JOSEPH B. RICHARDS
CPS-615 Fall 97
Email: jbrichar@mailbox.syr.edu

Date: 9/11/97

SIMULATION OF MARKETING SYSTEMS

INTRODUCTION:
 

Marketing system of a company is a complex interplay of factors involving market size, target customers, technology level, competition, financial strength, raw materials, distribution systems, pricing variables and a host of other variables. In most of the cases the information available for taking effective decisions is poor. In practice, decision making in real life marketing involves a lot of subjective judgment , and this is partly due to the difficulty in identifying an objective relationship between the innumerable variables involved. With the advent of high performance computers and information processing software , Marketing systems modeling and Simulation will enable real time and objective decision making.

ELEMENTS OF MARKETING MODEL
 

The most elementary marketing system is made up of the company and the market. The company is related to the market through a set of basic flows. The company dispatches goods, services, and communications to the market; in return it receives dollars and information.

Further going into the details, the marketing system can be divided into the following aspects. For illustration , I will here talk about a Bicycle manufacturing company.

The environment ,or, more precisely, those forces in the environment that effect Bicycle market such as the population growth , per capita income, attitude towards the Bicycle and so on.

The company and the competitor's marketing decision models

The major categories of decision making in the market- Product characteristics, price, sales force, physical distribution and service, advertising, promotion etc.

Distribution channels adopted, e.g.: direct sales( mail order, company sales), dealer sales, and other sales channels(Wal Mart etc.)

The buyer behavior model which shows the customer response to the activities of the Company , and the distribution channels as well as the environment.

Total industry sales and the market share of the company.

The Marketing decision maker has to grapple with above variables to arrive at the most ideal decision at any point of time. The system of these variables can be broken down to coherent components for analysis. In analysis of these components, we will describe one typical approach called Input-Output model.
 

Inputs of the model are the factors affecting the company marketing decisions and the outputs are the marketing decisions. For e.g.:

INPUTS
 
           Target Sales Growth Population Growth Market Response

Target return on sales Disposable Income Competitive outlook

Target ROI Cultural factors

Target profit Cost outlook

Demographic Factors
 

OUTPUTS

Wholesale price Product Characteristics

Trade allowances and Discounts Packing Characteristics

sales calls and service Retail Price

Trade Advertising and Promotion Consumer Advertising

Trade credit policies

Delivery Policies
 

Each input and output factors can be elaborated further. For example, Cultural factors can be further broken down into factors like Fitness consciousness, Energy consciousness, pollution consciousness, Automobile purchase attitude etc. All these sub factors will definitely determine the Number of Bicycles sold.

The product characteristics can be broken down into different product usage segments like , Utility transport bike , Competition Sports bike, Recreation bike, etc. The size of these products segments and buying characteristics will affect the number of bikes sold.

Having identified the major inputs and outputs of the company marketing decision model, we can proceed to trace how these data feed into other parts or components of the system. The output of one element of the system will form the input of the next element of the system.

For example, The Outputs stated before (Wholesale price , Service, Trade allowance etc. will form an input to the Distribution system model. The Distribution system model Can be composed of the following sub models.
 

• Dealer sales model
• Supermarket Model
• Mail Order model

The behavior of all the above sub models are dependent upon the above stated inputs. Further the outputs of the Distribution model will serve as one Input to the Ultimate Buyer behavior model. The buyer behavior model will have further inputs from the Product Characteristics model, and Advertising and promotions model.

The buyer behavior model will decide the number of bikes sold by the company to the targeted customers..
 

I have illustrated how each model component can be analyzed in greater detail to define its inputs and outputs , and how the output of one component become the Inputs to other components. The next task is to measure the functional relationships between various key elements. For instance, it is obvious that the retail price and advertising affect the rate of consumer purchase, the real task is to measure how much.

The role of high performance computing and simulation would become clear now. For example, the marketing manager, would like to know in real time ,what should be the price of a bike sold by mail ordering channel that will ultimately maximize the profits of the company. Increasing or reducing the price may not reduce or increase sales in same proportion at all times. In many occasions, it is possible to offset higher prices with more expenditure on promotion and advertising. The incremental revenue from higher prices can be higher than the incremental cost of promotion and advertising. Thus , if a decision maker is aware of the complex functional relationship between pricing and advertising expenditure, profit maximizing actions can be undertaken.

To derive the Functional relationships between the various variables, One way is to study large amount of historical data. Computations can be done for testing Hypothesis on large sample data . The power of High performance computing is relevant out here. It is a statistical truth that more the sample size, better the estimation of the population variables. The functional relationship deduced at any point of time after say, regression analysis , may be valid only for a set of fixed states of the system variables. For example, the relationship between the pricing and advertising for identical profit levels will be subject to constancy of variables like Disposable income of the people, import quotas(Competition from foreign manufacturers), and Gasoline prices( Competition from other means of transport). For the decision maker to take decision under alternative system scenarios (like reduced import quotas, Higher gasoline prices, or Less Disposable income), Modified functional relationship has to be computed. This is nothing but the simulation of the marketing system under changing input variables. From the discussion above it is obvious that even a rudimentary marketing model will deal with thousands of variables and vast amounts of accumulated data , that a realistic simulation can be carried out only with the help of a high performance computer. Parallel computing is one development which will enable the marketer to take real time decisions based on changes in the system variables.