Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Oh the Prices You'll See: Designing a Fair Exchange System to Mitigate Personalized Pricing

Aditya Karan, Naina Balepur, Hari Sundaram

TL;DR

The paper tackles the fairness of personalized pricing by proposing a fairness-centered exchange system that leverages private price dispersion to enable trading among consumers. It models centralized and decentralized price-setting within a matching framework and evaluates four fairness targets at individual and group levels. The results show that decentralized negotiation with a mean-net-cost objective yields the strongest fairness improvements, particularly when price dispersion is high, and that the system can be financially viable under certain dispersion and participation conditions. This approach offers a consumer-driven mechanism to counteract unfair personalization without seller cooperation, while also promoting seller accountability.

Abstract

Many online marketplaces personalize prices based on consumer attributes. Since these prices are private, consumers may be unaware that they have spent more on a good than the lowest possible price, and cannot easily take action to pay less. In this paper, we introduce a fairness-centered exchange system that takes advantage of personalized pricing, while still allowing consumers to individually benefit. Our system produces a matching of consumers to promote trading; the lower-paying consumer buys the good for the higher-paying consumer for some fee. We explore various modeling choices and fairness targets to determine which schema will leave consumers best off, while also earning revenue for the system itself. We show that when consumers individually negotiate the transaction price, and our fairness objective is to minimize mean net cost, we are able to achieve the most fair outcomes. Conversely, when transaction prices are centrally set, consumers are often unwilling to transact. When price dispersion (or range) is high, the system can reduce the mean net cost to each individual by $66\%$, or the mean net cost to a group by $69\%$. We find that a high dispersion of original prices is necessary for our system to be viable. Higher dispersion can actually lead to decreased net price paid by consumers, and act as a check against extreme personalization, increasing seller accountability. Our results provide theoretical evidence that such a system could improve fairness for consumers while sustaining itself financially.

Oh the Prices You'll See: Designing a Fair Exchange System to Mitigate Personalized Pricing

TL;DR

The paper tackles the fairness of personalized pricing by proposing a fairness-centered exchange system that leverages private price dispersion to enable trading among consumers. It models centralized and decentralized price-setting within a matching framework and evaluates four fairness targets at individual and group levels. The results show that decentralized negotiation with a mean-net-cost objective yields the strongest fairness improvements, particularly when price dispersion is high, and that the system can be financially viable under certain dispersion and participation conditions. This approach offers a consumer-driven mechanism to counteract unfair personalization without seller cooperation, while also promoting seller accountability.

Abstract

Many online marketplaces personalize prices based on consumer attributes. Since these prices are private, consumers may be unaware that they have spent more on a good than the lowest possible price, and cannot easily take action to pay less. In this paper, we introduce a fairness-centered exchange system that takes advantage of personalized pricing, while still allowing consumers to individually benefit. Our system produces a matching of consumers to promote trading; the lower-paying consumer buys the good for the higher-paying consumer for some fee. We explore various modeling choices and fairness targets to determine which schema will leave consumers best off, while also earning revenue for the system itself. We show that when consumers individually negotiate the transaction price, and our fairness objective is to minimize mean net cost, we are able to achieve the most fair outcomes. Conversely, when transaction prices are centrally set, consumers are often unwilling to transact. When price dispersion (or range) is high, the system can reduce the mean net cost to each individual by , or the mean net cost to a group by . We find that a high dispersion of original prices is necessary for our system to be viable. Higher dispersion can actually lead to decreased net price paid by consumers, and act as a check against extreme personalization, increasing seller accountability. Our results provide theoretical evidence that such a system could improve fairness for consumers while sustaining itself financially.
Paper Structure (23 sections, 1 theorem, 4 equations, 6 figures, 5 tables, 1 algorithm)

This paper contains 23 sections, 1 theorem, 4 equations, 6 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Statement
  4. Fairness
  5. The Model
  6. Model overview
  7. Consumer attributes
  8. Pricing algorithm $\mathcal{A}$ $(getPrice)$
  9. Exchange system $\mathcal{S}$ $(solveMatching \textmd{ and } computeM)$
  10. Theoretical claims
  11. Results
  12. RQ 1: Centralized versus decentralized $m$-setting
  13. RQ 2: Feasibility of our fairness measures
  14. RQ 3: Revenue to the exchange system $\mathcal{S}$
  15. Discussion
  16. ...and 8 more sections

Key Result

theorem 1

For $\epsilon_u = 0$$\forall u \in V$, the mean net cost over all consumers after trading is bounded below by $p_{min} (\frac{1 - \frac{|B|\gamma}{N}}{1 - \gamma})$, where $B$ is the set of consumers that will never engage as a buyer in the system, and $p_{min}$ is the minimum price assigned to any

Figures (6)

  • Figure 1: The relationship between the market (white square), the system (blue square), and consumers (gray circles). Consumers enter the market and are offered prices (a). They decide to participate in the exchange system (b). The system assigns a matching, and prices for these transactions are assigned or decided upon (c). The arrow direction represents the transfer of money. Of these matched transactions, only those that are mutually beneficial to both agents occur (d). This is determined by each agent's utility function.
  • Figure 2: Realization of four fairness definitions (a-d) under eight different optimizations (eight curves representing four fairness targets, two price-setting procedures). We set $N= 100, \gamma = 0.4$ and use pricing algorithm $\mathcal{A}_{0.95}$. Centralized methods are blue; decentralized orange. Variance bands show one standard deviation.$\mu_{I}$ and $\mu_{\mathcal{G}}$ can be effectively lowered by decentralized methods targeting their respective metrics. $\sigma_{I}$ and $\sigma_{\mathcal{G}}$ are less feasible for all metrics; in particular $\sigma_{I}$ for decentralized methods increases with $k$.
  • Figure 3: System revenue for four methodologies and intermediaries' profits for $\mu_{\mathcal{I}}^D$. System revenue changes with respect to $\gamma$ under high dispersion $\delta = 0.95$ and $k = 16$. The system at most earns $\approx \$2200$ in revenue, representing a large cut to the original seller's revenue without trading. For sufficiently high $\gamma$, system revenue falls as consumers find fees too high. Intermediary profits decline as a function of $\gamma$.
  • Figure 4: Impact of price dispersion ($\delta$) on outcomes ($\gamma = 0.4$). As $k$ increases, more consumers can access the lower price (a). Note that $\mu_I$ without trading is $\approx \$50$. If dispersion is too low for a given $\gamma$, trading will decrease (b). Assuming a financially viable $\gamma$, system revenue increases with $N$, and $\delta = 0.75$ maximizes revenue. At $\delta > 0.75$ nearly all consumers have incentive to trade; increasing $\delta$ lowers the Nash bargaining solution and thus system revenue. Increasing price dispersion lowers mean cost paid ($\mu_I$) and increases system revenue to a point.
  • Figure 5: Exchange system on an empirical pricing distribution. In (a) we see again that decentralized methods are able to reduce the gap between the best possible price and the prices paid by $62\%$. In (b) the cut taken by the system $\gamma$ must be low; at $\gamma = 0.01$ the system still earns $\$100$.
  • ...and 1 more figures

Theorems & Definitions (6)

  • theorem 1
  • proof
  • Claim 1
  • proof
  • Claim 2
  • proof