Introduction:

The relationship between marketing activity and sales is extremely messy. One particular interaction which has attracted a lot of research interest has been the “Share of Voice” vs “Share of Market” relationship. Share of voice is taken to be the percentage of the proportion of viewers of advertising across the category who viewed the adverts for the company in question. The Share of market is the percentage of the proportion of sales for the company in question vs the whole market.We can mark this as the first thesis from the research:

That there is a correlation between SoV and SoM, with SoV driving SoM. That there is a (positive) causal connection from SoV to SoM going forward.

The industry we have examined is the Australian Beer market. The Australian Beer market operates as a duopoly between Lion-Nathan which is owned by Kirin a Japanense conglomerate and Carlton United Brewry which is owned by British SABMiller. Beer consumption has actually declined compared to historical trends, and many strategists believe that there will be a movement to craft beer especially the highly touted Gair & McDonald label (Ferguson 2013).

Previous Research

The seminal paper on the relationship between Share of Voice and Share of Market was Jones 1990. He showed a strong relationship between the Share of Voice (SoV) and Share of Market (SoM) with some varience depending on the size of the market share. Those with low market share having excessive voice and those with high market share having a a SoV lower then their SoM.

In particular he found that depending on the maturity of the product the Share of Voice may be over or under the Share of Market. If you had the domiant brand you could invest less and maintain your bussiness (though there are limits). Conversley for small market share brands the equilibrium is to have excessive SoV.

That small companies tend to need to spend more on SoV and encumbents with big market share can spend less but still maintain their market.

Heimonen & Uusitalo (2009) studied the Finnish Beer market from 1994 - 1999. This is a particularly useful period to be studied as at the time the Finnish Beer Industry had a number of characteristics which amelioate some of the difficulties of other studies.

This removes a lot of the factors that might muddy the situation, making the impact of advertising volume clearer.

Some of their main findings were:

“Overall, the elasticity of market share with respect to advertising expenditure was low and statistically insignificant” (Heimonen & Uusitalo 2009 p. 967)

The dominant brand benefited from advertising expenditure by smaller competitors but the inverse was not true.

They also found that occasionally an increase in marketing expenditure could shrink the market share of the advertising company.

Futher research by Lalanananda (2007) and Eagle, Kitchen & Rose (2005) have shown that the shape of these Advertising Intensiveness Curves can vary depending on the type of market - High vs Low Voice markets and High Involvement vs Low Involvement markets. Interestingly the Australian beer market is undergoing potential structural change.

Our Findings

We struggled to find matching data across the two databases. We did however find that our data matched the previous research. The main difference is that with the trend for craft beers the major brewers are sinking alot of marketing spending into individual major brands. It is unclear which of these views will prove correct.

## 
## *********************************************************************
## Kruschke, J. K. (2015). Doing Bayesian Data Analysis, Second Edition:
## A Tutorial with R, JAGS, and Stan. Academic Press / Elsevier.
## *********************************************************************

As can be seen there is a general relationship between the excessive Voice and an increase of the next years Share of Market. However the p value is not significant (due to our small number of samples).

Frequentist Regression:

## 
## Call:
## lm(formula = Diff ~ VolMVoi2012, data = data)
## 
## Residuals:
##        1        2        3        4        5        6 
## -0.05149 -0.06343  0.11133 -0.06938  0.04711  0.02585 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.767e-17  3.340e-02   0.000    1.000
## VolMVoi2012 2.572e-01  1.826e-01   1.408    0.232
## 
## Residual standard error: 0.08181 on 4 degrees of freedom
## Multiple R-squared:  0.3315, Adjusted R-squared:  0.1643 
## F-statistic: 1.983 on 1 and 4 DF,  p-value: 0.2318

Bayesian Regression:

## Compiling model graph
##    Resolving undeclared variables
##    Allocating nodes
##    Graph Size: 34
## 
## Initializing model

##                  ESS      mean    median       mode hdiMass     hdiLow
## Param. Val. 11234.14 0.1561745 0.1286615 0.09977192    0.95 0.05071972
##               hdiHigh compVal pGtCompVal ROPElow ROPEhigh pLtROPE pInROPE
## Param. Val. 0.3376215      NA         NA      NA       NA      NA      NA
##             pGtROPE
## Param. Val.      NA

Example Adverts:

Heineken

James Squire

Peroni Nastro Azzurro

Marketing References:

Eagle, Kitchen & Rose (2005) “Defending brand advertising’s share of voice: A mature market(s) perspective” Journal of Brand Management Vol. 13 Iss. 1

Ferguson (2013) “Bitter battles in what looks to be an ailing market” Sydney Morning Herald 09/02/2013 (http://www.smh.com.au/business/bitter-battles-in-what-looks-to-be-an-ailing-market-20130208-2e3wx.html)

Hansen & Chrisensen (2004) “Long-term Advertising Effects and Optimal Budgeting” Center for Marketing Communication Research Paper Nr. 2

Hansen & Christensen (2005) “Share of voice/share of market and long-term advertising effects” International Journal of Advertising, Vol. 24 Iss. 3, pp. 297-320

Heimonen & Uusitalo (2009) “The beer market and advertising expenditure” Marketing Intelligence & Planning Vol. 27 No. 7

Jones (1990) “Ad Spending: Maintaining Market Share” Harvard Business Review, Iss. January 1990 (https://hbr.org/1990/01/ad-spending-maintaining-market-share)

Lalanananda (2007) “Advertising Expenditure as a Determinant of a Brand’s Share of the Market: A Case with Low and High Involvement Products” Master thesis in Business Administration Faculty of Economics and Social Sciences Agder University College

McDonald (2010) “A brief review of marketing accountability, and a research agenda” Journal of Business & Industrial Marketing Vol. 25 Iss. 5 pp. 383 - 394

Yoo & Mandhachitara (2003) “Estimating Advertising Effects on Sales in a Competitive Setting” Journal of Advertising Research

Technical References:

Gelman & Hill (2007) “Data Analysis using Regression and Multilevel/Hierarchial Models” (http://www.stat.columbia.edu/~gelman/arm/)

Kruschke (2015) “Doing Bayesian Data Analysis, Second Edition: A Tutorial with R, JAGS, and Stan”

Plummer (2013) “JAGS: Just another Gibbs sampler” (http://mcmc-jags.sourceforge.net/)

Plummer (2014) “rjags: Bayesian graphical models using Markov Chain Monte Carlo” (http://CRAN.R-project.org/package=rjags)

RStudio Inc (2015) “RPubs” (https://rpubs.com)

RStudio Inc (2015) “RStudio” (http://www.rstudio.com/)

Wickham (2015) “dplyr: A Grammar of Data Manipulation” (http://CRAN.R-project.org/package=dplyr)

Wickham (2009) “GGplot2: elegant graphics for data analysis” (http://ggplot2.org/)

Wickham (2014) “GGvis: Interactive grammar of graphics” (http://CRAN.R-project.org/package=ggvis)

Xie (2015) “knitr: A General-Purpose Package for Dynamic Report Generation in R” (http://yihui.name/knitr/)