Master Thesis
Effect of pricing and promotion policy on price and
promotion image of retailers: an empirical study
Marijn Bruggeman
S2365006
University of Groningen
Faculty of Economics and Business
MSc Marketing Intelligence
Radebinnensingel 38A
9711EM Groningen
+31(6)53554524
bruggemanmarijn@gmail.com
2
Acknowledgement
3
Abstract
This study investigates how retailers’ actual pricing and promotion policy affects their price and promotion image. Firstly, it is studied how pricing and promotion affects price image. The results indicate that retailers with higher prices have a higher price image and that both prices of national brands and private labels affect price image. In addition, a short-term effect of change in price level on price image is found. Secondly, it is studied how promotion intensity and promotion discount of both national brands and private labels affects a retailer’s promotion image. The results indicate that national brand promotions are perceived more attractive and that private label promotions negatively affects a retailer’s promotion image. Finally, the implications of the results for retailers are discussed and future research directions are suggested.
Keywords: Retail, Price Image, Promotion Image, Promotion Intensity, Promotion Discount,
4
TABLE OF CONTENTS
1. INTRODUCTION ... 5
2. THEORETICAL BACKGROUND ... 8
2.1 Price image ... 8
2.2 Promotions and Promotion Image ... 8
3. CONCEPTUAL DEVELOPMENT ... 10 3.1 Price image ... 10 3.2 Promotion image ... 12 4. METHODOLOGY ... 14 4.1 Data collection ... 14 4.2 Measurement ... 14 5. RESULTS ... 16 5.1 Descriptive statistics ... 16 5.2 Empirical results ... 20
5.2.1 Price image models ... 20
5.2.2 Promotion image models ... 24
6. CONCLUSION AND DISCUSSION ... 27
7. LIMITATIONS AND FURTHER RESEARCH ... 29
5
1. INTRODUCTION
Pricing is one of the main marketing instruments and is a crucial element in retail marketing. “Retail price is unique among marketing variables as it can be changed quickly, and it has immediate and dramatic effects on sales” (Mulhern, 1997). Price has also been the main element of competition in retailing traditionally (Dickson & Urbany, 1994) and has been an important factor for consumers when choosing a food retail store (Zielke, 2010). In recent years, price has become an even more important aspect of competition, because of the rapid expansion of discount retailers, especially in Europe (Zielke, 2010). Therefore, another aspect of retail marketing that is growing in importance is the price image of a store, because consumer decisions are not only influenced by the prices of individual products, but also by the consumers impression of the retailers’ overall price level (Hamilton & Chernev, 2013). “The notion that consumer decisions are influenced not only by a retailer’s actual prices but also by consumer perceptions of the retailer’s price image is gaining popularity among managers. This notion reflects managers’ beliefs that, faced with rapidly emerging new retail formats, an increasing number of retailer outlets in which to shop, and the ever-expanding number of product options from which to choose, consumers tend to rely on their overall impression of a store’s prices when making their purchase decisions” (Hamilton & Chernev, 2013). Price image is especially relevant for grocery retailers, because consumers buy convenience goods without much consideration and effort and often rely on an overall impression of a retailer that offers them good value for money in the long term (Feichtinger, Luhmer & Sorger, 1988). This is in line with findings of a Nielsen report from 2008: “85% of the worlds’ consumers ranked good value
for money the most important consideration when choosing a grocery store” and for 70% of
those consumers “it was important the store had a reputation for being cheaper than competitors - even if, in reality, this was not the case” (Nielsen 2008). Developments in the food retail sector confirm the growing importance of a store’s price image. In 2003, Dutch market leader Albert Heijn initiated a price war to improve their price image and stop their sliding market share, in which they succeeded (Van Heerde, Gijsbrechts & Pauwels, 2008) and also Walmart reduced prices of roughly 10.000 products to polish their price image in 2010 (Bustillo & Martin, 2010) which they repeated in 2017 (Haddon & Nassauer, 2017).
6 thus, more accurate. However, despite increased availability of price information, the price image of many retailers does not always accurately represent their actual prices” (Hamilton & Chernev, 2013) Price image formation is a very complex process and prior research provided several explanations why this will not necessarily result in a price image that reflects a retailer’s actual price level accurately: Consumers’ ability to process prices is limited and even if price knowledge is present, the evaluation of prices is a subjective process. Consumers can also derive a price image from nonprice-related cues like advertisements, store design, size and it’s service level (Zielke, 2006). Besides that, consumers’ impressions of price image have a tension to be resistant to change and consumers are unlikely to update their price image beliefs, unless they encounter a consistent accumulation of disconfirming price information (Hamilton & Chernev, 2013).
Another element of the marketing mix that is growing in importance are the promotions retailers have. The promotional pressure, which is the percentage of volume sales bought on promotion, has risen in Europe over the years. On average, 28.1% of all products were sold on promotion across Europe in 2016, especially in the UK where the promotional pressure was 51.5%. Also in the Netherlands, the promotional pressure grew with 1.1 percent point to 25.5% in 2016 (Eales, 2016). Prior research on the effect of promotions argued that promotions significantly increases sales and that they can increase store traffic, but they generally failed to provide evidence for positive long-term effects (Blattberg, Briesch & Fox, 1995). Although prior research did provide evidence that perceived promotions had a favorable impact on a store’s price image (Lourenco, Gijsbrechts & Paap, 2015), there is no research on how a retailer’s perceived promotions, or promotional image, is influenced by a retailer’s actual promotion policy.
7
Research question 1: To what extent are price image and promotion image related to actual differences in pricing and promotions of retailers?
Research question 2: How do retailers’ actual pricing and promotion changes affect their price and promotion image?
Research question 3: How can retailers alter their price and promotion image with their pricing and promotional policy and what are the time lags?
8
2. THEORETICAL BACKGROUND
In this chapter, the concept of price- and promotion image will be explained and relevant literature will be discussed.
2.1 Price image
Price image is defined as “the general belief about the overall level of prices that consumers associate with a particular retailer” (Hamilton & Chernev, 2013). Consumers’ initial beliefs are often shaped by nonprice cues, such as a store’s ambiance, service level, assortment and advertised promotions (Lourenco et al., 2015), based on consumers’ assumption that those nonprice cues correlate highly with a store’s actual price level. Those beliefs are gradually updated as consumers get exposed to actual prices when they visit a store (Büyükkurt, 1986). However, price image formation is a complex process and will not necessarily result in a price image that reflects a retailer’s actual price level accurately. Prior research provided several explanations: Consumers’ ability to process prices is limited and even if price knowledge is present, the evaluation of prices is a subjective process. Besides that, consumers can derive a price image from nonprice-related cues like advertisements, store design, size and it’s service level (Zielke, 2006). This implies that many factors influence a retailers price image, including price-related factors as well as nonprice-related factors. This study will especially focus on price-related factors.
2.2 Promotions and Promotion Image
9 retailer offers (promotion intensity) and with the depth of the promotions, i.e. the discount they offer in percentages.
10
3. CONCEPTUAL DEVELOPMENT
In this chapter, a conceptual model is developed that summarizes this paper (See figure 1). The model actually comprises two models, one for the dependent variable Price Image and one for the dependent variable Promotion Image. Firstly, the model for Price Image will be discussed and hypotheses will be formulated for the independent variables. Subsequently the model for Promotion Image with the corresponding hypotheses will be discussed.
Average price level
- National brands - Private labels Price dispersion Price Image Promotion Image Promotion intensity - National brands - Private labels Promotion discount - National brands - Private labels
Figure 1: Conceptual model
3.1 Price image
Variables that might influence a retailers’ price image and that are included in this study are the average price level of a retailer, the dispersion of prices a retailer has, the intensity of promotions a retailer offers and the average promotion discount a retailer offers. All of these variables have been suggested to have an influence on price image by prior research (Hamilton & Chernev, 2013; Blattberg et al., 1995). As indicated by the conceptual model, a division is applied between national brands and private labels for the average price level, the promotion intensity and promotion discount variables.
Average price level. Evidently, one major factor that influences a store’s price image is the
11 is, in reality, more or less expensive than the same basket of goods from another retailer)” (Hamilton & Chernev, 2013). A store’s average price level has been considered the main driver of price image formation traditionally (Feichtinger et al., 1988). It’s expected that retailers with higher prices have a higher price image. Because prices of national brands are easier to compare, it’s expected that national brands have a bigger influence on store price image formation.
H1A: A higher average price level leads to a higher price image
H1B: A higher average price level of national brands leads to a higher price image H1C: Average price level of private labels leads to a higher price image
H1D: Average price level of national brands has a bigger influence on price image than average price level of private labels.
Price dispersion. When two retailers have a similar overall price level, it doesn’t mean they
have comparable prices across all products. One retailer can, for example, have a fairly consistent price compared to the market average on most product categories whereas another retailer could offer some products cheaper and some products more expensive than the market average. Price dispersion is “the degree to which a retailer’s prices are competitive across different product categories” (Hamilton & Chernev, 2013). This might have an impact on price image, because empirically research showed that product categories differ in store price image signalling power. “SPIs are more strongly shaped by prices of expensive categories that are storable and purchased in larger volumes at a time, in support of the premise that consumers perceive prices in these categories as more important and relevant to monitor” (Lourenco et al., 2015). Another indication that price dispersion might influence price image, is that the frequency in which consumers encounter low prices when evaluating a retailer’s assortment has a big influence on price image formation (Alba, Broniarczyk, Shimp & Urbany, 1994). Therefore, it’s likely that consumers encounter more low prices in a store with a higher price dispersion and form a lower price image for that store.
H2: A higher price dispersion leads to a lower price image.
Promotion intensity and promotion discount. There are two main ways in which retailers differ
12 on a store’s initial price image (Lourenco et al., 2015). This may imply that promotions do affect a store’s price image. Another research found that “high frequency of printed retail advertising and large number of featured items per advertising are perceived to be associated with lower prices” (Büyükkurt, 1986). Another prior experiment tested the effect of shallow, but frequent promotions and the effect of deep, but infrequent promotions (with an identical average price over time) on the estimated price of a single product. The results showed that the brand with the deep, but infrequent promotions was perceived cheaper. However, an adjusted experiment showed that when prices are distributed more complexly, the product with the shallow, but frequent promotions was perceived cheaper. The experiments showed that both promotion strategies led to a lower estimated price than for a product with a constant price (Alba et al., 1999). Another study also proved that frequent price discounts could lower consumers’ expected prices (Grewal, Krishnan, Baker & Borin, 1998) and in this case, expected prices of the retailer. In the literature, no division is made between private label and national brand promotions. It’s expected that higher promotion intensity and higher promotion discounts for both private labels and national brands will result in a lower price image.
H3A: Promotion intensity for national brands has a negative effect on price image. H3B: Promotion intensity for private labels has a negative effect on price image.
H4A: Average promotion discount for national brands has a negative effect on price image.
H4B: Average promotion discount for private labels has a negative effect on price image.
3.2 Promotion image
In this study, the promotion image is defined as the perceived attractiveness of the promotions. Variables that might influence the perceived attractiveness of the promotions and that are covered by this study are the promotion intensity and the average promotion discount offered by retailers. For both variables, a division is applied between private labels and national brands and private labels.
Promotion intensity. Previous research provided evidence that price promotions affect
13 quality and internal reference price (Grewal et al., 1998). The sum of the positive and negative effects of price discount is still positive, implying that a higher promotion intensity will result in a higher perceived quality and thus perceived promotion attractiveness. The study also reports that the negative effects may not hold for products with a higher quality, because the positive effects of a brand name and a brand’s perceived quality counter-balance those negative effects. This could be an indication that promotions of national brands have a higher perceived value, and thus a higher perceived promotion attractiveness, than promotions of private labels.
H5A: Promotion intensity for national brands has a positive effect on promotion image. H5B: Promotion intensity of private labels has a positive effect on promotion image. H5C: Promotion intensity of national brands has a stronger positive effect on promotion image than promotion intensity of private labels.
Promotion discount. In addition to the findings showed above, Grewal et al. (1998) also found
that the higher the promotion discount, the stronger the above mentioned effects of price discount on perceived value, perceived brand quality and internal reference price were. Since the sum of the positive and negative effects of price discount on perceived value is still positive, it’s expected that a higher promotion discount will result in a higher promotion attractiveness. Another study affirms this and provided evidence for a significant positive effect of promotion discount on offer value (Alford & Biswas, 2002). Based on the argument that higher quality products counter the negative effects of promotion discount on perceived value (Grewal et al., 1998), it’s expected that promotion discount for national brands has a stronger positive effect on promotion image.
14
4. METHODOLOGY
4.1 Data collection
This research is conducted in the Dutch grocery retailing market by using two different datasets for a period of 26 months from April 2016 till May 2018. For the independent price and promotion variables, data from superscanner.nl was used. Superscanner.nl collects daily prices of products of all the grocery retailers with online prices as well as the weekly promotion intensity and average promotion discount per week of grocery retailers. For the dependent variables price image and promotion image this study made use of the ‘EFMI shopper monitor’, which is an extensive dataset based on monthly surveys filled in by persons who usually do the groceries in their household. The respondents are a representative sample of Dutch grocery shoppers. In the 26 months covered by this study, 4875 respondents ever did their grocery shopping at either Albert Heijn, Jumbo or Plus, averaging approximately 188 respondents per month. On average, approximately 150 respondents per month ever did their grocery shopping at Albert Heijn, approximately 111 respondents per month ever did their grocery shopping at Jumbo and approximately 46 respondents ever did their grocery shopping at Plus. Usually, the data collection takes place in week 2 of the month and takes about a week.
Because of the limited number of observations in the ‘EFMI shopper monitor’ for the smaller grocery retail chains and the limited availability of online prices for all supermarkets, only the three largest grocery retailers with the highest market shares have been included for this study: Albert Heijn (35,3%), Jumbo (18,7%) and Plus (6,4%).
4.2 Measurement
15
Table 1
Construct Measure
Average price level*
- National brands*
- Private labels*
Price of a basket containing 97 products
- 80 national brands
- 17 private labels
Change in price level*
- National brands*
- Private labels*
Percentual change in price level compared to previous month.
Price dispersion* Average absolute deviation of a retailers’ product price and the average
price for that product corrected with the general price level of the retailer (in %)
Promotion intensity national brands*
Number of national brand Stock Keeping Units (SKUs) in promotion per week divided by number of national brand SKUs in assortment (in %)
Promotion intensity private labels*
Number of private label SKUs in promotion per week divided by number of private label SKUs in assortment (in %)
Average promotion discount national brands*
Average discount of national brand SKUs in promotion per month (in %)
Average promotion discount private labels*
Average discount of private label SKUs in promotion per month (in %)
Price image “Suppose that a full shopping basket at the average Dutch supermarket
costs 100 euros. However, some supermarkets are cheaper, others are more expensive. How much do you think that shopping basket costs at the following supermarkets (estimate)?” (in euros)
Promotion image “How would you rate the following supermarkets on the attractiveness
of their promotions?” (1 = very poor, 10 = excellent) *Note: These variables also have a (t-1) and a (t-2) variant.
Missing data
16
5. RESULTS
5.1 Descriptive statistics
In this paragraph descriptive statistics, along with some explorative charts are discussed. Firstly, descriptive statistics are provided to get an overview of the data (Table 2). The mean values of all the variables, along with the standard deviation and the minimum and maximum value for each variable are provided. All the variables show substantial variation between the three supermarket formulas and also within supermarket formulas, most variables show substantial variation.
Table 2
Albert Heijn Plus Jumbo
Construct Mean(SD) Min Max Mean(SD) min max Mean(SD) min max
17
Price Image
To get a feel for the data, charts are developed to see the development of variable values over time. Figure 2 shows the development of the average price levels of the three supermarkets per chain. It’s notable that the price lines of Plus and Albert Heijn are really similar, especially from 2017 and that Jumbo’s price level is substantially lower. However, Jumbo’s price line shows some similar increases and decreases with the price lines of Plus and Albert Heijn. Figure 4 and 5 show the developments of the national brand price levels and private label price level respectively. The national brand lines look very similar to the overall price level lines, which is logical, because 80 out of 97 products in the dataset are national brands. The private label price levels however, show an interesting downwards trend, with especially Albert Heijn and Plus decreasing the distance to Jumbo’s price level. The development of the supermarkets’ price images over time (Figure 3), also show interesting developments. Firstly, it’s notable that there is quite some difference between the price image of Albert Heijn and the price image of Plus, even though their average price levels are almost similar.
Figure 2 Figure 3
18 This is in line with prior literature that price images formed by consumers are not always an accurate reflection of a retailer’s actual price level (Hamilton & Chernev, 2013; Zielke, 2006). This implies that more factors affect a retailers’ price image, other than just their actual prices. It‘s also notable that the different price image lines show similar increases and decreases. A possible explanation for this could be that general consumer sentiments are affected by publications from consumer organizations and other media about general increases and decreases in consumer prices.
A variable that might also affect a retailers’ price image is the dispersion of product prices offered by the different chains (Figure 6) . The figure shows that Albert Heijns price dispersion is quite low, whereas Jumbo’s price dispersion is higher, especially over time. This means that Jumbo offers some products cheaper than their general price level and some products more expensive than their general price level, while Albert Heijn offers their products on a fairly consistent price level compared to their general price level.
Promotions and Promotion Image
To get an overview on how the promotion policies of the three supermarket chains have developed over the past two years, the figures below show the promotion intensity and average discount of both national brand promotions and private label promotions over time. Figure X shows the development of promotion intensity of national brands over time. It shows that Albert Heijn substantially increased their promotion intensity of national brands over time and that Jumbo also increased their promotion intensity of national brands, while Plus kept their promotion intensity on a fairly consistent level. When looking at the development of the promotion intensity of private labels, there’s a less distinct difference visible between the three supermarket chains (Figure XX). Although it’s interesting to see that Jumbo substantially increased their promotion intensity for private labels over time and even surpassed Albert Heijn in 2018. Albert Heijn not only has a higher promotion intensity for national brands than Jumbo
19 and Plus, it also offers an higher average promotion discount on national brands (Figure 9). Plus however, offers an remarkably high average promotion discount on private labels (Figure 10).
The development of the promotion images of the retailers is presented in Figure 11. As described before, there was a limited amount of observations for the promotion images of Jumbo and Plus from April 2016 till May 2017 and therefore the decision was made to exclude these observation months of Jumbo and Plus. The figure shows that AH has the highest promotion image, followed by Jumbo and Plus, which is in line with their promotional efforts. A notable
Figure 7 Figure 8
20 inverse peak can be spotted for the promotion image of Albert Heijn in 2017. However, after further investigation this extreme value is not considered as an outlier.
5.2 Empirical results
This section will explain the models that have been developed in order to test the hypotheses. Firstly, models that explain the price image of the retailers will be discussed and afterwards, models that explain the promotion image of the retailers will be discussed. All the reported models have included the most optimal variant of the variable, that explains the dependent variable the best. In almost all cases, this is the (t-1) variant: the value for the variable one month before the value of the dependent variable.
5.2.1 Price image models
Firstly, two pooled models are created to analyse to what extent actual pricing and promotion variables can predict a retailers’ price image. A pooled model assumes that all parameters are the same across sections, in this case supermarket chains, which allows for more data points (Leeflang et al., 2015). It is particularly useful in this study, because of the limited observation months covered by this study. Table 3 presents the parameter estimates of the linear regression models. Model 1 is a pooled model with the average price level divided into national brands and private labels, whereas model 2 is a pooled model that only includes the total average price level. Both model 1 (F = 32.634; p <0.01) and model 2 (F = 34.634; p <0.01) are statistically significant, which means the included variables improved both model fits. Both models have a fairly high R2, with model 1 (R2 = 0.773) performing slightly better than model 2 (R2 = 0.753).
21 Model 2 shows that, in support of H1A, the average overall price level of a retailer (t-1) has a highly significant positive effect on price image (β = 0.320; p < 0.01). Model 1 provides evidence for H1B and H1C, that both the average price level of national brands (t-1) (β = 0.246; p < 0.01) and the average price level of private labels (t-1) (β = 2.390; p < 0.01) have a highly significant positive effect on price image. The estimates however are not comparable, because the estimates for average overall price level is based on a €1 increase in the price of a basket of 97 products, whereas the estimates for the average price levels of national brands and private labels are based on a €1 increase in the price of baskets of 80 and 17 products, respectively. Both model 1 and model 2 show a negative effect of price dispersion (t-1), however only model 2 provides an, admittedly, weak significant effect (β = -1.358; p < 0.1). The significant promotion variables are of unexpected directions and it’s likely that the variables serve as proxy variables in this model. This is probably due to the fact that Albert Heijn has a higher price image, while also having deeper and more frequent promotions. Therefore, the use of a partially pooled model is explored. Where a pooled model assumes that all parameters are the same across sections, partial pooling allows to accommodate heterogeneity between the cross
Table 3: Pooled regression model Price Image
Dependent variable: Price Image
(1) (2)
Estimate (St. Err.) Estimate (St. Err.)
Constant -35.628 (23.910) 8.514 (15.943)
Avg. price level national brands (t-1) 0.246*** (0.072)
Avg. price level private labels (t-1) 2.390*** (0.852)
Avg. overall price level (t-1) 0.320*** (0.068)
Price dispersion (t-1) -0.536 (0.850) -1.358* (0.805)
Promo intensity national brands (t-1) -0.096 (0.190) -0.167 (0.194) Promo intensity private labels (t-1) 0.544** (0.238) 0.485* (0.245)
Avg. promo discount national brands (t-1) 0.383*** (0.080) 0.381*** (0.082)
Avg. promo discount private labels (t-1) -0.017 (0.101) 0.067 (0.099)
Observations 75 75
R2 0.773 0.753
Adjusted R2 0.750 0.732
Residual Std. Error 2.437 (df = 67) 2.522 (df = 68)
F Statistic 32.634*** (df = 7; 67) 34.634*** (df = 6; 68)
Akaike Information Criterion 355.98 360.25
22 sections. In this study, an Ordinary Least Regression with Dummy Variables (OLSDV) is developed. This is a special case of partial pooling, where is assumed that the cross sections (supermarket chains) have different overall levels of the dependent variable (Price image), but that the response to marketing instruments is similar across the cross sections (Leeflang et al., 2015).
Because of multicollinearity issues between the estimated dummy’s and the price levels of the retailers (VIF >10), a new variable is developed that measures the monthly percentual
changes in the price levels of the retailers. The variable average promotion discount for national brands is removed, also because of multicollinearity issues with the estimated
dummy’s (VIF > 10). Table 4 presents the parameter estimates for the OLSDV models: model 1 divides the changes in price levels into national brand price changes and private label price changes, whereas model 2 only includes the change in the total average price level. Model 3 is a so-called intercept only model, without explanatory variables. Both model 1 (F = 38.966; p <0.01) and model 2 (F = 45.179; p <0.01) are statistically significant and both models explain more than 80% of the variance in price image (R2 = 0.832). Both models perform better than the pooled models, because of a higher R2 and lower AIC score, with model 2 performing slightly better than model 1. Model 2 is therefore used for the interpretation of the parameters both models have in common. Model 3, without explanatory variables, is also statistically significant (F = 139.564; p <0.01) and explains a large proportion of the variance in price image (R2 = 0.788). However, both model 1 and model 2 still outperform model 3
23
Table 4: Partially pooled regression model Price Image
Dependent variable:
Price Image
(1) (2) (3)
Estimate (St. Err.) Estimate (St. Err.) Estimate (St. Err.) Constant 108.749*** (3.108) 108.766*** (3.086) 107.296*** (0.446)
Dummy Plus -5.527*** (0.963) -5.585*** (0.940) -5.379*** (0.631)
Dummy Jumbo -11.538*** (1.262) -11.536*** (1.253) -10.545*** (0.631)
Change in price level
national brands (t-1) 1.196
** (0.569)
Change in price level
private labels (t-1) 0.277 (0.477) Change in overall price
level (t-1) 1.386** (0.621)
Price dispersion (t-1) 0.321 (0.738) 0.322 (0.734) Promo intensity national
brands (t-1) -0.123 (0.166) -0.131 (0.162)
Promo intensity private
labels (t-1) 0.245 (0.211) 0.239 (0.209)
Avg. promo discount
private labels (t-1) -0.083 (0.097) -0.080 (0.096) Observations 72 72 78 R2 0.832 0.832 0.788 Adjusted R2 0.811 0.813 0.783 Residual Std. Error 2.107 (df = 63) 2.092 (df = 64) 2.276 (df = 75) F Statistic 38.996*** (df = 8; 63) 45.179*** (df = 7; 64) 139.564*** (df = 2; 75) AIC 322.03 320.16 354.58 Note: *p<0.1 **p<0.05 ***p<0.01
24 between change in price level of national brands (t-1) and price image (β = 1.196; p < 0.05), which implies that a 1% increase in a retailers’ price level of national brands will result in a 1.196 increase in price image the month after. The change in price level of private labels (t-1) has no significant effect on price image. This variable only measures a short-term effect of change in price level on price image, so it only partially supports H1D that average price level of national brands has a bigger influence on price image than average price level of private labels. The other variables fail to provide significant effects on price image, however the direction of the promotion variables are more in line with expectations compared to the estimates of the pooled models.
5.2.2 Promotion image models
In order to analyse the relationship between the promotion policy of retailers and the perceived attractiveness of the promotions, a pooled model is developed first. Table 5 presents the parameter estimates of this model. The model is statistically significant (F = 7.682, p <0.01) and explains a decent proportion of the variance of a retailers’ promotion image (R2 = 0.411).
Table 5: Pooled regression model Promotion Image
Dependent variable: Promotion Image
Estimate (St. Err.)
Constant 7.948*** (0.386)
Promo intensity national brands (t-1) 0.062*** (0.022)
Promo intensity private labels (t-1) -0.082*** (0.028)
Avg. promo discount national brands (t-1) 0.013 (0.010) Avg. promo discount private labels (t-1) -0.029*** (0.011)
Observations 49 R2 0.411 Adjusted R2 0.358 Residual Std. Error 0.266 (df = 44) F Statistic 7.682*** (df = 4; 44) Note: *p<0.1 **p<0.05 ***p<0.01
25 promotion intensity of national brands will result in a 0.062 increase in perceived promotion attractiveness the month after. In contrast with H5B, the model shows a highly significant negative effect of promotion intensity of private labels (t-1) on promotion image (β = -0.082; p < 0.01). This implies that a 1% increase in promotion intensity of private labels will result in a 0.082 decrease in perceived promotion attractiveness the month after. The positive effect of promotion intensity of national brands on promotion image and the negative effect of promotion intensity of private labels on promotion image provide support for H5C, that promotion intensity of national brands has a stronger positive effect on promotion image than promotion intensity of private labels. In contrast with H6B, the model shows a highly significant negative effect of average promotion discount of private labels (t-1) on promotion image (β = -0.029; p < 0.01). This implies that a 1% increase in the promotion discount a retailer offers on private labels will result in a 0.029 decrease in perceived promotion attractiveness the month after. The model fails to provide significant evidence for H6A, although the positive direction of the relationship between average promotion discount of national brands (t-1) on promotion image is as expected. This also provides support for H6C, that promotion discount of national brands has a stronger positive effect on promotion image than promotion discount of private labels.
There is not enough difference in the overall level of promotion image across the
supermarkets for a partially pooled model to be appropriate. However, because Albert Heijn is the only supermarket chain with a full 26 months of observations, separate models are
estimated for Albert Heijn (Table 6). Model 1 is not statistically significant (F = 2.193; p > 0.10). This might be due to the ratio between the number of observations and the number of explanatory variables. Therefore, an improved model is developed with only significant variables. Model 2 is statistically significant (F = 5.596; p < 0.01) and explains a decent proportion of the variance in the promotion image of Albert Heijn (R2 = 0.456). Model 2 is
26
Table 6: Regression models Promotion Image Albert Heijn
Dependent variable: Promotion Image AH
(1) (2)
Estimate (St. Err.) Estimate (St. Err.)
Constant 7.757*** (1.019) 7.736*** (0.184)
Promo intensity national brands (t-1) 0.097** (0.038) 0.135*** (0.034)
Promo intensity private labels (t-1) -0.136** (0.058) -0.133** (0.050)
Promo intensity private labels (t-2) -0.101** (0.040)
Avg. promo discount national brands (t-1) -0.002 (0.034) Avg. promo discount private labels (t-1) -0.005 (0.029)
Observations 25 24 R2 0.305 0.456 Adjusted R2 0.166 0.375 Residual Std. Error (df = 20) 0.308 0.268 F Statistic 2.193 (df = 4; 20) 5.596*** (df = 3; 20) Note: *p<0.1 **p<0.05 ***p<0.01
Model 2 provides evidence for a highly significant positive effect of promotion intensity of national brands on the promotion image of Albert Heijn (β = 0.135; p < 0.01), which is in line with H5A. This implies that a 1% increase in the promotion intensity of national brands results in a 0.135 increase in perceived promotion attractiveness the month after. Similar to the pooled model, model 2 also provides significant effects that are in contrast with H5B. Both the promotion intensity of private labels (t-1) (β = -0.133; p < 0.05) and the promotion intensity of private labels (t-2) (β = -0.101; p < 0.05) have a significant negative effect on the promotion image of Albert Heijn. This implies that a 1% increase in the promotion intensity of private labels will result in a 0.133 decrease in AH’s promotion image the month after and a 0.101 decrease in AH’s promotion image two months later.
Model validation
27
6. CONCLUSION AND DISCUSSION
This paper has studied the effects of retailers’ price and promotion policies on their price and promotion images. In this chapter, the main findings of the analyses will be discussed and the research questions will be answered and implications for retailers will be discussed.
Price Image
First, the pooled models showed that the average overall price level of a retailer, as well as the average price level of national brands and the average price level of national brands have a significant positive effect on price image. This implies that retailers with a higher price level have a higher price image and that both national brands and private labels affects a retailers’ price image. The partially pooled models showed that a change in overall price level has a positive significant effect on price image. This implies that when a retailer decreases their prices in a particular month, it will result in a lower price image the month after. However, managers should note that this is not a long-term effect, because the effect disappears when prices stay stable the month after. This effect has also been found for national brands, but not for private labels. This implies that consumers are more aware of price changes of national brands, at least in the short-term. Another interesting finding was provided by the partially pooled model without explanatory variables. The model showed that only the name of the retailer already explained a lot of variation in the price images of Albert Heijn, Plus and Jumbo and that the differences are quite big. This implies that actual pricing and promotion variables only have a relatively small influence on price image formation, which is in line with a claim from prior research that price images tend to be resistant to change (Hamilton & Chernev, 2013). When looking back at the descriptive statistics, it’s also really striking that Plus has a lower price image than Albert Heijn, even though they have an almost similar price level as Albert Heijn, while scoring seemingly lower on promotions. This indicates that it’s hard for retailers to change their price image with pricing and promotion variables, because it appears that consumers estimate price images in an inaccurate way.
Promotion Image
28 intensity of national brands would have a stronger positive effect on perceived promotion attractiveness than promotion intensity of private labels, but the negative effect of promotion intensity of private labels was unexpected. The model also revealed that the average promotion discount of private labels had a negative effect on perceived promotion attractiveness. The separate model of Albert Heijn provided similar results and showed that promotion intensity of national brands had a positive effect on their promotion image, whereas promotion intensity of private labels had a negative effect on their promotion image. However, no significant effects were found for the average promotion discount they offered. Since retailers use promotions to attract consumers to their stores (Grewal et al., 1998), it’s important the promotions are perceived as attractive. The results imply that promotions of private labels are perceived as not attractive and that retailers should focus on offering more national brand promotions and less private label promotions to increase their promotion image.
Time lags
29
7. LIMITATIONS AND FURTHER RESEARCH
Although the study was able to provide some interesting insights, there are also a few limitations that need some attention. First of all, the study only focuses on the three largest traditional grocery retailers in the Netherlands and for example, doesn’t take hard discounters into account. Besides that, the study only covers 26 months, which limited the number of explanatory variables that could be used in regression models. This is due to a limited
availability of online price data, which a select group of retailers only started working with in 2016. Another limitation of this study are the determinations of the time lags. Because of the way the variables were calculated, it was hard to determine the exact time lags of retailers’ price and promotion changes. The way the price image variable is measured could potentially be another limitation and be improved for further research between the effect of promotions and price image. The question asked to the respondents was as follows: “Suppose that a full
shopping basket at the average Dutch supermarket costs 100 euros. However, some
supermarkets are cheaper, others are more expensive. How much do you think that shopping basket costs at the following supermarkets (estimate)?” In this question, promotions are not
mentioned, which could explain why promotion variables failed to have a significant effect on price image. Further research might find significant effects of promotions on price image by lightly pointing the respondents that promotions also could affect the price when buying a full shopping basket. Another limitation of this study might be that for the partially pooled model of price image, no division is made between price increases and price decreases. Prior
research stated that consumers react more strongly to price increases than to price decreases
(Kalyanaram & Winer, 1995). This is also mentioned by another study: “For retailers aiming
30
Reference list
Alba, J. W., Broniarczyk, S. M., Shimp, T. A., & Urbany, J. E. (1994). The influence of prior beliefs, frequency cues, and magnitude cues on consumers' perceptions of comparative price data. Journal of Consumer Research, 21(2), 219-235.
Alba, J. W., Mela, C. F., Shimp, T. A., & Urbany, J. E. (1999). The effect of discount
frequency and depth on consumer price judgments. Journal of Consumer Research, 26(2), 99-114.
Alford, B. L., & Biswas, A. (2002). The effects of discount level, price consciousness and sale proneness on consumers' price perception and behavioral intention. Journal of Business research, 55(9), 775-783.
Blattberg, R. C., Briesch, R., & Fox, E. J. (1995). How promotions work. Marketing science, 14(3), 122-132.
Bustillo, M. & Martin, T.W. (2010). Wal-mart bets on reduction in prices. Wall Street Journal.
URL:https://www.wsj.com/articles/SB10001424052702304198004575172271682347064
Büyükkurt, B. K. (1986). Integration of serially sampled price information: modeling and some findings. Journal of Consumer Research, 13(3), 357-373.
D'Andrea, G., Schleicher, M., & Lunardini, F. (2006). The role of promotions and other factors affecting overall store price image in Latin America. International Journal of Retail & Distribution Management, 34(9), 688-700.
Dickson, P. R., & Urbany, J. E. (1994). Retailer reactions to competitive price changes. Journal of Retailing, 70(1), 1–21.
Eales, T. (2016). IRI Special Report – Price and Promotion in Western Economies: A pause in promotion escalation, IRI Worldwide.
URL: https://www.iriworldwide.com/IRI/media/IRI-Clients/International/Price-Promotion-in-Western-Economies-a-Pause-in-Promotion-Escalation.pdf
Feichtinger, G., Luhmer, A., & Sorger, G. (1988). Optimal price and advertising policy for a convenience goods retailer. Marketing Science, 7(2), 187-201.
Grewal, D., Krishnan, R., Baker, J., & Borin, N. (1998). The effect of store name, brand name and price discounts on consumers' evaluations and purchase intentions. Journal of
retailing, 74(3), 331-352.
Haddon, H. & Nassauer, S. (2017). Wal-Mart brings price war to groceries, boosting pressure
on big food retailers. Wall Street Journal. URL:
31
Hamilton, R., & Chernev, A. (2013). Low prices are just the beginning: Price image in retail management. Journal of Marketing, 77(6), 1-20.
Kalyanaram, G., & Winer, R. S. (1995). Empirical generalizations from reference price research. Marketing science, 14(3), 161-169.
Leeflang, P.S.H., Bijmolt, T.H.A., Pauwels, K., Wieringa, J.E. (2015), “Modeling Markets", Springer
Lourenço, C.J.S., Gijsbrechts, E., and Paap, R. (2015) The Impact of Category Prices on Store Price Image Formation: An Empirical Analysis. Journal of Marketing Research: April 2015, 52(2), 200-216.
Mulhern, F. J. (1997). Retail marketing: From distribution to integration. international
journal of research in marketing, 14(2), 103-124.
Mulhern, F. J., & Padgett, D. T. (1995). The relationship between retail price promotions and regular price purchases. The Journal of Marketing, 83-90.
Nielsen, A.C. (2008), “Grocery Store Choice and Value for Money: A Global Nielsen Consumer Report,” technical report, January.
Van Heerde, H.J., Gijsbrechts, E., Pauwels, K. (2008). Winners and losers in a major price war. Journal of Marketing Research, 45(5). 499-518.
Zielke, S. (2006), “Measurement of retailers’ price images with a multiple-item scale”, The
International Review of Retail, Distribution and Consumer Research, 16(3), 297-316
Zielke, S. (2010). How price image dimensions influence shopping intentions for different store formats. European Journal of Marketing, 44(6), 748-770.
