The present disclosure is directed to apparatuses, systems, and methods for predicting item characteristic popularity—i.e., identifying item characteristics (e.g., item brands, item types, etc.) that are to eventually become popular. Something that is to eventually become popular is referred to herein as “pre-trend” or “cool.” In the embodiments described herein, electronic marketplace transaction data is analyzed to identify popular characteristics of items involved in recent transactions. The electronic marketplace transaction data is further analyzed to identify one or more users that executed transactions for items having these popular characteristics during a previous time period. These users' transaction histories are analyzed to determine what other item characteristics are prevalent in their more recent transactions, as these item characteristics can be identified as pre-trend/cool.