W5 Conjoint

Conjoint has a language that some may find to be quite unique. However, a basic understanding of common terminology is integral when learning and discussing conjoint. Here we include a basic list of useful terminology for those starting their conjoint education.

Attribute. A fundamental product or service characteristic such as brand, color, price, speed, etc. Some refer to attributes as factors or features. Each attribute included in conjoint analysis must have at least two levels. For example, the attribute color may be described using levels such as red, green, and blue. Each attribute should be as unique in meaning and independent from others as possible. While conjoint analysis cannot include all attributes that influence consumer preference, often most of the decision process can be modeled using a reasonably small number of attributes.

Conjoint analysis. A quantitative marketing research technique that asks respondents to rank, rate, or choose among multiple products or services, where each product is described using multiple characteristics, called attribute levels.

Full profile. Also known as multiple factor evaluation or profile method. Describes a conjoint analysis approach in which a product concept is fully defined using one level from each of the attributes in the study. Also see partial profile.

Hierarchical Bayes (HB) estimation. A computationally intensive method that may be used for estimating part-worth utilities for conjoint experiments. HB estimation allows researchers to estimate overall part-worth utilities averages and variances for the sample population and consider how well those part-worth utilities fit each respondent's choices or ratings. By reliably estimating individual part-worth utilities, W5 is able to identify valid segments that may exists within the population.

Importance. The maximum impact an attribute can exert upon product choice. Whereas, part-worth utilities provide consumer preference for different attribute levels (e.g., red, blue, green), importance gauges the impact of the attribute overall (e.g., how important is color overall).

Interaction effect. Describes situations in which the levels of two attributes combine to create something considerably better or worse that their independent values might suggest. For example, if we are studying cars, the combination of convertible with the color red may produce a synergistic effect upon utility that is not explainable by the preferences for the separate values of models and colors. Also, if limousine is combined with the color red, that combination may be considerably worse than might be expected from the separate utility scores for red and limousine.

Main effect. The independent preference or utility for the attribute levels, holding all other attributes constant. Main effects ignore the possibility of interactions between attributes. If interactions exist (and are not accounted for), main effect estimates are biased.

Market simulator. Using the part-worth utilities estimated from conjoint analysis experiments, W5 can build what-if simulators to predict how the market would choose among a set of competing product alternatives.

Part-worth. The utility associated with a particular level of an attribute in a multi-attribute conjoint analysis model. The total utility for the product is made up of the part-worths of its separate attributes.

Partial profile. A partial profile involves the presentation of a subset of the attributes in a product concept.

Share of preference. The respondent interest captured by product alternatives in a market simulation, expressed as percentages summing to 100 percent across competing product alternatives. The term "share of preference" is preferred to "market share" because the market simulator estimates which product preference all other attributes (e.g., distribution, advertising, availability) being equal.

Source: Getting Started with Conjoint Analysis, Bryan Orme, Sawtooth Software.