Mobile Edition
Print
Get the Mag
Weekly UpdatesExecutive Summary
There are many skin cream products for women, and it is very difficult for a single brand to capture the major share of the market. The competition is intense to get the attention of the target audience, and therefore each brand strives to increase the rate at which women try new products and make repeat purchases, through advertising and relative pricing. The purpose of the current study is to quantify the impact of advertising dollars, share of advertising voice, and relative price in order to establish the proper strategic direction of the marketing activities in the efforts to increase or defend the market share of a brand. A nonlinear regression approach is proposed here to study the elasticities of pricing and advertising efforts. The frequently changing marketing environment, with its many new product introductions, would make the long historical data irrelevant, and therefore a few quarters of the immediate past had to be analyzed in order to arrive at conclusions that could be appropriate to the immediate future.
Introduction
Marketing managers make strategic decisions almost every quarter regarding the fixing of proper price levels and advertising expenses, in order to maintain or enhance their brand's market share under the constant onslaught of changing competitive marketing activities of existing and new brands in the market place. The brand manager wants to know how factors such as pricing, advertising, distribution, and product quality affect his brand's share of the marketplace, so that he can determine the correct response to boost his brand's image in the eyes of consumers.
Most of the published research is too complex for easy implementation or easy interpretation. Moreover, the results were based on cross-sectional data and analysis, which most of the time is not relevant to individual brand decisions. The cross-sectional models lack the reliable predictive ability to be useful for a marketing manager of one product.
From the researcher's point of view, they require a large amount of data and complex mathematical functions to quantify the impacts and interactions of activities, and as a consequence the approaches proposed become unwieldy and impractical. The cross-sectional data across brands within an industry, or across brands and industries, may help managers understand the various factors better, but would give no relevant assistance to a single brand.
In order for studies to be helpful in making individual brand decisions, the studies should attempt to analyze individual brands within each industry, and should identify the similarities and differences among brands and industries.
Some Past Market Share Models
There were many modeling approaches to predict the impact of various factors on market share and sales in various industries. The popular ones are linear, additive and attraction models. The most common problem encountered by these marketing strategy researchers is the limited availability of appropriate time series data, especially on competition. As a consequence, the number of factors included in a model was restricted, so the study was small. The emphasis on most published studies has been on a few factors such as price, distribution, advertising and promotions. For example, Bass (1969) concentrated on advertising; Lambin (1972) studied distribution and advertising, and Kuehn, Mcguire and Weiss (1966) looked at price and advertising.
Some authors, instead of making market share the dependent variable, considered the change in market share from the previous period as the dependent variable. Buzzell and Wiersema (1981) utilized this approach to study the reasons for fluctuations in market shares, and they also resorted to the cross-sectional analysis to get around the problem of limited data availability over time. The cross-sectional analysis has its own limitations because it is not relevant to obtain predictions and inferences about individual brands or markets. As a result, most published literature that had a predictive emphasis had to rely on less than sixteen time series observations for their analysis.
The published attraction models are based on Kotler's (1971) fundamental theorem of market share determination, which states that market shares of competitors will be proportional to their marketing efforts. The attraction models faced the same difficulties as the linear predictive time series models because these models require accurate information about the competition, which is not easy to get for many time periods. Kuehn, Mcguire and Weiss (1966) took the attraction modeling approach to analyze market shares. Brodie and Kluyver (1984) compared the Linear and Attraction Models in terms of predictive power.
There were also some published models to identify optimal levels of marketing activity mix, using a game theoretic approach called the Lanchester model of combat for competing brands. Wang and Wu (1974) used the model for competing advertising decisions.
Some authors felt that marketing activities impact market share, and market share in turn influences the levels of factors such as advertising expenditures. Therefore, they suggested the simultaneous econometric equation approach for studying the interactions. Bass (1969) was one of the early researchers who attempted to estimate the effects of mutual dependence between market share and exogenous variables.
Cobb-Douglass Framework and Earlier Applications
The Cobb-Douglas equation is a function relating output and inputs, with output as a dependent variable and inputs as exogenous variables. The function was originally suggested by Knut Wicksell (1851-1926). It was tested by Cobb and Douglass in 1928 as a production function in the following form.
Y= A [L.sup.a] [K.sup.b] where
Y= Total Production
L= Labor Input
K= Capital Input
a and b are labor and capital elasticities of the output. It has been applied extensively in the production environment in various industries. The function, despite its nonlinear multiplicative form, can sometimes be linearized by logarithmic transformation. It depends upon the complexity of the format of the input variables, and in the case of involved input formats, complex parameter estimation procedures need to be utilized.
The researchers in other functional areas, such as marketing, are able to use the Cobb-Douglass framework to study the impact of various factors on dependent variables, such as sales. You can find extensive applications of it in the energy industry. Reinhard Madlener gives a good account of the literature on energy demand modeling. Reid and Bradford (1983) used the Cobb-Douglass functional form to predict the percentage of original value remaining for agricultural equipment as a function of factors such as age in years, and the horsepower of the machines. Parameters were estimated using a log-linear transformation and ordinary least squares.
Lott and Warner (1974), using the Cobb-Douglass function, studied the functional relationship between campaign expenditures, percentage of registered voters associated with a candidate's political party, and the election outcome.
In microeconomics, the Cobb-Douglass function has been proposed sometimes to study the utility as a function of consumption of different goods in order to establish the demand elasticities of the utility.
Current Research
Most of the published market share models in the past were not useful to the marketing manager who is confronted with the marketing activity decisions. The models are not relevant to him or her because either they are too general, with the analysis across brands and industries, or they are too complex to understand, implement and interpret. Most of the models are cross-sectional due to the difficulty in obtaining reliable data, especially about competition for a large number of time periods. As a consequence, the number of observations is much smaller than 20 observations for most of the time series studies of market share. Moreover, their emphasis has been large market-share brands, results of which may not be relevant to the small-share brands.
The objective of the current research is to propose an easily implementable and interpretable model with Cobb-Douglas framework, which can quantify the impacts of pricing and advertising strategies, and which can consider the interactions between pricing and advertising. The model is useful for both large-and small-share brands.
It will be demonstrated here that both the absolute advertising expenditures and the share of the advertising voice are relevant for market share. Previous research considered only one of these two advertising variables. Many previous studies used relative price by dividing a brand's price by the average competitive price of all competitive brands. The relative price in the present study is obtained by dividing its own price by the main competitor's price. This approach is more practical and less expensive, so it appeals to the brand management. Usually, a brand tends to concentrate on one main competitor.
The main purpose of the investigation here is also to demonstrate that Cobb-Douglas production function can be successfully and easily implemented to estimate relationship between market share and marketing activities. This model has an added advantage to the management because it readily gives elasticities of market share. The model presented here will be shown free from multicollinearity and heteroskedasticity.
The product that is being analyzed is a non-seasonal packaged, consumed product used by women. The brand has about 7% market share, and the objective of the analysis is to explore the reasons for the quarter-to-quarter unit share fluctuations in order to identify the key marketing strategies to improve unit share.
FEED