Timing of product allocation: Using probabilistic selling to enhance inventory management

Scott Fay, Jinhong Xie

Research output: Contribution to journalArticlepeer-review

40 Scopus citations


This paper examines probabilistic selling (PS) as an inventory-management mechanism, paying special attention to the impact of the timing of product assignment to buyers of probabilistic goods. In practice, sellers tend to offer probabilistic products only after major demand uncertainty has been resolved. By deferring product assignments, a firm is able to obtain more information about demand for each specific product before deciding which product to assign to consumers. However, our analysis demonstrates that PS can be an effective inventory-management mechanism even if the firm allocates products before knowing which product will be more popular and, thus, scarcer. Interestingly, we show that it can be more profitable for the firm to allocate products to consumers before, rather than after, learning the true demand for a product because, although early allocation imposes higher inventory costs (as a result of larger required inventory levels), it also enables the firm to charge higher prices. Our results also reveal that, when introducing probabilistic goods, the firm should order less inventory (relative to the case where probabilistic goods are not offered) if costs are very low but more inventory otherwise. Finally, we show that PS, as an inventory-management mechanism, can create a win-win situation, both improving profit and increasing social welfare.

Original languageEnglish (US)
Pages (from-to)474-484
Number of pages11
JournalManagement Science
Issue number2
StatePublished - Feb 1 2015


  • Inventory production
  • Marketing
  • Pricing
  • Retailing and wholesaling
  • Uncertainty

ASJC Scopus subject areas

  • Strategy and Management
  • Management Science and Operations Research


Dive into the research topics of 'Timing of product allocation: Using probabilistic selling to enhance inventory management'. Together they form a unique fingerprint.

Cite this