Market Model and Optimal Pricing Scheme of Big Data and Internet of Things (IoT)

Dusit Niyato, Mohammad Abu Alsheikh, Ping Wang, Dong In Kim, Zhu Han

Abstract

Big data has emerged as a critical approach for utilizing large datasets to optimize complex system operations, fueled by Internet-of-Things (IoT) services that generate immense sensory data. While technical research has focused heavily on machine learning and resource management, the economic modeling of these systems has been largely overlooked. This paper investigates the big data market model and proposes an optimal pricing scheme. By studying data utility from a machine learning perspective, we introduce a formal market model and develop a pricing strategy using a Stackelberg game approach. Numerical examples demonstrate that service providers can achieve maximum profit through this proposed framework.

Introduction

While the technical side of Big Data and the Internet of Things (IoT)—such as storage, processing, and machine learning—has seen explosive growth, the economic modeling of these systems has been largely ignored. Data is a unique commodity: its value is not merely in its volume, but in its ability to improve decision-making or model accuracy. However, more data does not always mean linearly better results; most algorithms hit a point of diminishing returns. This research addresses the challenge of pricing data when its utility is non-linear and determines how stakeholders in the supply chain can maximize their respective profits.

The Big Data Market Ecosystem

The authors define a four-tier market structure that captures the flow of value from raw observation to end-user service:

  1. Sensors: The primary source of raw data generation.
  2. Data Source: Aggregates raw data and sells it at a unit price ($p_b$).
  3. Service Provider: Purchases a specific volume of data ($n$), processes it (e.g., training a machine learning model), and offers a service to consumers for a subscription fee ($p_s$).
  4. Consumers: Subscribe to the service if the perceived utility exceeds the subscription fee.
The proposed four-tier market model for Big Data and IoT services.

Quantifying Data Utility

A core contribution of this work is defining the utility of data ($u$) as a function of the data size ($n$). Based on the behavior of machine learning models, utility is modeled as an increasing but concave function, reflecting diminishing marginal returns. The authors proposed and validated two candidate functions:

  • Fraction-based: $\phi(n) = 1 - \frac{\kappa}{1+gn}$
  • Exponential-based: $\phi(n) = 1 - \mu \exp(-hn)$

Using a Human Activity Recognition (HAR) dataset—consisting of accelerometer and gyroscope samples from Samsung Galaxy S II smartphones—the researchers trained Support Vector Machines (SVM) and Logistic Regression models. The experimental results confirmed that these mathematical models accurately fit the real-world performance curves of machine learning algorithms.

Optimal Pricing via Game Theory

To determine optimal pricing, the interaction is modeled as a Stackelberg game. In this scenario, the Data Source acts as the Leader, setting the price of raw data first. The Service Provider acts as the Follower, deciding how much data to purchase and what to charge consumers to maximize their own profit based on the Leader's price.

This strategic interdependence prevents the Data Source from setting arbitrarily high prices, as they must account for the fact that higher costs will lead the Service Provider to purchase less data, potentially reducing the Data Source's total revenue.

Key Results and Profit Optimization

The study derived mathematical proofs showing that for any given raw data price, there exists a unique, globally optimal subscription fee and data demand size. Numerical simulations demonstrated a clear "sweet spot" for profitability.

3D surface plot showing the optimal profit point for a service provider based on subscription fee and data demand.

Key insights from the results include:

  • Predictable Utility: Machine learning performance curves are consistent enough to serve as the basis for economic contracts.
  • Efficiency Gains: As the model parameter $g$ (the rate at which data improves the model) increases, service providers can achieve high utility with less data, allowing the data source to potentially command a higher unit price while maintaining market equilibrium.

Conclusion

This paper provides a formal economic framework for the "Data-as-a-Service" (DaaS) ecosystem. By linking machine learning performance to economic utility, it offers a robust method for stakeholders to optimize pricing and data acquisition. Future work may expand this model to include competitive markets with multiple providers and more complex consumer behavior models beyond uniform willingness-to-pay distributions.

Summary updated

Abstract

Big data has been emerging as a new approach in utilizing large datasets to optimize complex system operations. Big data is fueled with Internet-of-Things (IoT) services that generate immense sensory data from numerous sensors and devices. While most current research focus of big data is on machine learning and resource management design, the economic modeling and analysis have been largely overlooked. This paper thus investigates the big data market model and optimal pricing scheme. We first study the utility of data from the data science perspective, i.e., using the machine learning methods. We then introduce the market model and develop an optimal pricing scheme afterward. The case study shows clearly the suitability of the proposed data utility functions. The numerical examples demonstrate that big data and IoT service provider can achieve the maximum profit through the proposed market model.
BibTeX
@inproceedings{niyato2016market,
  author = {Dusit Niyato and Mohammad Abu Alsheikh and Ping Wang and Dong In Kim and Zhu Han},
  title = {{Market Model and Optimal Pricing Scheme of Big Data and Internet of Things (IoT)}},
  booktitle = {ICC},
  year = {2016},
  url = {https://openreview.net/forum?id=rbwAjFf3wG},
}
Figure