Aggregation processes in customer rating systems - Insights from an economic decision experiment

Area of Science:

• Consumer Behavior
• Information Systems
• Decision Science

Background:

• Reputation systems condense customer ratings into single scores to combat information overload.
• Platforms commonly use technical aggregation functions like the arithmetic mean.
• Customer aggregation patterns may differ from platform-employed functions.

Purpose of the Study:

• To investigate whether customers' innate rating aggregation principles align with typical platform functions.
• To identify diverse customer aggregation behaviors beyond simple averages.

Main Methods:

• A controlled economic decision experiment was conducted.
• Customer product ranking decisions were analyzed to elicit aggregation principles.
• Observed behaviors were contrasted with various reference aggregation functions.

Main Results:

• The majority of customers aggregate ratings using the arithmetic mean.
• Significant individual heterogeneity exists, with clusters showing binary (positive/negative focus) or negative-focused patterns (e.g., 1-star ratings).
• Aggregation patterns were stable across presentation formats and largely unaffected by demographics or experience, except for a link between risk attitudes and extreme rating focus.

Conclusions:

• While the arithmetic mean represents the majority, diverse aggregation behaviors necessitate consideration.
• Platforms could enhance reputation systems by offering customizable aggregation options.
• Catering to varied user preferences can improve decision-making quality and system effectiveness.