Conjoint analysis: Analyzing optimum pricing and the value of component parts

The origin of the model

Conjoint analysis is a statistical methodology developed by Gérard Debreu, a French economist, in 1960 and further developed by R Duncan Luce, a US mathematical psychologist and statistician, John Tukey, in 1964. The initial conjoint theory only had two attributes, but it didn't take long for other statisticians to create conjoint assessments with more. The implementation of conjoint modelling can be intimidating for a marketer who is not a statistician. It is aided by special software, the most well-known of which comes from Sawtooth. Even this, however, necessitates the assistance of a skilled practitioner. As a result, instead of being a marketing tool, it is a statistician's tool.

What the model looks like, and how does it work.

Conjoint analysis is a technique for determining how much consumers appreciate different offerings. Asking people simple questions about how much they would pay for a product has long piqued marketers' interest. Simple queries don't get to the heart of what people appreciate. Marketers want to know what consumers value since it affects their messaging to pique customer interest. The most valuable aspects of the offer can be singled out to generate distinct, attractive, and defendable communication. Marketers are likewise interested in determining the best price for a product. They don't want to leave money on the table, but they also don't want to overprice their goods to the point where it isn't purchased.?

Compromises and trade-offs are inherent in all decisions. We may desire a high-quality product with numerous features, but we will have to make do with less if we cannot afford it. Because the ideal is rarely achieved, we need a method for simulating decision-making in our questions. The conjoint analysis gives a framework for determining what consumers value in various offers. To formulate relevant queries, we must break down items and services into their characteristics and benefits, which we call attributes. These characteristics can be provided at various levels, such as high/low quality, delivered in an hour/shown in a week, and so on. Respondents have presented these traits and the levels of the characteristics that make up the conjoint offers, and they are asked to choose which they would prefer.?

The foundation of conjoint analysis is a list of an offer's fundamental characteristics. Consider the case of an envelope maker who wants to determine the value associated with the colour of the envelope, the sealing method, and whether or not it has a window. There are two concepts in the example, each with three qualities and two-level pricing. Which one would you pick?

The preceding example was limited to two concepts comprised of various levels of properties. More concepts may be added in the future. Concept C, for example, might be added, which is similar to concept A but is brown. Concept D, which is identical to A but is self-sealing, might be introduced, and so on. Each would be priced differently. The creation of concepts is an essential phase in a conjoint project, and it takes time to reduce them down to those that have an impact on purchasing decisions.

There are just two concepts in the case of envelopes, each of which has three qualities and two levels of variation. There can be up to seven qualities and four or five levels of each in much conjoint research. This means that the various traits and levels can have hundreds, if not thousands, of possible combinations. The conjoint researchers employ software to reduce these numerous variants to roughly 30 bundled offers presented to the public. The cost of each request is different. Respondents are shown four or five of these offers and asked to pick which ones they want to buy and which ones they don't. In a typical survey, a responder will see five or six different screens, each with four or five options — a total of 30 or so special offers, all with varying costs. The utility value for each of the attribute levels is calculated using sophisticated software that analyses the choices made by respondents. It is possible to see which of the combinations is the most favourable and how much people value the various traits and their levels in this way.

Returning to the envelope example, where there were only two concepts to pick from, we can see that a total utility figure may be determined for envelopes A and B. Out of the two options, envelope B is preferred and has a utility value of 85 (notice that the utility value is a relative value determined from the data replies, not a number out of 100). We can see that a white envelope is worth more than a brown envelope when we look at the details. This leads us to believe that a third concept, envelope C, which was white, self-seal, and included a window, would have been the preferable alternative if it had been included (depending on the price).

Respondents were provided printed cards that outlined the ideas in the early days of conjoint. The cards were jumbled and placed in favoured and rejected piles by the respondents. Coupling analysis has been done on computers since the 1980s, and it is now almost entirely done online. The way conjoint questions are asked and analyzed has undergone numerous changes. Before reviewing the concepts in a modern conjoint study, respondents are invited to share more about their preferences. As a result, the concepts can be tailored to the various needs of responses.?

The capacity to arrive at a scientific appraisal of individuals' value is one of the most appealing aspects of conjoint analysis. However, there are several drawbacks to conjoint that should be considered.

• Sample size: It usually is necessary to interview at least 100 target respondents, ideally 200, to achieve reliable results. In some B2B markets, large sample sizes are impossible to achieve.
• Several attributes: The conjoint concept's design is crucial. When there are a lot of various features, including price, conjoint analysis becomes unreliable. Similarly, the number of variables for each characteristic should be limited to three to five levels. Otherwise, the number of possible ideas combinations will be too huge to manage.
• Respondent fatigue: Respondents in a conjoint survey are asked to look at various screens, each with a different set of offers (with different prices). With so many requests to look at, it's easy for them to get "punch drunk." Unless the qualities and factors are highly different, the offers can blend and appear highly similar. Respondents become confused and exhausted due to this, and instead of giving careful thought to the options, they pick one at random to finish the boring interview.

The model in action

A company that makes carpet tiles for offices wanted to compare its current designs to some new ones. It was decided to concentrate the poll on critical decision-makers. These fit-out companies designed and installed new office spaces: The online survey had 100 participants. All of them were checked to ensure that they specified a considerable amount of carpet flooring per year.

?Online carpet tile testing has some drawbacks. The products are not touchable or feelable by the respondents. As a result, it was vital to accept this constraint and concentrate on two areas of the design: colour and visual texture. Both were captured on film. The length of the warranty, the carpet tiles' environmental friendliness, and the need for a stain protector were all evaluated factors. It was determined that the brand would not be included as an attribute or variable.?

According to the examination of the results, the attitudes toward the various carpet tiles vary depending on the projects completed by the companies. Those working with government offices had different options than those fitting out offices in larger cities. Customers with varying budgets affected the carpet tile choices used by the fit-out contractors.

For the most part, colour was the most crucial feature for specifiers. Texture and the wear guarantee came in second and third, respectively—contractors serving the public sector valued environmental friendliness more than anything else. The company was able to select new designs that might be targeted at different audiences due to the conjoint research. They could develop pricing that allowed the contemporary designs to sit easily with the existing ones in the portfolio. The company was able to design a carpet tile variety for commercial offices and another for public offices due to the research.