Product Pricing Solution for Firms Managing High-Variety Product Lines with Strong Life Cycle Effect or Diffusion Characteristics



Understanding a product’s diffusion—the process by which it is accepted by a market over time—is instrumental for effective implementation of pricing strategies. As a product is sold alongside substitutes or differentiated offerings by the same firm, diffusion modeling becomes increasingly complex. Hardcover, softcover, and electronic versions of the same book for example, all exhibit different but dependent diffusion characteristics. The same is seen in iPhones of the same model that vary only by memory storage capacity. Despite advancements in multiple-product diffusion modeling, solutions remain limited in their scalability and tractability. Overcoming these challenges will allow dynamic pricing to play a more exact and influential role in shaping product life cycles, guiding consumer focus, and maximizing a firm’s return-on-investment.


Invention Description

Research at Arizona State University has resulted in a new pricing solution based on a demand model that integrates multi-product diffusion with consumer choices. This method uses the multinominal logit (MNL) choice model to incorporate adoption decisions, product attributes, and demand interactions between products.


A series of novel mathematical transformations and an associated computer algorithm provide an unprecedented level of scalability and tractability, with very little increase in computational complexity as the number of modeled products grows. Flexibility is a primary emphasis, as the invention can optimize pricing for both simultaneous and sequential product introductions, all while synchronizing outputs with a firm’s desired price-adjustment frequency. 


Potential Applications

• Marketing analytics

• Pricing strategy

• Product development


Benefits and Advantages

• Non-restrictive – Optimizes pricing solutions for multiple products with different schedules

• Effective – Outperforms myopic pricing methods by 7-30% in numerical experiments

• Scalable – Computational burden is minimally affected by the number of modeled products


Homepage of Professor Hongmin Li

Case ID:
Last Updated:


Hongmin Li

For More Information, Contact