Is the placement of the products going to influence their price?
Minjun Yu
ECON 4980

Introduction
In today’s commodity economy, it is no longer only about manufacturing, transportation, cost, profit, supply, demand, etc. There are many other factors that matters in the sales of the commodity. For many years, retailers have been searching for methods to help them identify ways to improve their merchandise presentation (Ron Larson).One of the methods that retailers used to improve the sales is the display of the products in the aisles. Bezawada has said in his reach that “retailers can benefit substantially by having better placement of items in their aisles,” (Ram Bezawada). Bezawada’s research shows that aisle placements can influence sales across product categories as much as other marketing variables, such as price or how an item is displayed (Jacqueline Ghosen, 2009). According to some researches, changing the allocation of the products will increase the weekly sales by more than 9% (Jacqueline Ghosen, 2009). From all these research, we can see that products placement is one of the most important factor that going to influence retailer’s profits. Putting products in the right place will bring more benefits.
In marketing field, there are also some important theories about placement of products. It is called marketing mix which is usually refers to E. Jerome McCarthy's 4P classification for developing an effective marketing strategy. It encompasses product, price, placement (distribution) and promotion. These elements are adjusted until the right combination is found that serves the needs of the product's customers, while generating optimum income. Among the four Ps, place (distribution) is about getting the products to the customer. Although figures vary widely from product to product, roughly a fifth of the cost of a product goes on getting it to the customer. 'Place' is concerned with various methods of transporting and storing goods, and then making them available for the customer. Getting the right product to the right place at the right time involves the distribution system. The choice of distribution method will depend on a variety of circumstances. It will be more convenient for some manufacturers to sell to wholesalers who then sell to retailers, while others will prefer to sell directly to retailers or customers (McCarthy, J).
That brings my attention to the question that is the placement of the products really that important? Are we being charged higher price just because we pick the products from their “right” place? We all have to go grocery shopping and when I go along the aisles, I see the placement of the same products in different stores have some distinction and the price of them are quite different. I understand the ultimate goal for stores is to maximize the profit. So I think maybe they will put the most profitable products in the place that most people will buy. In here, the most profitable maybe means the highest price or the brands that are carried by the stores themselves so they will get other kinds of benefits. Finally I come up with my research question that is the placement of the products going to influence their price?

Abstract
I used the hedonic pricing model to do the regression analysis about the cereal products. I collected 152 cross section data in five different stores and have total 31 variables. The focus variables are about the locations in the shelf. After the regression analysis I found that the placements of products in the shelf do influence the price and the products that in the middle areas in the shelf have higher prices.
Literature review
1.
In 1979 Karen J. Morgan, Edward J. Metzen and S. R. Johnson mentioned in the paper An Hedonic Index for Breakfast Cereals that “According to the commodity characteristics theory of consumer demand, the price individuals are willing to pay for a product reflects the value they place on commodity characteristics. As an outgrowth of this theory, hedonic indices reflecting implicit prices of commodity characteristics have been estimated and applied in analyzing product quality change, pricing practices of firms, and buyer behavior”. They decided to use breakfast cereal as a research object because they are a labeled, regularly purchased food product and contain nutrient supplements. The data they used for analysis were supplied by state purchasing agents in the United States for the period of 1972 through 1976. Nineteen dietary components, five other product characteristics," and four market factors were selected as the independent variables to be estimated for the hedonic index. They had the assumption that buyers are processing the detailed dietary information available on the cereal label in an altered form. With the assumptions on grouping of dietary variables, the model for estimating the coefficients for the hedonic index was specified as:
Pit=j=1nβjt Xjit+r=1mδrt Zrit+ μit where Pit is the cereal price per ounce, Xjit represents the dietary characteristics, Zrit denotes the non-dietary characteristics and market factors, μit, is the structural disturbance, and i and t indicate the cereal and time period, respectively. The parameters βjt, and δrt are to be estimated for forming the hedonic index.
In the paper the variables the authors used have dietary component like protein, vitamins, minerals, vitamin variance, mineral variance, fiber, calories; package size; class of cereal like “Kids' stuff" or traditional; type of processing like flaked, shredded, granulated, oven-toasted, milled; production company like General Foods, General Mills, Kellogg, Ouaker Oats, Ralston Purina, Standard Milling; principal grain like corn, wheat, oats, rice, barley.
The results indicate that in the selection of breakfast cereals, buyers are not being guided completely by nutritional concerns in their use of product labels. Although purchasing agents appear to have a limited awareness of some nutritional information, as shown by their willingness to pay higher prices for the dietary components found in high nutritive cereals, their knowledge about nutrition and/or their application of such knowledge is limited. Recommendations for continued research in this area include collecting information from the retail market rather than the institutional market, sampling to permit results that can be more readily generalized, and extending the model to include additional characteristics of breakfast cereals.

2.
Valdimar Sigurdsson, Hugi Saevarsson and Gordon Foxall talked about products placement and consumer choice in the paper Brand Placement and Consumer Choice--An in-Store Experiment in 2009. They mentioned that one of the most influential models in marketing is the marketing mix which is designed to stimulate and influence consumer demand for brands by manipulation of variables associated with the brands. One variable that appears to affect consumer buying behavior is shelf placement. They also mentioned that the Wal-Mart consumer and market knowledge team used an eye tracking method to measure consumers’ observing behavior in stores and the results indicated that most shoppers failed to look at one third to one half of the brands on the shelf; shoppers looked mostly at the products in the center of the shelf. In fact, shoppers looked at the brands positioned in the center of the shelf nine times more than those placed in the corners. However, these findings are based on proprietary research. Therefore, the reliability and validity of these findings are not known. The authors tried many ways to try to get a reliable and valid result. One method of assessing the effects of shelf placement on consumer behavior is to manipulate the placement of items on shelves in stores and measure consumer behavior. Finally, they chose to use in-store field experiments which are well known in marketing practice in which simple between-groups and A-B-A research designs are more common.
The data they used refer to the sales volume of every brand in the potato chip product category (24 brands, including the target brand) sold in two budget stores in Reykjavik, Iceland. The target brand of potato chips was an international brand with a salty taste, and the other brands were of similar type (salty) or had other flavors. Shelf height was the same in both stores for the low (24 cm), middle (123 cm), and high (173 cm) shelves. The dependent variable was percentage of units sold of the target brand, which was calculated by dividing the units sold of the target brand by the units sold of the 24 brands of potato chips in the store. This proportion was grouped into periods that consisted of unit sales from Friday to Sunday or Monday to Thursday to control for fluctuations in sales due to day of the week.
The results showed that relative sales of the target brand were different, depending on shelf location (middle, M =7.5%, range, 6.6% to 8.6%; low, M 5 4%, range, 2.9% to 5.8%; high, M 5 3.3%, range, 2.7% to 3.6%). Relative sales of the target brand were highest when it was on the middle shelf.
They got the conclusion that the target brand’s relative sales against its product category were higher when it was placed in the middle shelf compared to the high or low shelf. One explanation for these results is that the response effort associated with looking at middle shelves is lower than that associated with looking at the high or low shelf.

3.
In the paper Core Principles for Supermarket Aisle Management the author Ron Larson mentioned that “aisle management, defined as dividing a store into clusters or zones (eg, aisles) and striving to increase traffic, sales, and profits from each of those zones, is one option several companies are using”. This is most of the grocery stores do nowadays. He also said that “aisle management shifts the focus from doing separate analyses for many small categories to doing one analysis for the group of products in a store”. That’s most of the products in the grocery store are allocated. In the paper Ron talked many researches that have done already. He mentioned that Dreze, Hoch, and Purk (1994) found that organizing ready-to-eat cereals by type reduced category sales by 5 percent and alphabetizing canned soup reduced sales by six percent. One common principle is that brands should always be grouped together to create “billboard” effects. He also said that what a store emphasizes can influence how shoppers think about products. For example, a headline for a display may encourage shoppers to focus on the attribute in the sign. For example Areni, Duhan, and Kiecker (1999) found that when shoppers saw a display for “Texas Wine,” they reduced their purchases of Wines from Texas and increased their purchases of Wines from other locations. These examples illustrate the importance of how products are organized on the shelf. About the location part Ron mentioned that Phillips and Bradshaw (1993) found that end-aisle displays gathered more attention if they did not include a large number of product lines

Theoretical section
Hedonic pricing model
Hedonic pricing model is a model that identifies price factors according to the premise that price is determined both by internal characteristics of the good being sold and external factors affecting it. It means that the starting point of every hedonic price index is the hedonic hypothesis. The core of this hypothesis is that each good is characterized by the set of all its characteristics. For any given good, let this set be ordered and denoted by x = (x1, ……,xn). It is assumed that the preferences of the economic actors with respect to any good are solely determined by its corresponding characteristics vector. Furthermore, it is assumed that, for any good, there is a functional relationship f between its price p and its characteristics vector x, i.e. p = f(x) . ------------------ (1)
Based on the functional relationship (1), the important concept of implicit or hedonic prices can be introduced. These prices are defined to be the partial derivatives of the hedonic function (1), i.e., they are defined through ∂p∂xnx=∂f∂xn(x) n=(1,2,……,n) ----------------(2)
The hedonic price ∂f /∂xn(x) indicates how much the price p of a good change if this good is endowed with an additional unity of the characteristic xn (Brachinger, Hans Wolfgang).
This idea of the hedonic hypothesis has been advanced long ago by Lancaster and Rosen in their famous articles. Sherwin Rosen proposed a main approach that contributed greatly towards the theoretical framework on hedonic prices in 1974. This approach aims to price products attributes, considering them as the elements generating utility for the consumer (Estrella Orrego).
Rosen suggests there are competitive implicit markets where implicit prices for embodied product attributes are defined and that consumers evaluate product characteristics when making purchasing decisions. Therefore, the observed price of a given good is the linear combination of the quality attributes where the implicit prices are the attributes' weights. Product Xj market implicitly reveals a function Pj = Pj (Z1, Z2,..., Zn) relating prices P and characteristics Z. Rosen supported his view on the idea that "when goods can be treated as tied packages of characteristics, observed market prices are also comparable on those terms" (Rosen).

Empirical section
Variables & Data
My research question is that “Is the placement of products going to influence the price?” I used cereal as my example, and I went to five grocery stores in Ogden area to collect the data. My research question applies to the Ogden Utah area where I live. I think this question can be studied in any area, but I decided to narrow it down so that it will easier for me to do it.
Obviously, my focus variables will be the locations in the shelves. According to my observation in stores, I found the shelves usually have five layers and about one whole aisle that are placed with cereal products. And the cereals are arranged longitudinally with brands. According to my literature review three, “organizing ready-to-eat cereals by type reduced category sales by 5 percent and alphabetizing canned soup reduced sales by six percent”. Stores always organized products vertically with the same brand. “One common principle is that brands should always be grouped together to create “billboard” effects” (Ron Larson). So I defined the vertical location of the shelf as “right”, “middle” and “left”. Although the length of the selves in different store is not the same I still decided to mark them as these three parts because I will do this with equal proportion. So the meaning of right, middle and left will be the same.
I separated the location variables into total eight. They are layer1, layer2, layer3, layer4, layer5, right, middle and left. The order of layers is from the top to bottom. These will be my focus variables that appear in my regressions. What I want to do is to find the relationship between prices and the location. According to my literature reviews and the hedonic pricing model, I think people will value the location of products as one of the characteristics that affect the price. In order to find other characteristics that matter with the prices, I have to find what other paper have already found. Because my research objects are cereal products, I think it will be necessary for me to understand something about this kind of product and its whole industry.
From my literature reviews, I have found many non-focus variables for my research. In the paper An Hedonic Index for Breakfast Cereals, the variables the authors used are dietary component; package size; class of cereals; type of processing; production companies; principal grains, etc. I chose my non-focus variables mainly based on this.
My non-focus variables are size; the dietary components: calories, total fat, sodium, protein, sugar, fiber; the processing type: flaked, shredded, granulated; the brands: Cheerios, Kellogg, Quaker; the stores: Smith’s, Walmart, Target, Fresh Market, Macy’s; the principal grains: corn, wheat, oats, rice, barley. I have total 23 non-focus variables. Among these variables, size and dietary components are continuous variables and the rest are dummy variables.
For size, I choose the unit of measurement as “cups” which is a common measurement in cereal products. So I measured the dietary components by the same amount, one cup. This will make the data have comparability.
For the brands, I did some researches and found that “the cereal industry is an oligopoly with four large companies and a few very small niche companies. The small niche companies hold very little market shares, approximately 13.6% in 1995, with the big four holding the other 86.4%. The big four manufacturers are Kellogg, General Mills, Post, and Quaker Oats” (Matthew Roy). According to this I chose three big brands for my data. They are Cheerios, Kellogg and Quaker. Because Kellogg has too many products and it’s located pretty separately, so I only choose one of its collections, the Special K as my research object.
For the stores, I chose five stores that located nearby. I found that same brand products have very different locations in different stores. But even they have similar locations, the prices are quite different. For example, the original flavor of Special K in target was in the layer3 and the left of the shelf. It was charged 2.99. But the same one in Walmart was charged 2.96 even though the location was almost the same. So I chose to contain stores as my non-focus variables because I believe the types of store will also influence the price. I also found the same products with different locations do have different prices. For example, the original flavor of Special K in Smith was located in the layer5 and right of the shelf. It was charged 3.57 dollars. The same product in the Macy’s was located in layer4 and the middle of the shelf. It was charged 3.69 dollars. This example seems can answer my research question, but I still have do the data analysis to make sure.
For the dietary components, I searched online for every product in my research and found the numbers of every component. Sometime the number it provides was based on different units of measurement like ¾ cup. I manually transformed them into the same units of measurement, one cup.
For the processing type, I mainly collected the data by observation. I only chose three type of processing types and the products I observed was almost fit the descriptions.
For the Principal grains, I read the ingredients table of every product and found the main grains it contains.
The type of my data is section data. I collected all the data in the same day and in different places. First I took photos of the whole aisle in one store, and then I input all the data into Excel to organize them. After I organized all my valid data, I got 152 data of all.
Regression analysis
My dependent variable Y will be the price and the independent variables are my focus variables and non- focus variables. The regression will be like following:
Y=B0+B1*BRANDS+B2*INGREDIENTS+……+Bn*Xn+Bn+1*left+Bn+2*middle+Bn+3*right+Bn+4*layer1+ Bn+5*layer2++ Bn+6*layer3+ Bn+7*layer4+ Bn+8*layer5
First I ran an overall regression includes all my variables. The results are like following:
Price = 9.51 + 0.142 kellogg - 3.65 cheerios - 0.338 Layer 1 - 0.270 layer 2 - 0.168 layer 3 - 0.182 layer 4 - 0.188 left + 0.133 middle + 0.0901 size - 0.0523 calories + 0.451 total fat + 0.00308 sodium + 0.121 protein + 0.256 suger - 0.148 fiber - 3.17 flaked + 5.19 shredded + 0.312 corn + 0.214 wheat + 0.379 oats + 0.061 rice - 0.178 smith - 0.807 walmart - 0.735 target + 0.163 fresh market Because many of my variables are dummy variables, so minitab automatically removed some variables that are highly correlated with other X variables, like quaker, layer5, right, granulated, barley and macy.
The results of every variable are as following:

As the result showed, among my non-focus variables, there are many ones showed with very small P value, which means the relationship between them and price is highly significant. Like Cheerios, size, calories, total fat, sugar, flaked, shredded walmart and target they have the P values equal to 0. Cheerios has negative relationship with the price. The coefficient is -3.647, which means keeping other variables constant, if the brand of the cereal products is Cheerios, the price of product will decrease by 3.647 dollars. Actually I think it’s strange because the average price of cereal products is around 3 to 4 dollars. Of the brand type will influence the price by 3 dollars; it will leave the producer no benefits. The reason for this strange result maybe because that the data I collected about Cheerios are not comprehensive enough. The variables like size, total fat, sugar and oats have positive relationship with the price. It’s easy for me to understand the size. Because more of the inside will cost more, so the price will be higher. About the total fat and sugar, I think these two ingredients make the cereal tastier even though they are not good for people’s health. People value the taste of cereal as one of the characteristics that determine cereal’s price. So the more fat and sugar the cereal contains, the higher it price will be. As calories, it is the measurement of food energy. It won’t necessarily make food more delicious and people especial women are avoiding absorbing too much calories so that they will gain weight. The coefficient of calories is -0.052278 which is small but still is negative. That means one unit increase in cereal’s calories; the price will decrease by –0.052278. And about my focus variables, I only found layer1 is significantly related to price. Because it is one of the dummy variables in my focus variables, so I will remain all of them. The regression analysis didn’t show any significant relationship between left, middle right with the price, so I decided to run second regression that contains my focus variables and all the non-focus variables which are significantly related with price. And the results are as following: price = 8.93 - 3.33 cheerios + 0.159 quaker + 0.009 layer 2 + 0.114 layer 3 + 0.151 layer 4 + 0.316 layer 5 + 0.331 middle + 0.136 right + 0.101 size - 0.0517 calories + 0.397 total fat + 0.00220 sodium + 0.125 protein + 0.260 suger - 0.119 fiber - 2.73 flaked + 5.52 shredded + 0.428 oats - 0.802 walmart - 0.740 target - 0.007 macy

The results of every variable are as following:

Form this table. I can see among my focus variables, layer5 and middle are significantly related to price and the coefficients are both positive, which means the products in those areas will have higher prices.
According to my literature reviews, the middle location is most important in the products placement. So I created five new variables to re-identify the middle location. They are middle*layer1, middle*layer2, middle*layer3, middle*layer4 and middle*layer5. They are the products of two dummy variables. So actually they represent five layers in the former middle area.
I put these new variables together with my focus variables and non-focus variables that are significantly related to price. The regression was showed as following: price = 8.95 - 3.08 cheerios + 0.297 quaker - 0.430 layer 2 - 0.351 layer 3 - 0.235 layer 4 + 0.080 layer 5 + 0.195 middle + 0.168 right + 0.100 size - 0.0520 calories + 0.358 total fat + 0.00262 sodium + 0.125 protein + 0.266 suger - 0.115 fiber - 2.53 flaked + 5.79 shredded + 0.392 oats - 0.749 walmart - 0.718 target - 0.037 macy - 0.452 middle*L1 + 0.579 middle*L2 + 0.537 middle*L3 + 0.305 middle*L4
Minitab automatically removed middle*layer5 because it was highly correlated with other new variables. And the results of every variable were showed as following:

The results showed that middle*layer2 and middle*layer3 have significant relationship with price and the coefficients are positive. This result means products that located in the layer 2 and 3 in the middle area will have higher prices.

Conclusion After I did all the work, I got some conclusions. First, placements do influence the price. Second, products in the middle of the shelves have higher prices. Third, products in the central location of the shelves have higher prices.

References

Brachinger, Hans Wolfgang, 2002. "Statistical Theory of Hedonic Price Indices," DQE Working Papers 1, Department of Quantitative Economics, University of Freiburg/Fribourg Switzerland, revised Aug 2003.
Estrella Orrego, M., Defrancesco, E., & Gennari, A. (2012). The wine hedonic price models in the Old and New World: state of the art. Revista De La Facultad De Ciencias Agrarias,44(1), 205-220.
Lancaster, K.J.: A New Approach to Consumer Theory. Journal of Political Economy, 74, 132–157 (1966)
Marketing theory,(2013), The times 100, from: http://businesscasestudies.co.uk/business-theory/marketing/marketing-mix-price-place-promotion-product.html#axzz2QqRkvRp4
Matthew Roy, 2004, Analysis of the Cereal Industry, reserved from: http://wwmr.org/ip marketing/cereal.htm
McCarthy, J. (1960). Basic marketing: A managerial approach. Homewood, IL: Irwin.
Rosen, S. 1974. Hedonic prices and implicit markets: product differentiation in pure competition. The Journal of Political Economy. 82(1); 34-55.