24. Choice-Based Conjoint Analysis

OBJECTIVE

Identify customers and potential customers’ preferences for specific attributes of a product. It can also be used to define the willingness to pay and market share of different products.

DESCRIPTION

This method is preferred to conjoint analysis because it represents a more realistic purchase situation and, in the case of having a large number of possible combinations, because it is sufficient to show only a certain number of combinations to each respondent. Then the responses are analyzed together and the utility is defined at the aggregated level (not at the individual level as in conjoint analysis).

For this method it is also very important to choose carefully the attributes (as a rule of thumb, no more than seven including the price) and the product profiles to present, that is, the combinations of attributes. Once the attributes and product profiles have been defined, choice scenarios are designed. A scenario is a combination of several products that is presented to the respondents. When defining scenarios, these recommendations should be followed:

• A “none” choice should be included among the products presented in each purchase scenario;
• Each scenario should not have more than 5 products;
• Between 12 and 18 scenarios are usually presented to each respondent.

Usually, all the combinations cannot be presented in the same scenario, and a good practice is to show from two to five products in each scenario. When choosing the combination of products for each scenario, it is important that all the products are shown an equal number of times and that each product is compared equally with other alternatives.

Once the data have been collected, utilities are estimated at the aggregate level. The market share of each product can be calculated using the “share of preferences”:

• Products’ utilities are calculated by summing all the attributes’ utilities;
• Products’ utilities are exponentiated;
• The market share is calculated as the product’s exponentiated utility divided by the sum of all the exponentiated utilities.

To obtain utilities at the individual level, a method called “hierarchical Bayes” is used. This method enables us to calculate a more reliable market share based on the choice of each respondent using three main techniques:

• First choice: each respondent chooses the product that maximizes her utility (this technique is suggested for expensive products that imply a careful evaluation, such as houses and cars);
• Share of preference: each respondent purchases a share of each product based on the share of utilities (suggested when a product is purchased several times during a certain period);
• Randomized first choice: each respondent chooses one product with a probability proportional to its utility.

This method is also useful for predicting variations in the market share compared with competitors by creating simulations in which prices or other products’ attributes are changed. For example, we can analyze whether a discount can attract a big enough market share to compensate for the reduction in price. In this kind of simulation, we assume that competitors are not modifying both attributes and price, but in reality this could not be the case. This is why we should at least simulate several scenarios including possible competitors’ reactions. A more complex approach would be to include a game theory model (see 76. GAME THEORYMODELS).

TEMPLATE

Here you can find a template in Excel which can help you both in the design of the survey (definition of combinations and surveys including optimal reduction of options) and in the analysis using conditional multinomial logit.

---

20. GABOR–GRANGER PRICING METHOD

OBJECTIVE

Define the optimal price range for a product or service.

DESCRIPTION

This method is useful in taking general pricing decisions. Data are collected through surveys in which each respondent is asked about his intention to purchase and shown several prices that move up or down depending on the previous answers. Alternatively, prices can be shown randomly or in a fixed series. The highest price at which a respondent reports that he would buy is considered to be his WTP. Once we have a specific price limit (WTP) for each interviewee, we can draw an accumulated demand curve.

Gabor–Granger Pricing Method

Since we have the information about demand and WTP available, we can calculate the revenue curve in the graph and establish the optimal price at which revenues are maximized.


Donwload the Gabor-Granger Excel Template

22. MONADIC PRICE TESTING

OBJECTIVE

Analyze people’s purchase intention at different price points and for alternative products.

DESCRIPTION

In monadic price testing, purchase behavior is tested for several price points, but each respondent is shown just a single price. Due to this method, a large base of respondents is necessary. A variation that needs a smaller sample is sequential monadic testing, in which the respondents are shown different price points, one at a time (usually no more than three price points are presented to each respondent). It is important to bear in mind that sequential monadic testing implies some biases and usually shows a higher purchase intention at the lower prices than monadic testing.

This is probably the best method for analyzing purchase behavior at a given price; however, it is only useful if we have an idea of the appropriate price points for a particular market. If this is not the case, we would need to obtain this information prior to the analysis, either through direct or indirect survey methods (see 18.INTRODUCTION).

Figure 20: Demand Curve Derived from Monadic Price Testing

Once the data have been collected, we can summarize the purchase behavior for the different price points (e.g. 11% of the market would purchase the product at €30, €32% at 20, etc.), and we can estimate a demand curve. The data are usually collected through surveys but can also be obtained from controlled experiments.

TEMPLATE

Download the Monadic Price Testing Excel Template

21. VAN WESTENDORP PRICE SENSITIVITY METER

OBJECTIVE

Determine consumer price preferences.

DESCRIPTION

People are asked to define prices for a product at four levels: too cheap, cheap, expensive, and too expensive. The questions usually asked are:

• -     At what price would you consider the product to be so expensive that you would not buy it? (Too expensive)
• -    At what price would you consider the product to be so inexpensive that you would doubt its quality? (Too cheap)
• -       At what price would you consider the product to start to be expensive enough that you could start to reconsider buying it? (Expensive)
• -       At what price would you consider the product to be good value for money? (Cheap)

The results are organized by price level, with the accumulated demand for each question. The demand is usually accumulated inversely for the categories “cheap” and “too cheap” to define crossing points with the other two variables (Figure below).

Van Westendorp’s Price Sensitivity Meter

From the four intersections, we have the boundaries between which the price should be settled (lower bound and upper bound). Although the other two price points are sometimes used, I prefer to use this model to define the lead prices and upper prices for a product, while the middle prices should not be static but should change based on several factors (period of purchase, place, conditions, etc.).

With this model we can define price boundaries, but we cannot estimate the purchase likelihood or demand. For the estimation of the demand (and revenues), we ask an additional question regarding the likelihood of buying the product at a specific price with a five-point Likert scale (5 = strongly agree, 1 = strongly disagree). The price to be tested can be the average of the “cheap” price and the “expensive” price for each respondent. A more comprehensive approach would be to ask the question for both the “cheap” and the “expensive” price. Then the results must be transformed into purchase probabilities, for example strongly agree = 70%, agree = 50%, and so on. With these results we can build a cumulative demand curve and a revenue curve (Figure below). The optimal price is the one at which the revenues are maximized (be aware that this approach aims to maximize revenues and does not take into account any variable costs).

Van Westendorp’s PMS Extension with Demand and Revenue Estimation

TEMPLATE

Discount code -40%: BLOG_ANALYTICS_MODELS

23. CONJOINT ANALYSIS

OBJECTIVE

Identify customers and potential customers’ preferences for specific attributes of a product. It can also be used to define the willingness to pay and the market share of different products.

DESCRIPTION

Conjoint analysis is a surveying technique used to identify the preferences of customers or prospective customers. The respondents are shown several products with varying levels of different attributes (e.g. color, performance) and are asked to rank the products. This ranking is then used to calculate the utility of each attribute and product at the individual level. The results can be used to define the best combination of attributes and price or to simulate market share variations with competitors (if competitors’ products are presented).

First of all it is very important to spend enough time designing the analysis, starting with the selection of the most important attributes and attributes’ levels.

There are three kinds of methods:

• -  Decompositional methods: the respondents are presented with different product versions, they rank them, and then the utilities are calculated at the attribute level by decomposing the observations;
• -   Compositional methods: the respondents are asked to rate the different attributes’ levels directly;
• -      Hybrid methods: compositional methods are used in the first phase to present a limited number of product versions in the second phase (they are useful when we have a large combination of attributes and levels).

In addition to the methods described above, several kinds of adaptive conjoint analysis are used to increase the efficiency of conjoint analysis, especially when the number of attributes is large.

In conjoint analyses the price is usually included as an attribute and the price utility is calculated. However, this creates several problems:

• -      By definition the price has no utility but is used in exchange for the sum of attributes’ utilities of the product;
• -     The price ranges, number of levels, and perception of the respondents can bias the answers;
• -    The purchase intention is not included, so we do not know whether the respondent would actually buy the product at the presented price (to avoid this problem partially, the respondents are usually asked to define a limit in the ranking below which products are not purchased).

The willingness to pay is calculated as the exchange rate between price utility and attribute utility. However, to avoid the abovementioned problems, we should consider a different approach, for example dividing the analysis into two phases:

• 1-      Perform a classic conjoint analysis for non-price attributes to define utilities;
• 2-      Ask for the purchase intention of full product profiles with varying prices to define the lower and upper boundaries between which the respondent would agree to purchase the product.

With this information a linear function can be estimated in which the price is the dependent variable and the utility is the independent variable.

In the example we present a classic conjoint analysis that includes the price as an additional attribute. It includes one three-level attribute, one two-level attribute (color), and three levels of price. Full-profile products are presented to the respondents (compositional method), and they are asked to give a preference on a scale from 0 to 10 (10 being the most preferred product) instead of ranking the products.

Combinations and Stated Preferences of Conjoint Analysis

The utility of a respondent is calculated by removing one level for each attribute to perform a multiple linear regression with dummy variables. The removed variables will have a utility of “0,” while the attributes included in the regression will have the utility corresponding to the regression coefficients. After verifying the significance of each attribute (p-value < 0.05; see 38. LINEAR REGRESSION), the coefficients can be summed to build the utility equation.

The utility equation at the individual level can be used to define the most profitable combination of attributes and price. It also allows the building of scenarios in which shifts in the market share are calculated due to changes in the price or products’ attributes compared with the products offered by competitors. Especially for the market share scenarios, it is important to define the purchase intention by asking the respondents to state a “limit” beyond which they will not purchase the product.

In the template a second sheet is presented in which the price is not included as an additional attribute but the respondents are asked about it separately, either directly or by showing them different price–product combinations and asking for their purchase intention. The last example usually performs better, but if we have numerous combinations, we cannot show all of them.

Price–Utility Linear Relation

There are two main approaches when creating surveys for conjoint analysis:

• -     Classic conjoint: the respondents are shown all the combinations of attributes’ levels and are asked either to rank them or to define their preferences on a certain scale (e.g. 0 to 10). If the number of combinations is too large, we should either split the combinations and present them several times to the respondents or present only a certain percentage of all the possible combinations (randomly selected). We should also ask for a “limit,” that is, the ranking position or preference level at which the respondent would change his purchase intention.
• -     Conjoint in which the price is not an attribute: the process is the same as the classic conjoint analysis, but the price is not included as an attribute. After asking the respondents to rank or set their preferences concerning several combinations of attributes’ levels, they are asked whether they would purchase a specific combination at a specific price. Depending on the response, either the utility or the price is modified to identify the WTP. If the number of combinations is limited, each one can be tested; if the number is large, not all combinations can be tested and the WTP must be calculated for different levels of utility and can then be estimated for all the combinations.

TEMPLATE

Download the Conjoint Analysis Template