on sales for hedonic brands (3.75%) compared to utilitarian brands (1.66%). The effect of feature only advertising gives a different comparison, as on average, a stronger effect is reported for utilitarian products of (9.76%), compared to the effect with hedonic products of (9.62%). The difference between the two is 0.14%. This difference is so small that not only based on the comparison, but also based on the size of the difference between the
Hypotheses
H1a Advertisement by the manufacturer has a positive immediate effect on sales of fast
moving consumer goods.
H1b Feature only advertising has a positive immediate effect on sales of fast moving
consumer goods.
H1c Display only advertising has a positive immediate effect on sales of fast moving
consumer goods.
H1d The combination of display and feature advertising has a positive immediate effect
on sales of fast moving consumer goods.
H2 Decreasing price has a positive immediate effect on sales of fast moving consumer goods.
22
two, hypothesis H3b is rejected. The results on display only advertising show a stronger effect for hedonic products (7.22%) than for utilitarian products (6.03%) making a difference of 1.19%. On average, the combination of display and feature advertising shows a stronger effect for utilitarian brands than for hedonic brands. The effect that is reported is 12.70% for hedonic products compared to 14.04% for utilitarian products. The difference between the two is 1.34%, and therefore hypothesis H3d is rejected. The strength of price on average is reported to be stronger for utilitarian brands 15.08%) compared to hedonic brands (-12.58%). The difference between the two product types is 2.50%. Hypothesis 4 cannot be accepted because of this finding. Hypothesis H3a and H3c were not rejected as for the comparisons that have been made. However, it is also difficult to accept them as the effects can not exactly be compared. Nevertheless, they will be accepted as they did have larger effects based on the interpretation used in this section.
Hypotheses
H3a Immediate effects caused by advertisement by the manufacturer are stronger for
hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
H3b Immediate effects caused by feature only advertising are stronger for hedonic fast
moving consumer goods compared to utilitarian fast moving consumer goods. X
H3c Immediate effects caused by display only advertising are stronger for hedonic fast
moving consumer goods compared to utilitarian fast moving consumer goods.
H3d Immediate effects caused by the combination of display and feature advertising are
stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
X
H4 Immediate effects caused by increasing the price are stronger for hedonic fast moving consumer goods compared to utilitarian fast moving consumer goods.
X
5.3 Persistent effects
23
24 Hypotheses
H5a There is a persistent positive effect on sales caused by advertisement by the
manufacturer for fast moving consumer goods of small brands.
H5b There is a persistent positive effect on sales caused by feature only advertisement
for fast moving consumer goods of small brands.
H5c There is a persistent positive effect on sales caused by display only advertisement
for fast moving consumer goods of small brands.
H5d There is a persistent positive effect on sales caused by the combination of display
and feature advertisement for fast moving consumer goods of small brands.
As mentioned in section two, only effects from advertising were expected; however, also the effect of price was present. With eight significant estimates for small brands, the unweighted average results in an effect of 129.37%. This result is positive and therefore the opposite of what would be expected from the relation of price to sales. Gillette Gii Plus and Domo Dessert Vla have a positive effect and the other brands have relatively small negative effects compared to the positive effects of these two brands. This causes the unweighted average to be positive as well. The interpretation of these two brands would mean that by increasing the price, consumers would buy more of that product. The only argument that would be valid for this to be true is the one from Griskevicius et al., which was mentioned in the literature review (2010). This argument would than state that these two brands are seen as a sort of status products and a higher price would therefore evoke a perception of more quality, and status. On the other hand, just as with Coberco it might be a bias caused by other omitted variables.
6. Conclusion
25
This study has showed that there is a positive relationship between the advertising variables advertisement by the manufacturer, feature only, display only, and the combination of display and feature advertising on the sales volume. This finding is in line with previous studies (Zhang et al., 2009; Allenby and Ginter, 1995). The unweighted average of the combination of display and feature was the highest among the advertising variables overall, which shows that based on this comparison it has the largest effect on sales. The effects that were found are immediate effects, and therefore only answer one part of the main research question. However, the research also focuses on the difference between hedonic and utilitarian fast moving consumer goods. Although it is difficult to exactly describe differences between these two groups, some conclusions can be drawn with a side note that the exact effects cannot be interpreted. Advertisement by the manufacturer, and display only advertising both have a higher unweighted average for the effect on sales volume for hedonic products compared to utilitarian products. This is the other way around for the combination of display and feature advertising. Price has shown to have a negative relationship with sales volume, which is in line with earlier findings (Pauwels et al., 2002; Sethuraman et al., 2002). In terms of the difference of the effect between hedonic and utilitarian products, it can be stated that based on the unweighted average, the negative effect of price on utilitarian products is larger compared to hedonic products, but no specific statement can be given. Furthermore, some conclusions can be drawn upon the results for persistent sales effects. First, more unit roots were found in the sales variable for small brands (18) compared to large brands (5) which seems to be in line with the study of Slotegraaf and Pauwels (2008). From those analyses, the results could be interpreted as persistent effects. Advertisement by the manufacturer, feature only, display only, and the combination of display and feature advertising all have a positive persistent effect on the sales volume. Also price seemed to have a persistent negative effect on the sales volume when price would be increased. In addition, the unweighted averages of all advertising variables reported are higher for small brands compared to large brands. This infers that advertising could be more effective for small brands.
26
manufacturer and display only advertising. In terms of the effect of price changes, more impact is with utilitarian products.
7. Managerial implications and limitations
7.1 Managerial implications
Based on the discussion and conclusion some managerial implications can be given. These implications can be used by marketing managers to make decisions on the marketing mix and try to influence consumers’ buying decisions. Advertisement seems to have more influence on durable goods than on the nondurable fast moving consumer goods (Sethuraman et al., 2011), which makes it very important to know what works best so the marketing budget is spent in the best way. In terms of immediate effects on sales volume, expenditures on advertisement by the manufacturer, feature only, display only, and the combination of display and feature advertising all have shown to have a positive influence. However, it is important to know that there are quite some large differences between the effectiveness of advertising for different brands and categories. Overall, the combination of display and feature advertising has the largest positive effect. Based on the comparisons made between hedonic and utilitarian brands, some other suggestions can be made.
27
investigate how large the effect would be for their own brands. Next to these immediate effects, marketing managers should know that the effects of marketing on sales could also achieve persistent effects on their sales. Focusing on small brands with marketing efforts is recommended, as the effects seem to be stronger for those compared to large brands. An example for a small brand would be that when increasing the combination of weighted display and feature advertising of Domo Dessert Vla Custard desserts by 0.10 (one standard deviation) leads to an increase in sales volume of 26.73%. An example for a large brand would be that when increasing the combination of weighted display and feature advertising of Becas Liquid soups by 1.16 (one standard deviation) leads to an increase in sales volume of 5.73%.
7.2 Limitations and avenues for future research
28
References
Ailawadi, K.L. Neslin, S.A. 1998. The effect of promotion on consumption: buying more and consuming it faster. Journal of Marketing Research, 35:390-398.
Allenby, G. M. Ginter, J. L. 1995. The effects of in-store displays and feature advertising on consideration sets. International Journal of Research in Marketing, 12:67-80.
Arnold, M.J. Reynolds, K.E. 2003. Hedonic shopping motivations. Journal of Retailing, 79:77-95.
Assmus, G. Farley, J.U. Lehmann, D.R. 1984. How advertising affects sales: meta-analysis of econometric results. Journal of Marketing Research, 21:65–74.
Ataman, M.B. Van Heerde, H.J. Mela, C.F. 2010. The long-term effect of marketing strategy on brand sales. Journal of Marketing Research, 47:866-882.
Babin, B.J. Darden, W.R. Griffin, M. 1994. Work and/or fun: measuring hedonic and utilitarian shopping value. Journal of Consumer Research, 20:644–656.
Bodapati, V.A. Srinivasan, V. 2006. The Impact of Feature Advertising on Customer Store Choice. Working Papers Stanford Graduate School of Business, 1-44.
Briesch, R.A. Chintagunta, P.K. Fox, E.J. 2009. How does assortment affect grocery store choice? Journal of Marketing Research, 46:176-189.
Cameron, G.T. Haley, J.E. 1992. Feature advertising: policies and attitudes in print media.
Journal of Advertising, 21:47-55.
29
Collins, A. Kavanagh, E. Cronin, J. George, R. 2014. Money, mavens, time, and price search: modelling the joint creation of utilitarian and hedonic value in grocery shopping. Journal of
Marketing Management, 30:719.
CPB 2002. Factsheet miljoenennota 2002. (accessed February 12, 2016) [ available at
http://www.rijksbegroting.nl/archief-minfin/2002/default36cf.pdf%3FCMS_TCP=tcpAsset&id=F53854B84703496EAAF41B1A5 0CFE736?CMS_TCP=tcpAsset&id=F53854B84703496EAAF41B1A50CFE736]
Dekimpe, M.G. Hanssens, D.M. 1995. The persistence of marketing effects on sales.
Marketing Science, 14: 1-21.
Dekimpe, M.G. Hanssens, D.M. Silva-Risso, J. M. 1999. Long-run effects of price promotions in scanner markets. Journal of Econometrics, 89:269–291.
Dekimpe, M.G. Hanssens, D.M. Nijs, V.R. Steenkamp, J.E.M. 2005. Measuring short- and long-run promotional effectiveness on scanner data using persistence modelling. Stochastic
Models Bus. Ind., 21:409-416.
Griskevicius, V. Van den Bergh, B. Tybur, J.M. 2010. Going green to be seen: status, reputation, and conspicuous conservation. Journal of Personality and Social Psychology, 98:392-404.
Janssens, W. De Pelsmacker, P. 2005. Advertising for new and existing brands: the impact of media context and type of advertisement. Journal of Marketing Communications, 11:113-128.
Johnson, E.J. Bellman, S. Lohse, G.L. 2003. Cognitive lock-in and the power law of practice.
Journal of Marketing, 67: 62–75.
Kahneman, D. Tversky, A. 1979. Prospect theory: an analysis of decision under risk.
30
Kan, C. Lichtenstein, D.R. Grant, S.J. Janiszewski, C. 2014. Strengthening the influence of advertised reference prices through information priming. Journa; of Consumer Research, 40: 1078-1096.
Kotler, P. 1997. Marketing Management (9th edition). Englewood Cliffs, NJ: Prentice Hall.
Mishra, H. Mishra, A. Nayakankuppam, D. 2007. Seeing through the heart’s eye: the interference of system 1 in system 2. Marketing Science, 26:666-678.
Osborne, D.K. 1964. On the goals of the firm. Journal of Economics, 78:592-603.
Osinga, E.C. Leeflang, P.S.H. Wieringa, J.E. 2010. Early marketing matters: a time-varying parameter approach to persistence modeling. Journal of Marketing Research, 47:173-185.
Pauwels, K. Hanssens, D.M. Siddarth, S. 2002. The long-term effects of price promotions on category incidence, brand choice, and purchase quantity. Journal of Marketing Research, 39:421-439.
Pieters, R. Wedel, M. Zhang, J. 2007. Optimal feature advertising design under competitive clutter. Management Science, 53: 1815-1828.
Price, L.L. Feick, L.F. Higie, R.A. 1989. Preference heterogeneity and coorientation as determinants of perceived informational influence. Journal of Business Research, 19:227-242.
Sethuraman, R. Tellis, G.J. Briesch, R.A. 2011. How well does advertising work? Generalizations from meta-analysis of brand advertising elasticities. Journal of Marketing
Research, 48: 457-471.
Slotegraaf, R.J. Pauwels, K. 2008. The impact of brand equity and innovation on the long-term effectiveness of promotions. Journal of Marketing Research, 45:293–306.
Vakratsas, D. Ambler, T. 1999. How advertising works: what do we really know? Journal of
31
Van Heerde, H. J. Leeflang, P. S. H. Wittink, D. R. 2000. The estimation of pre- and postpromotion dips with store-level scanner data. Journal of Marketing Research, 37:383-395.
Van Heerde, H. J. Gijsenberg, M. J. Dekimpe, M. G. Steenkamp, J. B. E. M. 2013. Price and advertising effectiveness over the business cycle. Journal of Marketing Research, 50:177-193.
Verhoef, P.C. Leeflang, P.S.H. 2009. Understanding the marketing department's influence within the firm. Journal of Marketing, 73:14-37.
Zhang, J. 2006. An integrated choice model incorporating alternative mechanisms for consumers’ reactions to in-store display and feature advertising. Marketing Science, 25:278-290.
Zhang, J. Wedel, M. Pieters, R. 2009. Sales effects of attention to feature advertisements: a bayesian mediation analysis. Journal of Marketing Research, 46:669-681.
32
Appendices
Appendix 1 Product categories Hedonic product categories Brands Utilitarian product categories Brands
Apple juice Bari, Euroshopper, Jufri, AK Cat food Euroshopper, Belamy, Whiskas, Kazan
Cornflakes Pablopepe, Euroshopper, Nobrand, Oldfarmercornflakes
Coffee milk Campina, Friesmeisje, Menkomel, Saliter
Crème rinse Nivea, Euroshopper, Andrelon Cream Care, Nobrand
Instant coffee A Vogel Bambu, Buisman, Grandos, Douwe Egberts Cappuccinoshake
Crisps Tuttotondo, Bahlsenpicanterie, Croky Bubblechips, Croky Vakantiepak
Ketchup Woodmans, Cereal, Heinz, Heinz Tomato Ketchup
Custard dessert Naarmann, Domo, Domo Dessertvla, Domo Lang Lekker
Liquid soups in tins & cartons
Becas, Lacroix, Menziterrine Culinaire, Campbells Tastes Of The World
Instant soup Gasa Ayam Special, Mamee, Knorr Quicksoup, Weilihmien
Razor blades Bic, Gillette, Gillette Gii Plus, Gillette Sensor
Mustard Amora, Lowensenf, Colmans, Gulpener
Spaghetti Antonio Deniro, Barilla, Epifine, Grand Italia
Popcorn Bards, Nobrand, Brandwijk, Borden Crackerjack
Tea Kanis Gunnink, Pickwick, Van Nelle Twee Thee, Duikertjes
Rice Inter Rice, Flits Snelrijst, Nobrand, Oryza
Toilet tissue Liny, Pura Ninfea, Eldou, Lotus Comfort Plus
Shampoo Selsun, Albertovo5, Crinopex, Dicom Cosmetics
Yoghurt Melkunie, Coberco, Danone, Fage
Appendix 2 Unit root tests hedonic FMCG’s
Product category Included in test equation Unit-Root Apple juice Intercept Trend and intercept None
LNBariSales .000 .000 .657 -
33
LNBariFeatuerOnly .000 .000 .015 -
LNBariAdvertising Near singular matrix
LNBariPriceindex .000 .000 .711 - LNBaridisplayandfeature .000 .000 .104 - LNEuroshopperSales .000 .000 .545 - LNEuroshopperdisplayOnly .000 .000 .054 - LNEuroshopperFeatuerOnly .000 .000 .001 - LNEuroshopperAdvertising .000 .000 .000 - LNEuroshopperPriceindex .000 .000 .742 - LNEuroshopperdisplayandfeature .000 .000 .006 - LNJufriSales .003 .007 .511 - LNJufridisplayOnly .000 .000 .000 - LNJufriFeatuerOnly .000 .000 .000 -
LNJufriAdvertising Near singular matrix
LNJufriPriceindex First difference .669 .000 987 .000 .656 .000 -
LNJufridisplayandfeature Near singular matrix
LNAKSales First difference .741 .000 .810 .000 .112 .000 - LNAKdisplayOnly .000 .000 .000 -
LNAKFeatuerOnly Near singular matrix
LNAKAdvertising Near singular matrix
LNAKPriceindex First difference .611 .008 .424 .040 .125 .001 -
LNAKdiplayandfeature Near singular matrix
Product category Included in test equation Unit-Root Cornflakes Intercept Trend and intercept None
34
LNEuroshopperSales .248 .000 .492 -
LNEuroshopperdisplayOnly .000 .000 .000 -
LNEuroshopperFeatuerOnly .000 .000 .000 -
LNEuroshopperAdvertising Near singular matrix
LNEuroshopperPriceindex First difference .471 .000 .341 .000 .576 .000 - LNEuroshopperdisplayandfeature .000 .000 .000 - LNNobrandSales First difference .461 .000 .792 .000 .694 .000 - LNNobranddisplayOnly .000 .000 .000 - LNNobrandFeatuerOnly .000 .000 .000 -
LNNobrandAdvertising Near singular matrix
LNNobrandPriceindex First difference .465 .000 .785 .000 .680 .000 - LNNobranddisplayandfeature .000 .000 .000 - LNOldfarmercornflakesSales .007 .041 .817 - LNOldfarmercornflakesdisplayOnly .000 .000 .000 - LNOldfarmercornflakesFeatuerOnly .000 .000 .098 -
LNOldfarmercornflakesAdvertising Near singular matrix LNOldfarmercornflakesPriceindex First difference .1238 .000 .5141 .000 .3063 .000 - LNOldfarmercornflakesdisplayandfeature .000 .000 .000 -
Product category Included in test equation Unit-Root Creme Rinse Intercept Trend and intercept None
35 LNEuroshopperAdvertising .000 .000 .000 - LNEuroshopperPriceindex .005 .000 .749 - LNEuroshopperdisplayandfeature .000 .000 .000 - LNAndreloncreamcareSales First difference .158 .000 .283 .000 .432 .000 - LNAndreloncreamcaredisplayOnly .000 .000 .000 - LNAndreloncreamcareFeatuerOnly .000 .000 .000 -
LNAndreloncreamcareAdvertising Near singular matrix LNAndreloncreamcarePriceindex First difference .358 .000 .325 .000 .300 .000 - LNAndreloncreamcaredisplayandfeature Near singular matrix
LNNobrandSales First difference .860 .000 .167 .000 .334 .000 - LNNobranddisplayOnly .000 .000 .000 - LNNobrandFeatuerOnly .000 .000 .000 -
LNNobrandAdvertising Near singular matrix
LNNobrandPriceindex First difference .655 .000 .347 .000 .832 .000 - LNNobranddisplayandfeature Near singular matrix
Product category Included in test equation Unit-Root Crisps Intercept Trend and intercept None
LNTuttotondoSales .000 .000 .774 - LNTuttotondodisplayOnly .000 .000 .280 - LNTuttotondoFeatuerOnly .000 .000 .039 - LNTuttotondoAdvertising .000 .000 .001 - LNTuttotondoPriceindex .000 .000 .000 - LNTuttotondodisplayandfeature .000 .000 .043 -
LNBahlsenpicanterieSales Data have zero variance
LNBahlsenpicanteriedisplayOnly .000 .000 .190 -
LNBahlsenpicanterieFeatuerOnly .000 .000 .097 -
LNBahlsenpicanterieAdvertising .000 .000 .000 -
LNBahlsenpicanteriePriceindex .000 .000 .000 -
36 Croky Bubblechips and Croky Vakantiepak missed all sales figures.
Product category Included in test equation Unit-Root Custard dessert Intercept Trend and intercept None
LNNaarmannSales .000 .000 .735 - LNNaarmannDisplayOnly .000 .000 .000 - LNNaarmannFeatuerOnly .000 .000 .000 - LNNaarmannAdvertising .000 .000 .000 - LNNaarmannPriceindex .045 .122 .227 LNNaarmannDisplayandfeature .000 .000 .000 - LNDomoSales First difference .417 .947 .000 .022 LNDomoDisplayOnly .000 .000 .000 - LNDomoFeatuerOnly First difference .104 .242 .000 .010
LNDomoAdvertising Near singular matrix -
LNDomoPriceindex First difference .558 .753 .000 .076 LNDomoDisplayandfeature .000 .000 .000 - LNDomodessertvlaSales First difference .087 .000 .097 .000 .188 .000 - LNDomodessertvlaDisplayOnly .000 .000 .000 - LNDomodessertvlaFeatuerOnly .000 .000 .000 -
LNDomodessertvlaAdvertising Near singular matrix LNDomodessertvlaPriceindex First difference .095 .000 .114 .000 .397 .000 - LNDomodessertvlaDisplayandfeature .000 .000 .000 - LNDomolanglekkerSales First difference .410 .000 .577 .000 .541 .000 - LNDomolanglekkerDisplayOnly .000 .000 .000 - LNDomolanglekkerFeatuerOnly .001 .000 .000 -
LNDomolanglekkerAdvertising Near singular matrix
LNDomolanglekkerPriceindex .037 .053 .324
37
Product category Included in test equation Unit-Root Instant soup Intercept Trend and intercept None
LNGasaayamspecialSales .000 .000 .784 - LNGasaayamspecialDisplayOnly .000 .000 .031 - LNGasaayamspecialFeatuerOnly .000 .000 .000 - LNGasaayamspecialAdvertising .000 .000 .000 - LNGasaayamspecialPriceindex .000 .000 .602 - LNGasaayamspecialDisplayandfeature .000 .000 .000 - LNMameeSales .000 .000 .776 - LNMameeDisplayOnly .000 .000 .000 - LNMameeFeatuerOnly .000 .000 .000 - LNMameeAdvertising .000 .000 .000 - LNMameePriceindex .000 .000 .557 - LNMameeDisplayandfeature .000 .000 .000 - LNKnorrquicksoupSales .000 .000 .664 - LNKnorrquicksoupDisplayOnly .000 .000 .000 - LNKnorrquicksoupFeatuerOnly .000 .000 .000 -
LNKnorrquicksoupAdvertising Near singular matrix LNKnorrquicksoupPriceindex First difference .618 .000 .909 .000 .623 .000 - LNKnorrquicksoupDisplayandfeature .000 .000 .000 - LNWeilihmienSales .000 .000 .566 - LNWeilihmienDisplayOnly .000 .000 .000 -
LNWeilihmienFeatuerOnly Near singular matrix
LNWeilihmienAdvertising Near singular matrix
LNWeilihmienPriceindex .000 .000 .622 -
LNWeilihmienDisplayandfeature Near singular matrix
Product category Included in test equation Unit-Root Mustard Intercept Trend and intercept None
LNLowensenfaSales .000 .000 .000 -
LNLowensenfaDisplayOnly .000 .000 .000 -
LNLowensenfaFeatuerOnly .000 .000 .000 -
LNLowensenfaAdvertising .000 .000 .000 -
38
LNLowensenfaDisplayandfeature .000 .000 .000 -
LNGulpenerSales .000 .000 .014 -
LNGulpenerDisplayOnly .000 .000 .000 -
LNGulpenerFeatuerOnly Near singular matrix
LNGulpenerAdvertising .000 .000 .014 -
LNGulpenerPriceindex .780 .457 .902
LNGulpenerDisplayandfeature Near singular matrix
Amora and Colmans had no sales figures
Product category Included in test equation Unit-Root Popcorn Intercept Trend and intercept None
LNBardsSales .001 .000 .776 -
LNBardsDisplayOnly .000 .0002 .006 -
LNBardsFeatuerOnly .000 .000 .000 -
LNBardsAdvertising Near singular matrix
LNBardsPriceindex .000 .000 .672 -
LNBardsDisplayandfeature .000 .0001 .000 -
LNNobrandSales .000 .000 .804 -
LNNobranddisplayOnly .000 .000 .000 -
LNNobrandFeatuerOnly .000 .000 .000 -
LNNobrandAdvertising Near singular matrix
LNNobrandPriceindex First difference .536 .000 .589 .000 .391 .000 - LNNobranddisplayandfeature .000 .000 .000 - LNBrandwijkSales .000 .000 .000 - LNBrandwijkDisplayOnly .000 .000 .000 -
LNBrandwijkFeatuerOnly Near singular matrix
LNBrandwijkAdvertising Near singular matrix
LNBrandwijkPriceindex .000 .000 .000 -
LNBrandwijkDisplayandfeature Near singular matrix
LNBordencrackerjackSales .040 .249 .651
LNBordencrackerjackDisplayOnly .000 .000 .000 -
39 LNBordencrackerjackAdvertising Near singular matrix
LNBordencrackerjackPriceindex .000 .000 .000 -
LNBordencrackerjackDisplayandfeature Near singular matrix
Product category Included in test equation Unit-Root Rice Intercept Trend and intercept None
LNInterriceSales .000 .000 .664 -
LNInterriceDisplayOnly .000 .000 .001 -
LNInterriceFeatuerOnly .000 .000 .137 -
LNInterriceAdvertising Near singular matrix
LNInterricePriceindex .000 .000 .600 -
LNInterriceDisplayandfeature .000 .000 .008 -
LNFlitssnelrijstSales .148 .000 .557 -
LNFlitssnelrijstDisplayOnly .000 .000 .203 -
LNFlitssnelrijstFeatuerOnly .000 .000 .084 -
LNFlitssnelrijstAdvertising Near singular matrix
LNFlitssnelrijstPriceindex .210 .000 .844 -
LNFlitssnelrijstDisplayandfeature .000 .000 .001 -
LNNobrandSales .000 .000 .955 -
LNNobranddisplayOnly .000 .000 .000 -
LNNobrandFeatuerOnly .000 .000 .024 -
LNNobrandAdvertising Near singular matrix
LNNobrandPriceindex .752 .039 .246 - LNNobranddisplayandfeature .000 .000 .000 - LNOryzaSales .000 .000 .653 - LNOryzaDisplayOnly .000 .000 .090 - LNOryzaFeatuerOnly .000 .000 .000 - LNOryzaAdvertising .000 .001 .000 - LNOryzaPriceindex .000 .000 .445 - LNOryzaDisplayandfeature .000 .000 .000 -
Product category Included in test equation Unit-Root Shampoo Intercept Trend and intercept None
LNSelsunSales .000 .000 .708 -
40 LNSelsunFeatuerOnly .000 .000 .002 - LNSelsunAdvertising .000 .000 .000 - LNSelsunPriceindex .024 .064 .673 LNSelsunDisplayandfeature .000 .000 .000 - LNAlbertovo5Sales .000 .000 .766 - LN Albertovo5DisplayOnly .000 .000 .033 - LN Albertovo5FeatuerOnly .000 .000 .000 - LN Albertovo5Advertising .000 .000 .015 - LN Albertovo5Priceindex .000 .000 .694 - LN Albertovo5Displayandfeature .000 .000 .003 - LNCrinopexSales .001 .000 .537 - LNCrinopexdisplayOnly .000 .000 .018 - LNCrinopexFeatuerOnly .000 .000 .000 - LNCrinopexAdvertising .000 .000 .000 - LNCrinopexPriceindex .000 .000 .580 - LNCrinopexdisplayandfeature .000 .000 .000 - LNDicomcosmeticsSales .000 .000 .806 - LNDicomcosmeticsDisplayOnly .000 .000 .002 - LNDicomcosmeticsFeatuerOnly .000 .000 .000 - LNDicomcosmeticsAdvertising .000 .000 .000 - LNDicomcosmeticsPriceindex .000 .000 .737 - LNDicomcosmeticsDisplayandfeature .000 .000 .044 -
Appendix 3 Unit root tests utilitarian FMCG’s
Product category Included in test equation Unit-Root
Cat food Intercept Trend and intercept None
LNEuroshopperSales .211 .001 .605 -
LNEuroshopperdisplayOnly .000 .000 .001 -
LNEuroshopperFeatuerOnly .000 .000 .002 -
