# Decision-Based Research Budgeting - Marketing Research

If the preceding examples outline approaches not to use to budget research, then how should you do it? It will come as no surprise that the answer advocated here is to begin with the decision or decisions at is- sue. This, of course assumes that by the point at which the question of how much to budget is raised, the potential researcher will have framed the basic research issue in terms of one or more decisions to be made by the organization. If this is not the case, then it is essential that they or the managers involved go back and make this determination.

Structuring the Decision

What is it about the decision that will affect what should be spent on the specific research project—indeed, whether to spend anything at all? To answer this question, it will be useful to make a brief detour into the academic world of decision theory, which is simply a structured way of thinking about management problems. It says that all decision situations have five major characteristics:

1. The decision alternatives
2. A decision environment
3. The expected outcomes
4. The probabilities of various future scenarios coming to pass
5. A decision rule

Decision Alternatives. The first step management must take is to set out all the realistic decision alternatives. These alternatives can be very simple, such as whether to leave a price as it is or raise or lower it 6 percent, or relatively complicated, such as whether to (1) introduce product A containing raw materials 1, 2, and 3, at a price of P, an advertising budget of $M, and an advertising theme of Q, or (2) introduce product B with raw materials 1, 4, and 6, at a price of 2P, an advertising budget of$1/2M, and the same advertising theme Q. Nevertheless, the only requirement is that each alternative should be mutually exclusive of every other alternative. The alternatives needn’t be exhaustive, but they should cover the major alternatives management is considering.

A Decision Environment. The next step is to specify the major dimensions of the future environment that will affect whether the choice of alternative 1, 2, or 3 is a good or a bad idea. These could be factors outside management’s control, such as future interest rates or the proportion of a market already owning a particular electronic appliance or computer software. Alternatively, they could be factors over which management could have some influence through the lternatives it chooses, such as competitors’ prices and advertising budgets.

These environmental influences, formally called states of nature in the decision theory literature, also can be complex. For example, one future scenario facing a marketer of a new privacy software could be that interest rates will be at 6 percent, 35 percent of the market will own privacy software, competitors’ average price will be $312, and their average advertising budget will be$3.5 million. Another future could be an interest rate of 7 percent, 20 percent privacy software penetration, $350 average price, and advertising at$4.5 million. These alternative futures need not be exhaustive, but they do need to cover the major possibilities.

Expected Outcomes. For each combination of decision alternative and future environment, management must estimate an outcome. Typically, this outcome will be a dollars-and-cents estimate of the consequences of taking some action in the face of some particular environment, such as net profits before taxes. For nonprofit organizations, the outcomes may be stated in other terms, such as votes for a candidate or for a referendum or the number of individuals taking up a good health habit or visiting a museum.

Probabilities of Future Scenarios Coming to Pass. Managers are usually not neutral about the future. They have hunches or intuitions, often based on years of experience and sometimes on a recent similar situation, that some futures are more likely than others. Formal decision theory simply requires that management write down these estimates as probabilities, called prior probabilities. They must collectively cover the most likely possible futures.

A Decision Rule. This complex mix of alternatives, future environmental states, outcomes, and subjective probabilities must now be combined in order to come up with a final decision. This requires that management decide on a rule that it is going to use to choose among the alternatives. There are several rules that management could use. A weighted average rule (called the expected value rule in formal decision theory) is somewhere between the two extremes outlined It requires that the manager explicitly use the probabilities of the various future environments to weigh the outcomes under each decision alternative. The recommended course of action then becomes the one that yields the best weighted sum of all possible future outcomes, with each outcome weighted by the probability of its occurrence. This doesn’t guarantee one will be correct in any specific case. But in the long run, the weighted average rule will yield the best average payoffs. This is likely because in each decision instance, use is made of all that management knows about the decision situation. The framework also gives needed direction in deciding when to do research and how much to spend on it.

Determinants of the Cost of Uncertainty

Two factors directly affect the cost of uncertainty. The more obvious one is the stakes in the decision. the cost of uncertainty would also be ten times as great.

The other determinant is the manager’s uncertainty. But this uncertainty is not what you might think. It is not how uncertain the manager is about the state of the current market or what it will be in the future. Rather, it is the manager’s uncertainty about the better course of action to take. There are times when the manager will not need research in spite of great gaps in knowledge about the market if the organization is already quite committed to a specific program or course of action. Alternatively, research may be very important if the manager has strong feelings about the nature of the market but is quite unsure about what to do. The latter case often comes about when the manager has strong forebodings about the market but still sees a particular venture as worth a try. For these reasons, we will refer to this second determinant of the research budget as decision uncertainty.

Imperfect Research

Of course, the manager cannot acquire a perfect study. We are thus still left with the question of what the manager should do about buying a research study that will not be perfect. The first step in the example is to ask whether there is any way to get a reasonable estimate of the truth for less than $1,440. Since this is the upper boundary on research expenditure, it serves as a quick checkpoint for determining whether the study is feasible under any circumstances. Suppose the manager, after having read through the rest of this book, thinks that there are at least some techniques that might be useful for this problem at a reasonable cost, such as convenience sampling or telephone research using existing center staff. The question now is how much should be budgeted for such research given that it will be imperfect. There is a formal answer to this question using decision theory and a spreadsheet computer program with which experienced decision makers are familiar. However, most managers at this point simply use judgment and experience in the light of the stakes and degree of decision uncertainty involved to decide just how much less than the maximum (perfect study) amount should be committed. Other Factors Although dollars and cents of profit have been used as the measure of the stakes in the example, it is possible that the stakes may have some no monetary elements, such as risks to a personal career or to a company’s Wall Street image or the possibility that a government agency will question a particular action. These consequences are difficult to quantify, and a consideration of techniques to attempt to do so is beyond the scope of this book. However, it is perfectly valid, and indeed sound management practice, to augment or decrease a particular dollars-and-cents outcome to take account of no quantifiable yet important outcomes in the decision environment. For example, in the example, the manager might wish to add a value of$1,000 to the payoffs under the new brochure option to reflect an imputed value for the positive effects on staff motivation of trying any new approach.