Maximum Difference (MaxDiff) Scaling
A client was interested in presenting a cooking oil product as a key ingredient in table spreads and needed to conduct research to determine what table spread products consumers would like to see cooking oil as a key ingredient. During an online survey, respondents experienced eight shopping exercises and made purchase decisions based on the available products.
The responses from the choice exercise were modeled using Hierarchical Bayes estimation. The model results were used to calculate the percentage who would purchase each table spread product given each claim tested. The research findings identified the table spread products most likely to experience an increase in sales by including the cooking oil as a key ingredient.
The cooking oil was launched overseas with great success, but had only received slow acceptance domestically. The business team wanted to be able to look at presenting this product as a key ingredient in table spreads. The client was interested in collecting some research data that would allow the team to approach various spread manufacturers to suggest that by using their oil as a key ingredient, more consumers would buy their product.
The main objectives of this study were:
- To understand how the addition of a branded ingredient affects consumer purchase interest in key brands within the spread category.
- To be able to approach key spread manufacturers with the idea of how the addition of a strong, client-branded claim could increase consumer purchase interest in their product.
Research Design and Methods
Five claims, along with nine table spread products, were tested in the study. Surveys were conducted with 418 respondents using Decision Analyst’s proprietary American Consumer Opinion® online panel. The nationally representative sample was augmented to obtain readable bases for four health-focused consumer segments. There were 32 MaxDiff scenarios optimally divided into four blocks of eight scenarios each. Each scenario contained four table spread concepts, each described by a claim. Each respondent evaluated (in random order) all eight scenarios within their assigned block.
Half of the respondents viewed table spread products described by unbranded claims or no claim. The other half viewed products described by the client’s branded claims or no claim. For each scenario, respondents selected the table spread they would be most likely to buy for their households and the spread they would be least likely to buy. The responses from the MaxDiff exercise were modeled using Hierarchical Bayes estimation. The model results were used to calculate the percentage who would purchase each of the nine table spread products, given each claim tested.
Pairing a table spread with a claim (either branded or unbranded) in most cases increased consumers’ interest in purchasing the spread. The impact of the client’s cooking oil name on purchase interest varied by table spread. While pairing a table spread claim increased consumer’s purchase interest, the client decided that the impact on purchase interest was not significant enough to warrant taking the product to market.
Choice Modeling Services
For more information, please contact please contact Jerry W. Thomas, President/CEO, by emailing him at firstname.lastname@example.org, or by calling 1-800-ANALYSIS (262-5974) or 1-817-640-6166.