Attitude Change Models

In general, attitude change models all have three phases. First, consumers’ attitudes are measured based on the existing and similar products. Then, secondly, consumers are given the new product and, after use, attitudes are measured again. Then, using the data gathered, we can translate the attribute measure into purchasing probability and correct that probability with awareness and availability factors. The final outcome is the prediction of purchasing probability.

For collecting the preferences, there are two main models. First is the COMP and the second is Silk and Urban. The procedures are the same and the difference is the estimation methods they use to collect the preference data.

COMP - Linear Modeling Procedure
COMP uses the linear model to collect consumer preferences and states the importance of each attribute. The attribute measure of the product under test equals to: (Preference of measured product)/(Sum of the preferences for All Products)

Silk and Urban- Logit Modeling Procedure
Silk and Urban doesn’t use the linear modeling method but instead uses the logit model. Consumers’ preferences are measured from a constant sum paired comparison task in which consumers allocate a fixed number of chips among each pair of products in their consideration set. Then we can translate the result into ratio-scaled preferences.
To forecast the purchase probability, Silk and Urban define it as the multiply of unadjusted purchase probability and consideration percent. Unadjusted purchase probability is the result of preference analysis and the consideration percent, which is the percent of consumers who will consider the new product, is the estimation made by manager based on the product types. For new brand or product, the consideration percent is an estimate and, for existing brand, the consideration percent can be obtained from the market and past data.

The main advantages of this model are:
Avoid some of the laboratory effects. That the attitudes are measured in real world not in laboratory can prevent the system biases.
Intervening effects such as awareness and availability are incorporated directly in the model which makes the results more reliable and accurate.
Product characteristics or consumer perceptions are readily incorporated in the model through the preference analysis.
Easy to translate into actionable decision. Since attitude change models use the preference analysis concept, manager can easily find out which attribute to improve and know the outcome.

The disadvantages of the model are:
As the preference analysis, it’s an indirect measure. The prediction depends on the accuracy and completeness of the model used to estimate choice behavior from the measured attitude.
Attitude change models assume the product enters into an existing category to compare the prior and after preferences change to estimate the purchasing probability. So, if there is no existing category fits the product or there is no right reference product in the market, the use of attitude change models are limited and the results are doubtful.
The main limitations associated with this approached are mentioned below:

Need references
Because of the assumptions that attitude change models have, the tested product needs to have existing or similar references in the market. If not, the prior test result is valueless because consumers maybe can’t even relate the new product with the others as providing the same function or being in the same consideration set. Also, the preference result can mislead the manager to the wrong direction and the consideration percent that manager estimate can depart far from the real situation.

Costs of trials can’t be too high
Here we measure the attitude change after consumers use the product, so products for trial can’t be too high if the value of pretest can’t cover the costs.

Trial period can’t be too long
The benefits of the product should be able to be sensed by consumers in short time otherwise the timeline will be too long and lost the opportunity.
The attitude change model is appropriate for forecasting for different types of products:

Consumer package goods
Based on the limitations mentioned above, attitude change models are suitable for consumer package goods because in this industry most categories are defined and company can easily find reference products to compare with the new product. Also, the product benefits can be sensed by consumers quickly and managers don’t have to estimate the availability or awareness since the in this mature market the data is easy to collect.

Consumer durables
In this industry, managers have to be careful of using attitude models since the trial costs of some of the products in this industry are too high and the benefits of the products can’t be found in short period. But, if these issues can be solved, attitude change analysis is valuable to use. Consumers spend more time on, are more careful of, and are less impulsive of choosing durable goods. Attitude change measurements can effective tell the purchasing probability, the attributes needed to be improved, and the expected results of improvements.

Service industry
For existing service or service improvement, attitude change models are perfect fit. For innovative service, consumers may not have enough references to compare. In this case attitude change models can still provide valuable information but we need to incorporate other measurements to better estimate the purchasing probability.