Revenue Maximization for Communication Networks with Usage-Based Pricing

Shuqin Li, Jianwei Huang, Shuo-Yen Robert Li

TL;DR

This paper provides a mathematical framework for optimizing usage-based pricing in resource-constrained communication networks. The authors demonstrate that by strategically differentiating prices among user groups, service providers can increase revenue by up to 50%, and they offer a polynomial-time algorithm to find the "sweet spot" between simple pricing tiers and maximum profit.

What Problem Does This Paper Address?

In the world of telecommunications, pricing has shifted from "flat-fee" (unlimited data) to "usage-based" (pay-per-GB). While economists have long proposed complex nonlinear pricing schemes to squeeze every bit of revenue from consumers, these are often too complex to implement or too "unfriendly" for customers.

The authors tackle two main challenges:

  1. The Complexity-Performance Tradeoff: How can a provider design a simple pricing scheme (with just a few tiers) that still captures most of the revenue of a perfectly differentiated system?
  2. The Information Gap: How does a provider maximize revenue when they don't know exactly which user belongs to which "willingness-to-pay" group?

Methods

The researchers model the interaction between a service provider and users as a two-stage Stackelberg model. In Stage 1, the provider sets prices; in Stage 2, users decide how much resource to buy to maximize their own "surplus" (utility minus cost).

Key technical approaches include:

  • Convex Optimization: Transforming non-convex revenue problems into equivalent convex formulations to find global optima for resource allocation.
  • Clustering Algorithm: For "Partial Price Differentiation," they developed a polynomial-time algorithm ($O(I^J)$ complexity, where $I$ is the number of user groups and $J$ is the number of price tiers) that determines which groups should be bundled into the same price bracket.
  • Incentive Compatibility: Under incomplete information, they designed a quantity-based price menu. This forces users to "self-select" into the correct category based on their usage patterns.

Datasets

The paper does not use external real-world datasets. Instead, it relies on numerical simulations and stylized communication network models (e.g., 3-group and 5-group scenarios) to validate their theoretical findings. They use logarithmic utility functions, which are standard in the field for modeling "proportionally fair" resource allocation.

Key Insights and Ideas

  • The "Effective Market" Threshold: When resources are limited, the optimal strategy is often to completely exclude low-value users. The provider focuses resources on a subset of high-value users—the "effective market."
  • The Minority Rule: Price differentiation is most beneficial when high-willingness-to-pay users are a minority and resources are relatively scarce. If everyone is willing to pay a lot, or if resources are infinite, simple pricing works just fine.
  • Reversing the Volume Discount: Interestingly, to mimic perfect price differentiation under hidden information, the provider should charge a higher unit price for larger quantities. This is the opposite of a traditional "volume discount" and serves as a signal of a user's high valuation.
  • Consecutive Partitioning: They proved that in an optimal multi-tier pricing system, users should always be grouped by consecutive willingness-to-pay levels. You never want to put the highest-paying and lowest-paying users in one tier while putting middle-payers in another.

Key Results

  • Revenue Gains: Complete price differentiation can achieve a 50% revenue increase over a single-price scheme in scenarios where high-value users are rare.
  • The "Sweet Spot": In a 5-group simulation, moving from one price to just two price tiers captured nearly 15% more revenue, coming within 2.4% of the theoretical maximum possible with five distinct prices.
  • Information Parity: The authors derived a specific condition under which a provider can earn the exact same revenue without knowing user identities as they would with perfect information, provided the user groups are sufficiently different.

Limitations and Future Work

  • Monopoly Assumption: The paper assumes a monopolist provider. In a competitive market with multiple ISPs, price differentiation strategies would likely be constrained by "price wars."
  • Utility Functions: While the authors provide an appendix for general utility functions, the core results focus on logarithmic utility. Different traffic types (like real-time video vs. background downloads) might require different models.
  • Static vs. Dynamic: The model is static. Future work could explore how these pricing tiers should shift in real-time as network congestion fluctuates throughout the day.

Summary updated

Abstract

We study the optimal usage-based pricing problem in a resource-bounded network with one profit-maximizing service provider and multiple groups of surplus-maximizing users. We first analytically derive the optimal pricing mechanism that the service provider maximizes the service provider's revenue under complete network information. Then we consider the incomplete information case, and propose two incentive compatible pricing schemes that achieve different complexity and performance tradeoff. Finally, by properly combining the two pricing schemes, we can show that it is possible to maintain a very small revenue loss (e.g., 0.5% in a two-group case) without knowing detailed information of each user in the network.
BibTeX
@inproceedings{li2009revenue,
  author = {Shuqin Li and Jianwei Huang and Shuo-Yen Robert Li},
  title = {{Revenue Maximization for Communication Networks with Usage-Based Pricing}},
  booktitle = {GLOBECOM},
  year = {2009},
  url = {https://openreview.net/forum?id=tjEYZAOX2x},
}

Loading related content...

Figure