A pure 15-minute city is a dead city

X-minute city (XMC) policy aims to increase the provision of two local public goods: neighborhood-level agglomeration benefits (pull-effects, PE) and daily-life convenience (self-containment, SC). We argue, a priori, that XMC objectives may be in tension with the non-linear, hierarchical, and cumulative mechanisms that structure urban systems. To test this, we measure PE and SC for both work and non-work trips using over 3 billion of observed trips across 426 neighborhoods in Seoul, Korea. Although both indicators capture ‘attraction-type’ phenomena, we find that they are statistically distinct and governed by contrasting spatial dynamics. PE exhibits a strongly hierarchical distribution following a power-law dominated by a small number of major attractors, whereas SC for non-work trips (‘local life’) is diffusely distributed, weakly spatially autocorrelated, and largely decoupled from the city’s pull hierarchy. We therefore propose a four-type functional neighborhood classification based on work–life attraction profiles to support more granular and effective re-localization policy design. Our results show that XMC policy cannot rely on a single universal lever, and we conclude by outlining possible directions for a more empirically informed and context-sensitive XMC policy.