The modern industrial firm increasingly relies on software to support its competitive position. However, the uncertain and dynamic nature of today's global marketplace dictates that this software be continually evolved and adapted, to meet new business challenges. This ability – to rapidly update, improve, remove, replace, and reimagine the software applications that underpin a firm's competitive position – is at the heart of what has been called IT agility. Unfortunately, we have little understanding of the antecedents of IT agility, specifically with respect to the choices that a firm makes when designing its portfolio of software applications.In this paper, we explore the relationship between software portfolio architecture and IT agility. In particular, we use modular systems theory to examine how different types of coupling impact the ability to maintain, retire and commission new software applications. We test our hypotheses with a unique longitudinal dataset from a large financial services firm. Our sample comprises information on over 2,000 software applications observed over a 4-year period.We find that applications with higher levels of coupling cost more to maintain, are less likely to be retired, and are less likely to be commissioned. However, we show specific types of coupling present greater challenges than others, in terms of their impact. In particular, applications that are cyclically coupled (i.e., mutually interdependent) are the most difficult to manage, in terms of maintaining and updating the software portfolio. Our results suggest that IT managers have a critical design role to play, in firms that seek enhanced digital agility