by

Scott M. Smith

and

William R. Swinyard

Internet Text January 1999

May Not Reproduced without the Permission of the Authors

This chapter examines two advertising models, each of which addresses a critical area of advertising decision-making. The first is conceptually similar to a model developed by John D. C. Little, and his colleague Leonard Lodish; it is called MEDIAC (Media Evaluation using Dynamic and Interactive Applications of Computers). MEDIAC is a linear programming model which helps management overcome the ordinarily enormous task of developing an optimum media schedule. The second model is ADBUDG (an Advertising Budgeting model), a decision calculus model developed by Little that provides intriguing guidance toward setting appropriate advertising budgets.

Contributions like these to advertising decision making should be welcomed, since expenditures for advertising are staggering. Annual U.S. spending for advertising is well over 100 billion dollars. But despite such huge expenditures, the potential contributions of management and behavioral science are just beginning to be tapped. It is safe to say that in no other field of business do so few people with so little knowledge spend so much money.

Advertising is an area filled with ironies and contrasts. It has the troublesome habit of posing questions whose answers serve only to revise the question, or of requiring solutions before the problem has even been articulated. This is so despite the fact that it has been a focus of intense study for at least 50 years. In harmony with the approach of other chapters, in this chapter we will first review this context in which ADBUDG and MEDIAC must function and the challenges which they might be able to overcome. Then each model will be briefly discussed.

Advertising is expected to stimulate sales, increase profits, or otherwise help attain company goals. While this is clear to everyone, no one yet knows just how advertising works to accomplish these things. Let us consider a simple proposition that advertising works directly on sales (or other company goal). That is,

This relationship could be modified to include logarithmic functions or lagged time effects for advertising, or other sophistications, but the relationship would still be a fundamental one. We would all like to believe in that kind of simplicity. For example, suppose we were shown a relationship between advertising and sales like that in Figure 9-1. We might look at their close relationship and conclude that advertising is doing just what it is supposed to -- causing sales. Or we might suspiciously wonder if the advertising levels are being set (as they often are) as a percentage of sales. Or perhaps we'll suggest that a relationship this close would be found only with direct marketing, in which the sales can only come from the direct-mail advertising and sales.

In fact, the first explanation is the least likely one for the simple reason that many variables other than advertising affect sales. That is,

This is a cumbersome set of concepts to model quantitatively, which have typically made the simpler assumption that advertising affects sales directly. But this has received a good deal of attention from behavioral scientists. Indeed, for more than 70 years advertisers have believed that the key effect of advertising is not on sales directly, but on factors that mediate or cause sales.

In the 1920s this idea was expressed as the "AIDA" model, which proposed that advertising first influenced an audience's Attention, then its Interest, then its Desire, and finally Action would follow. The AIDA model has been refined and modified in the intervening 60 years, including new forms of it which explain advertising effects when the audience is not involved with the product. But the fundamental premise of mediating effects is still accepted.

This modeling approach has been called a "hierarchy of effects"; each step is believed to be a necessary but insufficient condition for the succeeding step. Advertising was not perceived as having a direct influence on sales, but on factors that mediate sales.

It was natural for someone to suggest that advertisers should look directly to sales as a measure of advertising effectiveness. They should measure its effects on the mediating factors. This was tackled a 1961 report to the Association of National Advertisers, "Defining Advertising Goals for Measured Advertising Results" (DAGMAR). The DAGMAR report suggested a precise method for selecting and quantifying advertising goals. Most important, it proposed that advertisers should collect feedback measures to determine if their advertising met those goals.

This was not a well-received suggestion. DAGMAR was in conflict over sales-oriented measurement approaches, for it made a careful distinction between sales goals and advertising goals, but it lit a controversy in the advertising community that still smolders. Among the popular recommendations which have followed from it, however, has been the concept of the "advertising decision sequence," which provides a framework for the rest of this chapter.

As the term "advertising decision sequence" suggests, well-informed advertising decisions will usually result from a sequence of activities. Although the sequence has many variants, a popular one is illustrated in Figure 9-2.

Before any advertising money is spent, some important homework must be completed. And even after the advertising has been placed, the advertiser's work is not through. Let's look briefly at each step to consider the model-building efforts there.

The question to be answered during the situation analysis stage is, "What should we communicate?" This is a paramount data collection, assimilation, and correlation stage that has few of the boundaries desirable for model building. However, many firms are now using mainframe and personal computer database-management-systems to more efficiently serve the recurrent needs for data collected during the situation analysis stage.

In this step the parameters of the communication problem are examined, the problems faced by the advertiser are discovered, and likely approaches to solve them are identified. For example, suppose our client were "Instatune" -- a regional chain of quick car tune-up centers. Sales have been slipping recently and the client has asked for an advertising campaign to reverse the decline.

A preliminary situation analysis might reveal that the service departments of a number of automobile dealerships have added their own "fast tune-up" bays, bleeding off Instatune customers. The data suggest several possible solutions: increase the promotional budget, with no strategy change; temporarily reduce prices; add new services not now offered by the dealerships; or, a reposition the "product" offered by Instatune. We might conclude that the latter holds the most promise -- reposition the product offered by Instatune as an Engine Maintenance Agreement (EMA).

While measurement of mediating factors might make more sense than measuring sales, there is plenty of evidence to suggest that advertising has a measurable impact on sales. A large body of research has used distributed lag models of advertising effect to show the short- and long-term effects of advertising on sales.

Advertising objectives most often are defined as sales objectives, but often also include others, such as audience awareness, beliefs and understanding, attitudes, purchase intentions, etc. These objectives are best set following research which establishes the "baseline" (pre-campaign) levels.

Traditionally, copy-testing services such as Starch, Gallup, Robinson, Burke Research, and others have focused on the power of an ad to generate awareness. In recent years, more attention has been given to the generation of beliefs, however.

For example, research for Instatune may lead to any or all of the following baselines and objectives whereby the objectives of an advertising campaign are clear, and its effectiveness can be clearly measured.

                                 Baselines     Objectives 
       Awareness of Instatune        53%            70%
         Belief that Instatune ...
            does excellent work      35             40    
	    does tuneups fast        38             40
            protects your engine     30             55 
            offers an EMA             0             20
            is reasonably priced     31             35
            has friendly service     38             40
         Preference for Instatune
            over competitors         22             30
         Purchase intentions for
            Instatune                14             20
         Revenues for Instatune      $5M            $6M

Message strategy is the most creative aspect of advertising. This is the creative execution, answering the questions, "What shall we say?" and "How shall we say it?" Decisions about the type of appeals (puffery or refutational, comparative or non-comparative, etc.) are made here, as well as actual copy content and format. Because of the qualitative nature of this area, quantitative modeling has contributed little. However, the impact of management science is felt in providing input to and in evaluating advertising's creative quality and impact.

In recent years a number of widely used behavioral approaches and analytical techniques have stimulated better creative strategy by providing new ways to look at customers and their reactions to products and advertising. Among them are muliti-attribute attitude modeling, multidimensional scaling, and conjoint analysis. A number of studies have used regression techniques to make recommendations about desirable ad characteristics.

Some noteworthy work by Irwin Gross has led to recommendations about the number of ads that should be created and pre-tested. Most budgets for creating and pre-testing alternative campaigns represent three to five percent of the total advertising budget. By means of a stochastic modeling approach, Gross concluded that this was far too small a sum; that the optimum budget here is about 15 percent of the total budget. This larger investment would fund development and pre-testing of alternative creative executions, from which an advertisement having greater impact could be selected.

The quantitative and iterative nature of media planning makes it an attractive area for computer-assisted decisions. In developing the media plan we answer the question, "Where, when, and how often shall we say it?" The complexity of media scheduling can take on impossible proportions. For example, a decision to use spot television advertising in any 20 of the top 100 television shows provides 4.29 x 10²² different ways to schedule the advertising. Because of its complexity, selection of media is usually made on the simple basis of cost-per-thousand exposures.

Three main exposure criteria affecting the media decision are total exposures, frequency, and reach (or, coverage). Total exposures is simply the number of times the target segments see or hear an advertisement in a given time period. An adjustment to this that reflects the potential of the segment is "weighted total exposures." (For example, compared with those seen by the working class, Instatune advertisements seen by up-scale people may have greater impact because they have more money and inclination to use Instatune. "Frequency" refers to the average number of advertisements for the product that each target group member sees during a given period. And "reach" is defined as the total number of people exposed to at least one of the advertisements during a given time period.

The most comprehensive and detailed management science approach to analyzing media selection is micro-analytic simulation, in which audience members, or segments, are represented in probabilistic units, usually based on real-life response data. Media simulations can include factors of repeated exposure, forgetting, media type, cost discounts, duplication, and many other effects. A significant problem with this approach is that it does not develop an optimum media plan. It only evaluates submitted media schedules.

Linear programming approaches have been used in media scheduling, but most of these are constrained by some unrealistic assumptions. They typically assume that audience responses to advertising are linear, they assume that the number of media exposures is a continuous variable, they assume constant media costs (no quantity discounting), and they ignore time as a variable.

Optimization models, like MEDIAC, largely overcome these weaknesses. MEDIAC is unique in that the input data can be audience response functions. The input consists of media characteristics, market and segment characteristics, and market potential with any number of exposures up to a saturation level. Its output is a media schedule, including which media to use over time, the realized potential market response by segment. It has been frequently used by decision-makers, with reported improvements of 5-20 percent in starting schedules.

A preliminary budget must be established early in the advertising sequence to guide the choice of media and its scheduling. Revisions of both that budget and the media schedule will probably be made several times. That budget is finalized after all major strategy plans have been made - this is the final budget allocation step.

Ad budgeting is a critical area in which important modeling contributions have been made. Among them are competitive models under uncertainty, game theory models, decision theory models, stochastic models, and decision calculus models. Probably the most intuitive and easiest to parameterize are the decision calculus models, of which ADBUDG is the best known. ADBUDG uses a variety of readily available input data to provide, for each advertising budget, the brand market share, sales and profit contribution.

The impact of ad budget on advertising effects is influenced by floor and ceiling (saturation) effects. As advertising spending rises, its effects (e.g., audience awareness of the advertising) also rise to the point of diminishing returns and then until a ceiling or saturation point is reached. Spending more on advertising usually produces diminished audience effects and results in inefficiency.

Carryover effects also influence the impact of advertising spending. Consumers who have become new users because of advertising, or who have increased their usage because of it, will probably continue to do so after the advertising stops. Due to this carryover, reductions in ad spending will usually reduce audience response until a floor level of response is reached. At least in the short term, further reductions in spending will not reduce audience response below the floor level. A response function typical of this is shown in Figure 9-3.

Several budget-setting approaches are popular. The "marginal economic approach" views advertising spending as an investment, and focuses on optimizing its rate of return. Although elegant and seemingly simple, it is not a practical method since it requires audience-response data of impossible precision.

The "percentage of sales" approach leads advertisers to set sales budgets at a particular percentage of a previous period's (or of expected) sales. This is an easy and popular though illogical approach. It makes advertising a function of sales, ignores the possibility that advertising will have an effect on future sales, and discourages thinking about what the firm wants its advertising to do.

The "competitive parity" approach is similar to the percentage of sales approach except that the percentage chosen is the average industry outlay.

The "affordable" and "historical spending" approaches suggest that firms should set advertising budgets based on what they can afford, or on the basis of what they have traditionally spent. The effect in both cases is to relegate advertising to the insignificant; its budget is what is left after other "more important" expenditures have been made.

Finally, the "objective and task" method sets advertising budgets at a level sufficient to permit it to reach its objectives, as discussed in the preceding "advertising objectives" section. Advertising targeted at increasing sales by 30 percent will need a different budget than that targeted at increasing sales by 10 percent. The objective and task method is both an appealing and popular approach to setting advertising budgets.

Finally, the planned advertising must be implemented. It is created in finished form and scheduled with the planned media. Then the effectiveness of the advertising is monitored through internal data or consumer research on the variables established in the objectives-setting stage.

While much of advertising is wholly qualitative, in the sections that follow, we will discuss two quantitative models: one for assisting in media planning, and one for aiding advertising budget planning.

MEDIAC is a media optimization model for frequently purchased consumer products. It uses a number of variables to search for the media schedule that will maximize total market advertising exposure. It is designed to be a straightforward model to use, following a premise established by one of its developers (John D. C. Little) for "decision calculus" models.

MEDIAC relies on a "leaky bucket" model of how advertising works. With the level of water in a bucket representing the mental state (e.g., awareness of the advertising) of the segment toward the advertising, advertising struggles to fill it while audience forgetting continues to empty it. The concept of diminishing returns is introduced by a function that links the audience's mental state -- a mediating variable -- with purchasing decisions.

The input data for MEDIAC include

The model divides the population into segments and characterizes each segment by its sales potential and media habits. It models exposure decay (forgetting) and determines the pattern of exposure in each segment by media coverage and duplication (reaching the same segment with multiple media). Its output is a media schedule which recommends which media to use and when, and realized potential market response (exposures) by segment.

The worksheet model we call MEDIAC is an extensive modification of the model developed by Little. It uses similar input variables but, unlike the original, is not an optimization model and does not use linear programming to determine a solution. The worksheet model operationalizes the value of a total campaign as a function of the maximum exposure response given insertion costs, the number of insertions, the exposure value of an individual advertisement, and the efficiency of the exposure (measured as the probability that the audience member is a member of the targeted market segment). This is expressed as,

where

To restate, K_ijt is the exposure efficiency, or the expected number of exposures produced in market segment i by one insertion in media vehicle j during time t. And X_jt is the number of insertions in media vehicle j during time period t.

The MEDIAC problem is, then, to maximize total sales for the period through the purchase of the proper number of media insertions X_jt, for all media vehicles j during time period t. The solution of the model is subject to constraints in exposure value, use of specific media, and advertising budget.

A limitation of MEDIAC, and of the current worksheet model, is that exposures are treated in the aggregate, which could hide the fact that one person in a segment could get five exposures and another (in the same segment) gets none. But this approach has the advantage of being more efficient and more adaptable to available sources of data. MEDIAC fits best when the objective of the advertising is to maintain loyalty by "reminding" the segment of the product, rather than when introducing a new product or a new use for an old product. Used appropriately, MEDIAC is an excellent model which will both select media options and schedule them over time. The current model, which is based on the optimization of a simple objective function, is provided for illustrative purposes and may be less than ideal in actual usage situations.

The MEDIAC worksheet is divided into three sections.  We will look at each section in turn, examining it separately and in connection with the other sections, to provide an idea of how the three sections work together.

The highlighted sections of the worksheet indicate the areas where input is to be made by the user.  Inputs for the first section include the maximum budget to spend, and those variables that influence media exposure efficiency.  Media exposure efficiency is computed as the product of the four columns in the first section; this media efficiency shows the expected number of exposures produced in the market segment by one airing of the advertisement.

The second section of the worksheet reports the number of exposures purchased by the optimization model. Note that the number of exposures is not the same as the number of media inserts.  User inputs include the value of exposure to an individual consumer, the maximum sales response, the average sales per capita, the number of exposures purchased for each media vehicle option, and the media efficiency index computed in the first section of the worksheet. The simulation in the third section of the worksheet computes the optimal number of insertions.  When the optimal number of insertions, the value of exposures, and media efficiency are multiplied together, the result is the number of individual exposures that are recommended for purchase. 

The third section displays the model cost and the composite index that provides the evaluation components of the model.  Given user input for the maximum number of insertions, cost per insertion, and the marginal return constant, the model cost is computed as the product of the cost per insertion and the optimal number of insertions reported in the second section.  The composite index is analogous to a marginal return index, in that as each purchase of a media vehicle option is made, the composite index declines.  Because the return is greatest for the media vehicle option with the highest composite index, it is this media vehicle option that is chosen for the next purchase. 

Load Excel with a backup of the worksheet on the hard drive. When the blank worksheet appears on your computer screen, continue with the File Open commands and specify the MEDIAC file.

After a few moments, the MEDIAC model will be loaded into your computer, and will display its command menu:

The MEDIAC commands are invoked by selecting the desired command and then pressing OK. Once you invoke a menu function a secondary menu appears or the selected feature is initiated. You can initiate another command function by simply selecting another option from the command menu.

Each function in the command menu invokes a predefined macro sequence which performs a series of operations. The purpose of each function is described below:

Invoking "ENTER" is the typical starting point for MEDIAC. The "ENTER" command prompts you to enter new data into the worksheet model, which will replace the sample data that has already been entered. "ENTER" starts a series of prompts, beginning with your entry of an advertising media budget, as shown in Figure 9-4. After entering the budget, the macros will run for several seconds before the next step appears, which is a secondary menu:

This menu is a prompt for any changes you wish to make in data for each of several predefined media vehicles. See Figure 9-5. If you wish to change data for any magazine, invoke that magazine and then enter all data for that media option. For each vehicle you select, you will be prompted for new data.

To view screens one, two, and three, select the "VIEW" option. A secondary menu appears, permitting you to examine each of the sections of the spreadsheet, each showing either entered or calculated data for the magazines being considered in the media schedule:

1 2 3 MAIN

FIGURE 9-4
MEDIAC ENTRY SCREEN
                                                                                 
    A     B     C    D     E    F     G     H      I     J      K    L     M     
1   |                            MEDIAC TYPE MODEL      Media Efficiency:  |     
2   |INPUT MAXIMUM BUDGET:    $15,000                   K = (N)*(H)*(G)*(S)|     
3   |______________________________________________________________________|     
4   |   Media   |   N,s    |   H,j    |    G,sj    |    S,jt    |  K,sjt   |     
5   |  Option   |Population| Prob. of |% of Segment|Seasonality |  Media   |     
6   | (Vehicle) |of Segment| Exposure |In Audience |   Factor   |Efficiency|     
7   |-----------|----------|----------|------------|------------|----------|     
8   |LIFE       |   15,000 |      0.4 |        0.7 |        0.9 |    3,780 |     
9   |TIME       |   35,000 |      0.7 |        0.6 |        0.9 |   13,230 |     
10  |NEWSWEEK   |   30,000 |      0.6 |        0.4 |        0.9 |    6,480 |     
11  |TV GUIDE   |   10,000 |      0.7 |        0.8 |        0.9 |    5,040 |     
12  |BUSINESS WK|   20,000 |      0.5 |        0.5 |        0.9 |    4,500 |     
13  |AZ HIGHWAYS|   25,000 |      0.2 |        0.2 |        0.9 |      900 |     
14  |SCHOOL PAPE|    8,000 |      0.1 |        0.1 |        0.9 |       72 |     
15  |DAILY HERAL|    8,500 |      0.2 |        0.1 |        0.9 |      153 |     
16  |SPORTS ILL.|    9,500 |      0.4 |        0.1 |        0.9 |      342 |     
17  |FIELD&STREA|    9,000 |      0.3 |        0.1 |        0.9 |      243 |     
18  |-----------|----------|----------|------------|------------|----------|     
19  |        10 |  170,000 |          |            |            |   34,740 |     
20  |___________|__________|__________|____________|____________|__________|

Invoking the first page (1) shows population, readership, seasonality, and media efficiency. The second page (2) shows exposure values, minimum and maximum numbers of insertions, media efficiency, and exposures purchased. And the third page (3) shows the forecasts: sales per capita, sales response, cost per insertion, marginal return constants, and a composite index of media efficiency. See Figures 9-6, 9-7, and 9-8 for samples of these three screens.

Invoking this command runs the model using either the data which you have entered (or, if you have entered none, it uses the sample data). This is a large model, and running it will takes several minutes, depending on the speed of your PC, typically, and more if the budget is very large or insertions are highly constrained. The greater the number of insertions to be evaluated, the longer the model will take to run.

Invoking "PRINT" takes you to a secondary menu like that found in "VIEW", and permits you to print out any of the three pages ("1", "2", or "3"). Be sure your printer has paper and is on-line.

 QUIT

Invoking this command quits the worksheet and returns to Excel spreadsheet, where you can explore the worksheet and its macros. If you wish to re-invoke MEDIAC at this point, type <ALT><M>. If you made changes to the model that you wish to save, you may save them using the Excel File Save command. However, rename the file you save to avoid overwriting the original worksheet file.

FIGURE 9-5
MEDIAC VEHICLE MODIFICATION SCREEN
 
    A     B     C    D     E    F     G     H      I     J      K    L     M     
1   |                            MEDIAC TYPE MODEL      Media Efficiency:  |     
2   |INPUT MAXIMUM BUDGET:    $15,000                   K = (N)*(H)*(G)*(S)|     
3   |______________________________________________________________________|     
4   |   Media   |   N,s    |   H,j    |    G,sj    |    S,jt    |  K,sjt   |     
5   |  Option   |Population| Prob. of |% of Segment|Seasonality |  Media   |     
6   | (Vehicle) |of Segment| Exposure |In Audience |   Factor   |Efficiency|     
7   |-----------|----------|----------|------------|------------|----------|     
8   |LIFE       |   15,000 |      0.4 |        0.7 |        0.9 |    3,780 |     
9   |TIME       |   35,000 |      0.7 |        0.6 |        0.9 |   13,230 |     
10  |NEWSWEEK   |   30,000 |      0.6 |        0.4 |        0.9 |    6,480 |     
11  |TV GUIDE   |   10,000 |      0.7 |        0.8 |        0.9 |    5,040 |     
12  |BUSINESS WK|   20,000 |      0.5 |        0.5 |        0.9 |    4,500 |     
13  |AZ HIGHWAYS|   25,000 |      0.2 |        0.2 |        0.9 |      900 |     
14  |SCHOOL PAPE|    8,000 |      0.1 |        0.1 |        0.9 |       72 |     
15  |DAILY HERAL|    8,500 |      0.2 |        0.1 |        0.9 |      153 |     
16  |SPORTS ILL.|    9,500 |      0.4 |        0.1 |        0.9 |      342 |     
17  |FIELD&STREA|    9,000 |      0.3 |        0.1 |        0.9 |      243 |     
18  |-----------|----------|----------|------------|------------|----------|     
19  |        10 |  170,000 |          |            |            |   34,740 |     
20  |___________|__________|__________|____________|____________|__________|     




FIGURE 9-6
MEDIAC SCREEN 1
 
    A     B     C    D     E    F     G     H      I     J      K    L     M     
1   |                            MEDIAC TYPE MODEL      Media Efficiency:  |     
2   |INPUT MAXIMUM BUDGET:    $15,000                   K = (N)*(H)*(G)*(S)|     
3   |______________________________________________________________________|     
4   |   Media   |   N,s    |   H,j    |    G,sj    |    S,jt    |  K,sjt   |     
5   |  Option   |Population| Prob. of |% of Segment|Seasonality |  Media   |     
6   | (Vehicle) |of Segment| Exposure |In Audience |   Factor   |Efficiency|     
7   |-----------|----------|----------|------------|------------|----------|     
8   |LIFE       |   15,000 |      0.4 |        0.7 |        0.9 |    3,780 |     
9   |TIME       |   35,000 |      0.7 |        0.6 |        0.9 |   13,230 |     
10  |NEWSWEEK   |   30,000 |      0.6 |        0.4 |        0.9 |    6,480 |     
11  |TV GUIDE   |   10,000 |      0.7 |        0.8 |        0.9 |    5,040 |     
12  |BUSINESS WK|   20,000 |      0.5 |        0.5 |        0.9 |    4,500 |     
13  |AZ HIGHWAYS|   25,000 |      0.2 |        0.2 |        0.9 |      900 |     
14  |SCHOOL PAPE|    8,000 |      0.1 |        0.1 |        0.9 |       72 |     
15  |DAILY HERAL|    8,500 |      0.2 |        0.1 |        0.9 |      153 |     
16  |SPORTS ILL.|    9,500 |      0.4 |        0.1 |        0.9 |      342 |     
17  |FIELD&STREA|    9,000 |      0.3 |        0.1 |        0.9 |      243 |     
18  |-----------|----------|----------|------------|------------|----------|     
19  |        10 |  170,000 |          |            |            |   34,740 |     
20  |___________|__________|__________|____________|____________|__________|     




FIGURE 9-7
MEDIAC SCREEN 2
 
    N      O      P     Q      R    S    T    U     V    W     X     Y     Z     
1   |                                                                      |     
2   |             Exposures  Purchased   Formula:  (E,j)*(X,jt)*(K,sjt)    |     
3   |______________________________________________________________________|     
4   |    Media    |    E,j     | Maximum | Average  |  K,sjt   |Exposures  |     
5   |   Option    |  Value of  |  Sales  |  Sales/  |  Media   |Purchased  |     
6   |  (Vehicle)  | Exposures  |Response |  Capita  |Efficiency|           |     
7   |-------------|------------|---------|----------|----------|-----------|     
8   |LIFE         |        0.5 | $75,000 |    $5.00 |    3,780 |         0 |     
9   |TIME         |        0.6 |$210,000 |    $6.00 |   13,230 |    15,876 |     
10  |NEWSWEEK     |        0.5 |$210,000 |    $7.00 |    6,480 |     6,480 |     
11  |TV GUIDE     |        0.6 | $90,000 |    $9.00 |    5,040 |     3,024 |     
12  |BUSINESS WK  |        0.5 | $40,000 |    $2.00 |    4,500 |         0 |     
13  |AZ HIGHWAYS  |        0.4 |$200,000 |    $8.00 |      900 |         0 |     
14  |SCHOOL PAPER |        0.2 | $32,000 |    $4.00 |       72 |         0 |     
15  |DAILY HERALD |        0.3 | $59,500 |    $7.00 |      153 |         0 |     
16  |SPORTS ILL.  |        0.5 | $28,500 |    $3.00 |      342 |         0 |     
17  |FIELD&STREAM |        0.4 | $27,000 |    $3.00 |      243 |         0 |     
18  |-------------|------------|---------|----------|----------|-----------|     
19  |   Totals    |            |$972,000 |    $5.72 |   34,740 |    25,380 |     
20  |_____________|____________|_________|__________|__________|___________| 



    
FIGURE 9-8
MEDIAC SCREEN "3"
 
    AA   AB   AC  AD  AE   AF    AG   AH   AI    AJ    AK AL  AM    AN     A     
1   |Exposure Value    Formula:  Exposures / $Per Insertion              1 |     
2   |Maximum Budget:     $15,000                                         0 |     
3   |__________________________________________________________          0 |     
4   |  Media  |Optimal| MAX X,jt |  Cost   |Model Cost |Margin|          0 |     
5   | Option  | # of  | Maximum  |   Per   |($/Insert)*|Return| Composite  |     
6   |(Vehicle)|Inserts|Insertions|Insertion|(# Inserts)|Const.|   Index    |     
7   |---------|-------|----------|---------|-----------|------|------------|     
8   |LIFE     |     0 |       20 |  $1,000 |        $0 | 0.90 |    2679071 |     
9   |TIME     |     2 |       15 |  $5,000 |   $10,000 | 0.90 |    1028877 |     
10  |NEWSWEEK |     2 |       10 |  $1,000 |    $2,000 | 0.90 |     887319 |     
11  |TV GUIDE |     1 |        6 |  $3,000 |    $3,000 | 0.90 |     622954 |     
12  |BUSINESS |     0 |       14 |  $7,000 |        $0 | 0.90 |     289285 |     
13  |AZ HIGHWA|     0 |       30 |  $2,000 |        $0 | 0.90 |     129600 |     
14  |SCHOOL PA|     0 |        1 |    $100 |        $0 | 0.90 |        664 |     
15  |DAILY HER|     0 |        0 |    $900 |        $0 | 0.90 |          0 |     
16  |SPORTS IL|     0 |        0 |    $600 |        $0 | 0.90 |          0 |     
17  |FIELD&STR|     0 |        0 |    $400 |        $0 | 0.90 |          0 |     
18  |---------|-----------------------------------------------|------------|     
19  | Totals  |     5 |       96 |         |   $15,000 |      |            |     
20  |_________|_______|__________|_________|___________|______|____________|

TABLE 9-1
Composite Index of Non-advertising Effects

                                  Period
Index of Effect on Share     1     2     3     4   
  Promotions                1.00  1.10   .98  1.00
  Price                     1.00  1.00  1.00  1.00
  Package                   1.00  1.05  1.05  1.05
  Competitive Action        1.00   .98   .95  1.00
  Other                     1.00  1.00  1.00  1.00

  Composite Index           1.00  1.13  .978  1.05

FIGURE 9-11
ADBUDG INPUT SCREEN
 
                             A                                 B        C       
1                     ADBUDG:  ADVERTISING  BUDGETING  MODEL                     
2   =====================================================================        
3   |                       Number of periods (Max 10) =             4  |        
4   |                                                                   |        
5   |                        REFERENCE CASE CONDITIONS                  |        
6   |                     Mkt share at start of period =         0.054  |        
7   |           Adv rate to maintain share ($M/period) =             1  |        
8   |              Mkt share at period end if adv is 0 =        0.0454  |        
9   |   Mkt share at period end if incr. to saturation =         0.063  |        
10  |      Mkt share at period end if adv increase 20% =        0.0554  |        
11  |                Mkt share in long run if adv is 0 =             0  |        
12  |                           Index media efficiency =             1  |        
13  |                         Index copy effectiveness =             1  |        
14  |              Contribution profit before adv exp. =          2.25  |        
15  |                              Average brand price =           8.6  |        
16  |                     Mkt share in previous period =         0.055  |        
17  |            Product sales rate at start of period =            22  |        
18  |               Average price for product category =         $8.60  |        
19  |                                                                   |        
20  |MULTI-PERIOD BUDGET HORIZON CONDITIONS   (ENTER 1 IF YES, 0 IF NO) |