Online auctions have become a major electronic commerce channel in terms of revenue, reaching an enormous and very diverse population of participants all over the world. Modeling the factors that drive the dynamics of an auction, that is, the interactions that happen during its negotiation, is crucial to improve the customer experience for both sellers and buyers. Auction dynamics can be seen as complex and non-isolated interactions, where successive interactions become a 1oop-feedback mechanism. In what we call reactivity, the user behavior affects the auction negotiation and vice-versa. In this work, we develop a characterization approach for online auctions to capture the reactivity concept. Using the characterization, we are able to model both auction negotiation patterns and bidding behavior and correlate them. Our approach is novel, and the results start to shed light into measuring reactivity in online auctions. Our aim is to explain how the auction negotiation affects the bidders' behavior and vice-versa, and to relate the correlation auction pattern - bidding behavior to outcome measures.