# Conjoint analysis

Conjoint analysis (also Conjoint Measurement, Conjoint analysis, multi-attributes compositional models, dt. translated about “group measurement”) is a statistic method, which was developed in the scientific psychology. The term designates a proceeding for the measurement of the evaluation (if necessary. fictitious) property. In addition certain characteristics of the property (stimuli) are provided with certain meaning meanings, in order to be able to derive from it a as generally as possible valid total preference judgement of the consumers over the property.

In the work by Luce and John the W. Tukey (1964) for the first time mentioned Conjoint analysis was introduced in the 70's as procedure and is today the most frequently assigned analysis method for the collection of the preferences of consumers. With the help of the Conjoint analysis is examined, in which measure individual characteristics and/or. Characteristic combinations, which distinguish a certain product, by the user to be preferred. For a car manufacturer would be for example important it to determine, which meaning “manufacturer”, “HP number” and “car color” for the purchase decision of the user have the characteristics. In the context of a Conjoint analysis from these characteristics a set of total products would be combined (e.g. a red AUDI with 170 HP). The asked one delivers now to these total concepts in each case a vote. In the context of the Conjoint of procedure it is possible to conclude from the data of the user on its preferences concerning the individual characteristics and characteristic developments. In our example it could result for example that the pro gangs orient themselves with the purchase of a new vehicle primarily at the manufacturer, whereby the manufacturer BMW is preferred. Since now everyone of an enterprise provided property as combination of product properties with certain characteristic developments be understood can, the procedure Conjoint found the far spreading, which possesses it in the market and marketing research today.

For the understanding of the Conjoint analysis the decompositional principle in such a way specified of this procedure is substantial: The evaluations of the asked ones refer first to holistic product combinations, which are divided during the evaluation and converted to those characteristics and their developments, which flowed into the evaluation with. By this proceeding the Conjoint analysis corresponds to a considerable degree to the actual evaluation process of a material purchase situation, in which the consumer is likewise confronted with holistic products. Since these products of asking view have both certain pre and certain disadvantages, it is brought to weigh the meaning of the different characteristics relatively to each other and to make themselves the actual meaning of the individual characteristics conscious.

## operational areas of the Conjoint analysis

as the three most important operational areas of the Conjoint analysis in the market study can apply the ranges product development, Preisbestimmung and market segmenting. Within the range of the product development Conjoint analyses play particularly during the introduction on the market (Launch) of new products and/or. the Relaunch already more existing and modifying products a large role. “Which characteristics of my product are it, those the buyer the maximum use donate?” for example a question could be in this connection, which could be clarified by the execution of a Conjoint analysis in the appropriate sample. Ergebnisse solcher Befragungen zu Neuprodukten können neben der Steigerung des Absatzes AUCH helfen，并且Kosten zu sparen，并且da模子相连分析u.U。 those product characteristics as for the buyer irrelevant identifies, who is connected with high costs in the production.

In the range of the Preisbestimmung Conjoint analyses are frequently used, over the database for the computation of the prospective price paragraph function for a product on a given market and/or. to supply in a competition surrounding field. With the data of the Conjoint analysis thereby a market simulation in such a way specified can be accomplished, over which for a given product that price can be calculated, which brings in the profit optimum for the manufacturer.

## two standard techniques of the Conjoint analysis

the basic form of the Conjoint analysis in the course of the years into numerous variants one transferred, which are to serve for it, of certain weaknesses of the traditional procedure to overcome. Under the Conjoint methods, which developed in the course of the years by modifications and specialization of the existing procedures, became generally accepted two procedures, with which is to be dealt here more in greater detail: the adaptive Conjoint analysis (ACA) and the Choice Based Conjoint analysis (CBC).

in contrast to the “classical” Conjoint methods represents the adaptive Conjoint analysis a procedure, which is only computer-assisted feasible. When adaptive designated this procedure, because the inputs pro gangs already during the interview by the computer are converted and used to develop the next in each case questionnaire side. The interview adapts thus to the individual preference structure of the individual user, in order to draw as meaningful an information as possible from the interviews. For the asked one the adaptive Conjoint presents itself analysis as a quite varied kind of the questioning, since a ACA consists of altogether 5 questioning phases, which he has to go through. The computer becomes acquainted with the preference structure pro gangs from phase to phase better and arranges the questionnaire sides in such a way in each case that they bring the maximum information value.

Contrary to the classical Conjoint analysis is not the ACA a filling profile method, i.e. that the rehearsing and in the course of the interview must never evaluate product combinations, which consist of ALL characteristics. Everyone of the products which can be evaluated consists rather of an only small number of characteristics, whereby in the course of the interview nevertheless information about the preference structure pro gangs concerning all characteristics results. The length of the interview rises with a higher number of characteristics, since of pro gangs more preference judgements are required. The interview lengths move however with large investigation Designs in moderate orders of magnitude. In practice ACA studies become usually with 8-15 characteristics and approx. in each case. Accomplished, theoretically the possibility exists 5 developments of taking up up to 30 characteristics into the questioning Design with.

### the choice-based Conjoint analysis

the second important development within the range of the Conjoint procedures represents the choice-based Conjoint analysis, the CBC in such a way specified. Contrary to the ACA it evaluated thus always concerns with the CBC a filling profile procedure, the rehearsing and products, which consist of all possible characteristics. A further difference to the ACA consists of the fact that the user cannot gradate its evaluation concerning the submitted products. The user gets rather per questioning side a set of products submitted, from which he can select only one (or none) as of it the Präferierte.

Since the rehearsing and must against each other weigh for example 4 products with all in each case their characteristics with such a situation, the CBC makes much higher demands against the attention of the investigation participant than a ACA. Nevertheless the information yield from this kind of the collection is relatively small, since each new questionnaire side brings the information, for which the products the asked one decided, but not, as clearly its decision failed. CBC interviews are therefore usually longer than interviews with the ACA procedure. The advantage of the CBC procedure is however therein to be seen that the Befragungssetting has very large similarity to an actual purchase decision situation for example in the department store, in which the rehearsing and mostly likewise for or none of the dargebotenen products decide must. Concerning the data quality is for the CBC to be marked that the user by those relatively are enough for interviews into a certain automatism purge can, in which it the products no more in their whole does not notice, but only due to a certain characteristic (e.g. Price) decides. On the other hand the effect can be observed with the ACA that characteristics, which are actually” unimportant to the pro gangs for “from it tendentious overrated are however tendentious underestimated important characteristics.

## an advancement: Of Christian's the cock and Markus Voeth

justified the limits Conjoint analysis (LCA ), extends as limit Conjoint designated methodology the traditional Conjoint beginnings by a further process step. In the first step a certain number of stimuli is submitted to the pro gangs, which must be evaluated in accordance with condition of the collection Design and brought into an order of rank. A stimulus is here a combination of different characteristic developments. Im zweiten Schritt werden模子individuellen Kaufbereitschaften erhoben的尼姑，并且indem模子刺激在kaufenswerte und nicht-kaufenswerte Alternativen unterteilt werden。 This happens through platzieren a limit Card (process card) behind the last still buy-worth rank. The process card can be set not exclusively between two stimuli, but rather also before first or behind the last rank. So the rehearsing and can express that it concerning none and/or. all stimuli a consumer acceptance exhibits. The process card is interpreted as use zero point. Purchase values stimuli do not take positive, buy-worth accept negative use values. This procedure makes the acceptance necessary that the pro gangs judge the use distances between ranks as same lasting. Also the groups as „buy-worth “and „not buy-worth “classified stimuli must be accepted as same scaled. In this way, contrary to the classical Conjoint analysis, absolute use values are determined, in place of bare changes of use.

## example

of a group of participants is presented a property (MP3-Player), which exhibits four characteristics (running time, capacity, equipment and warranty) in in each case two developments (highly, low).

In the following the property with different in each case developments of the four characteristics is introduced to the pro gangs. Examples of these configurations would be

• high capacity and running time, but small equipment and warranty
• high capacity and equipment, but small running time and warranty
• high warranty and equipment, but small running time and capacity

the participants are to indicate now their preference order, by arranging the differently out-arranged goods in the order, which corresponds to its preferences.

## literature

• revised one and shortened version of Daiber, Achim; Hemsing, Werner: On-line Conjoint: A proven method in the new garb. In: planning & analysis, 1/2005, S. 47-53, German specialized publishing house, Frankfurt/Main. Publication with friendly permission of the publishing house.
• Baking house, K. /Erichson, B./Plinke, W. more /Weiber, R.: Multivariate analysis methods: an application orientated introduction. Berlin 2003
• Green, Paul E. /Srinivasan, V., 1978: Conjoint analysis in Consumer Research: Issues and Outlook. In: The journal OF Consumer Research, 5, S. 103-122.
• Luce, R.D. & Tukey, J.W. (1964): Simultaneous conjoint measurement. Journal OF Mathematical Psychology, (1), 1-27.
• Melles, T. (2001): Framing effects in the Conjoint analysis. An example of problems of the characteristic definition. Aachen.
• Pond ore, Thorsten, 1999: Conjoint analysis. In: Gentleman man, Andreas/Homburg, Christian (Hrsg.): Market study: Methods, applications, practical examples. Wiesbaden: Gabler, S. 472-511.
• Boutellier, R. & peoples R. (1997): Success by innovative products. Munich Vienna: Carl Hanser publishing house
• Thaden, Christian of (2001): Conjoint analysis with many characteristics. Peter long publishing house
• Voeth, Markus (2000): Use measurement in the purchase behavior research: The hierarchical individualized limits Conjoint analysis (HILCA). Dt. Univ publishing house