Conjoint Analysis - Overview of Conjoint Analysis
Conjoint Analysis - Overview of Conjoint Analysis
Conjoint analysis is an extremely powerful and useful analysis tool. Its acceptance and level of use have been remarkably high since its appearance around 1970. One study concluded that over 1000 conjoint studies were undertaken between 1971 and 1978.4
As the previous examples indicate, a major purpose of conjoint analysis is to help select features to offer on a new or revised product or service, to help set prices, to predict the resulting level of sales or usage, or to try on a new-product concept. Conjoint analysis provides a quantitative measure of the relative importance of one attribute as opposed to another. In Chapter 9, other methods to determine attribute importance weights were introduced. The most direct was simply to ask people which attribute is important. The problem is that respondents usually indicate that all attributes are important. In selecting a car, they want good gas mileage, sport appearance, lots of room, a low price, and so forth. In conjoint analysis the respondent is asked to make trade-off judgments. Is one feature desired enough to sacrifice another? If one attribute had to be sacrificed, which one would it be? Thus, the respondent provides extremely sensitive and useful information.
Some characteristics of situations where conjoint analysis has been used productively are:
1. Where the alternative products or services have a number of attributes, each with two or more levels (e.g., automatic versus manual transmission)
2. Where most of the feasible combinations of attribute levels do not presently exist
3. Where the range of possible attribute levels can be expanded beyond those presently available
3Clark Material Handling Group—Overseas: Brazilian Product Strategy (A), HBS Case Services, 1981.
4Phillipe Cattin and Richard R. Wittink, "Commercial Use of Conjoint Analysis: A Survey," Journal of Marketing (Fall 1982).
4. Where the general direction of attribute preferences probably is known (travelers want less noise, faster travel, more comfort, and so on)
The usual problem is that preferences for various attributes may be in conflict (a large station wagon cannot get into small parking spaces) or there may not be enough resources to satisfy all the preferences (a small price tag is not compatible with certain luxury features). The question usually is to find a compromise set of attribute levels.
The input data are obtained by giving respondents descriptions of concepts, which represent the possible combinations of levels of attributes. For example, one of the credit card concepts for retailers to evaluate would be:
1. Discount rate of 6 percent
2. Payment within ten days
3. Credit authorization by telephone
4. 0.75 percent of billings to support payments for local retailer advertising
5. No rebates
Respondent retailers then evaluate each concept in terms of overall liking, intentions to buy, or rank order of preference compared to other concepts.
The computer program then assigns values or "utilities" for each level of each attribute. When these utilities are summed for each of the concepts being considered, the rank order of these total value scores should match the respondents' rank ordering of preference as closely as possible. This process can be illustrated with the utilities from the credit card study shown in Figure 19-1. The combination with the highest total utility should be the one that originally was most preferred, and the combination with the lowest total utility should have been least preferred.
Interpreting Attribute Importance
The greater the difference between the highest and the lowest valued levels of an attribute, the more important the attribute. Conversely, if all the possible levels have the same utility, the attribute is not important, for it has no influence on the overall attitude. In the credit card study, the size of the discount was clearly the most important attribute. While this is not a surprising finding it should be kept in mind that the magnitude of the difference in utilities for the two discount levels is strongly influenced by the choice of extreme levels. Had the chosen discount levels been 2.5 percent and 4 percent, the difference in utilities would have been much less. For this reason it is often desirable to use three or even four levels of a complex attribute.
Usually the measures of attribute importance obtained from a trade-off study are only a means to an end. The real pay-off comes from using the results to identify optimal combinations of levels of attributes for new products, services, or policies. To see how this is done we first need to look more closely at the way the trade-off data are collected from respondents, analyzed, and interpreted.
Conjoint analysis is an extremely powerful and useful analysis tool. Its acceptance and level of use have been remarkably high since its appearance around 1970. One study concluded that over 1000 conjoint studies were undertaken between 1971 and 1978.4
As the previous examples indicate, a major purpose of conjoint analysis is to help select features to offer on a new or revised product or service, to help set prices, to predict the resulting level of sales or usage, or to try on a new-product concept. Conjoint analysis provides a quantitative measure of the relative importance of one attribute as opposed to another. In Chapter 9, other methods to determine attribute importance weights were introduced. The most direct was simply to ask people which attribute is important. The problem is that respondents usually indicate that all attributes are important. In selecting a car, they want good gas mileage, sport appearance, lots of room, a low price, and so forth. In conjoint analysis the respondent is asked to make trade-off judgments. Is one feature desired enough to sacrifice another? If one attribute had to be sacrificed, which one would it be? Thus, the respondent provides extremely sensitive and useful information.
Some characteristics of situations where conjoint analysis has been used productively are:
1. Where the alternative products or services have a number of attributes, each with two or more levels (e.g., automatic versus manual transmission)
2. Where most of the feasible combinations of attribute levels do not presently exist
3. Where the range of possible attribute levels can be expanded beyond those presently available
3Clark Material Handling Group—Overseas: Brazilian Product Strategy (A), HBS Case Services, 1981.
4Phillipe Cattin and Richard R. Wittink, "Commercial Use of Conjoint Analysis: A Survey," Journal of Marketing (Fall 1982).
4. Where the general direction of attribute preferences probably is known (travelers want less noise, faster travel, more comfort, and so on)
The usual problem is that preferences for various attributes may be in conflict (a large station wagon cannot get into small parking spaces) or there may not be enough resources to satisfy all the preferences (a small price tag is not compatible with certain luxury features). The question usually is to find a compromise set of attribute levels.
The input data are obtained by giving respondents descriptions of concepts, which represent the possible combinations of levels of attributes. For example, one of the credit card concepts for retailers to evaluate would be:
1. Discount rate of 6 percent
2. Payment within ten days
3. Credit authorization by telephone
4. 0.75 percent of billings to support payments for local retailer advertising
5. No rebates
Respondent retailers then evaluate each concept in terms of overall liking, intentions to buy, or rank order of preference compared to other concepts.
The computer program then assigns values or "utilities" for each level of each attribute. When these utilities are summed for each of the concepts being considered, the rank order of these total value scores should match the respondents' rank ordering of preference as closely as possible. This process can be illustrated with the utilities from the credit card study shown in Figure 19-1. The combination with the highest total utility should be the one that originally was most preferred, and the combination with the lowest total utility should have been least preferred.
Interpreting Attribute Importance
The greater the difference between the highest and the lowest valued levels of an attribute, the more important the attribute. Conversely, if all the possible levels have the same utility, the attribute is not important, for it has no influence on the overall attitude. In the credit card study, the size of the discount was clearly the most important attribute. While this is not a surprising finding it should be kept in mind that the magnitude of the difference in utilities for the two discount levels is strongly influenced by the choice of extreme levels. Had the chosen discount levels been 2.5 percent and 4 percent, the difference in utilities would have been much less. For this reason it is often desirable to use three or even four levels of a complex attribute.
Usually the measures of attribute importance obtained from a trade-off study are only a means to an end. The real pay-off comes from using the results to identify optimal combinations of levels of attributes for new products, services, or policies. To see how this is done we first need to look more closely at the way the trade-off data are collected from respondents, analyzed, and interpreted.
Comments
Post a Comment