Research model

 

Introduction:

Products typically come with numerous attributes, such as price, color, style, and unique features specific to the product. So: Which product attributes are likely to appeal to consumers? Further analysis reveals that the core issue is: Among the product's key features, how important is each attribute to consumers? And given the same opportunity cost, which product features would consumers actually find acceptable? Traditional market research on this type of issue often relies only on qualitative methods or asks respondents to provide direct statements, making it difficult to conduct accurate analyses. Conjoint Analysis, however, was developed precisely to address these challenges, offering a robust approach to market analysis.

Conjoint analysis can simulate the trade-off process in which consumers sacrifice certain attributes to satisfy needs in other areas—just as they would in real life. The results of this analysis are more objective and accurate.

 

 

CBR utilizes the industry-leading conjoint analysis software suite from Sawtooth, a trusted authority in the field, enabling analyses that include various research methodologies such as Adaptive Conjoint Analysis (ACA) and Choice-Based Conjoint (CBC).

 

? CBR's joint analysis can be conducted either through traditional paper-and-pencil questionnaires or entirely on a computer interface.

 

 

 

None: I won’t purchase any of the above.

 

Through the conjoint analysis method, we can accurately determine which type of product consumers prefer most. Each product’s varying price levels and distinct brands, among other factors, play their own unique roles in influencing consumer choices.

 

 

· CBC is a consumer-choice-based analytical model where testers simply need to make choices among a series of simulated products to uncover consumers' underlying preferences.

· CBC is suitable for most research settings, with its main advantage being that it creates more realistic test scenarios that closely mirror actual consumer choices. However, a drawback of CBC is that, to ensure accurate analysis, researchers typically require larger sample sizes to achieve reliable results.

· ACA is particularly well-suited for situations where analyzing multiple attributes (6 or more) or attribute levels is required, significantly saving respondents time and reducing their burden. At the same time, ACA can deliver strong results even with smaller sample sizes.

· The ACA's estimation of price factors contains certain biases and is generally not suitable for dedicated price testing.

 

 

The Gap Model is a CBR-specific model designed to analyze consumer needs. Its core idea is to identify the key demand drivers that motivate consumers to make purchases by examining the link between current unmet needs and future buying behavior. Additionally, based on the size of the demand gap, the model can predict potential purchase rates under varying degrees of unmet need.

 

 

Case Analysis – What Factors Are Driving Consumers to Upgrade Their Current DC?

Research shows that the primary reasons driving consumers to upgrade their current DC products are dissatisfaction with pixel quality and a desire for improved zoom capabilities.

 

 

Note: To protect the client's interests, the data in this case has been adjusted and is intended solely for demonstrating potential model outputs.

 

The demand gap of under 3 megapixels isn’t enough to encourage consumers to buy digital cameras again;

Only when the gap reaches over 3 million will the likelihood of repurchasing significantly increase. Meanwhile, when the gap hits 5 million, the pixel-gap effect reaches its peak. The gap is defined as the respondents' ideal DSC configuration minus their current DSC configuration.

 

 

Note: To protect the client's interests, the data in this case has been adjusted and is intended solely for illustrative purposes regarding the model's potential output.

 

 

· Decision Tree Analysis is one of the foundational methods in modern Data Mining, enabling a comprehensive comparison of independent variables and automatically identifying those that have the greatest impact on the target outcome—thereby uncovering effective classification patterns.

· Market Segmentation: Based on the decision tree output, help customers choose a suitable market segmentation approach.

· Gain Analyses: For each segment, we analyze market potential and marketing effectiveness to help clients identify the most promising segments.

· Long-term application: Utilizing a decision tree, we can generate a discriminant rule for segmenting the market, enabling us to directly determine the likely category of unknown entities in the future—even without relying on these rules. This approach provides robust support for database marketing tailored to our customers.

 

Phase 1: Select all possible segmentation variables → Choose the target variable that will serve as the basis for decision-making

Phase 2: Select Decision Analysis Method → System Operation with Automatic Detection

Phase 3: Output the Decision Tree Diagram → Output the Gain Table

 

 

· In a study of a certain IT product, customers are faced with several possible segmentation approaches: geographic segmentation, income-based segmentation, gender segmentation, and more. So, from which perspective would segmentation help identify potential consumers?

· Based on the objective, we selected "whether planning to purchase in the next 6 months" as the target variable and categorized those who plan to buy into the Target group.

· For the independent variables, we included as many demographic factors as possible that could influence the decision to purchase: age, income, city, gender, marital status, education…

· The organized analysis results are provided below.