Abstract:The sorting and filtering capabilities offered by modern e-commerce platforms significantly impact customers' purchase decisions, as well as the resulting prices set by competing sellers on these platforms. Motivated by this practical reality, we study price competition under a flexible choice model: Consider-then-Choose with Lexicographic Choice (CLC). In this model, a customer first forms a consideration set of sellers based on (i) her willingness-to-pay and (ii) an arbitrary set of criteria on items' non-price attributes; she then chooses the highest-ranked item according to a lexicographic ranking in which items with better performance on more important attributes are ranked higher. We provide a structural characterization of equilibria in the resulting game of price competition, and derive an economically interpretable condition, which we call gradient dominance, under which equilibria can be computed efficiently. For this subclass of CLC models, we prove that distributed gradient-based pricing dynamics converge to the set of equilibria. Extensive numerical experiments show robustness of our theoretical findings when gradient dominance does not hold.
What problem does this paper attempt to address?
This paper attempts to solve the problem of how sellers set prices according to customers' purchasing behaviors in modern e - commerce platforms. Specifically, the article explores the price competition among sellers and its equilibrium results under the "Consider - Then - Choose" (CLC) model formed after customers use the filtering and sorting functions of the platform.
### Research Background and Problems
1. **Changes in Customer Choice Behaviors**:
- E - commerce platforms offer a variety of goods and services, resulting in a burden of choice for customers.
- Customers rely on the platform's filtering and sorting functions to narrow down their choices and often use "fast and frugal" decision - making heuristics for their final selection.
2. **Research Objectives**:
- Predict the market equilibrium price results under the CLC model, helping platform operators design more effective search functions, monitor abnormal situations, and guide sellers to reach a welfare - optimal equilibrium.
3. **Research Challenges**:
- Due to product diversity, customer preference diversity, and the complexity of platform design, it becomes particularly difficult to analyze the results of price competition.
- Even in a simple two - player game, calculating the equilibrium is a daunting task (Daskalakis et al., 2009).
### Main Contributions
1. **Pseudo - competitive Property**:
- The article discovers a unified pseudo - competitive property, that is, any seller's revenue remains unchanged when its competitor's price is strictly lower than its own price.
- This property makes the equilibrium result have a sequential - move structure even in a simultaneous - move game.
2. **Existence and Computation of Local Nash Equilibrium**:
- Under the condition of satisfying the gradient dominance, the existence of the local Nash equilibrium (LNE) is guaranteed, and the time complexity of equilibrium computation is polynomial.
- The gradient dominance condition applies to cases where some sellers have a competitive advantage in non - price attributes such as quality.
3. **Price Dynamic Convergence**:
- Theoretically, when sellers use simple local price update strategies (such as the distributed gradient ascent algorithm), under mild regularity conditions, price dynamics will converge to the local Nash equilibrium.
- Numerical experiments show that even when the gradient dominance condition does not hold, price dynamics still often converge to the LNE.
### Conclusions
By introducing the CLC model, this paper successfully captures various actual purchasing behaviors on e - commerce platforms and provides feasible methods to analyze the equilibrium results of price competition. In addition, the research also shows how the natural distribution mechanism promotes the emergence of these equilibrium results, thus providing valuable theoretical basis and practical guidance for platform operators.
### Related Formulas
- **Pseudo - competitive Property**: Let \( p_i \) be the price of seller \( i \), and \( R_i(p_{-i}) \) be the revenue of seller \( i \) given the prices \( p_{-i} \) of other sellers. Then the pseudo - competitive property can be expressed as:
\[
R_i(p_{-i}, p_j < p_i) = R_i(p_{-i}, p_j' < p_i) \quad \text{for any } p_j < p_i \text{ and } p_j' < p_i
\]
- **Gradient Dominance Condition**: Let \( \nabla R_i(p) \) be the gradient of seller \( i \)'s revenue with respect to price. Then the gradient dominance condition can be expressed as:
\[
\nabla R_i(p) \cdot (p_i - p_j) \geq \alpha \| p_i - p_j \|^2 \quad \text{for some } \alpha > 0
\]
These formulas help to understand the relationship between sellers' revenue changes and price adjustments, thus supporting the analysis of price competition equilibrium.