Massive stars dominate the space surrounding them, as the source of ionising flux, by the feedback from their winds and from dramatic cataclysmic events such as supernova. Understanding the physics of massive star winds is critical to various fields of astrophysics. A large area of tension is the so-called weak wind problem, with measured mass-loss rates from stellar winds from late O stars being lower than hydrodynamical predictions. In particular, it is suggested that the weak wind problem affects earlier O types within the low-metallicity galaxies, such as the SMC, compared to the Milky Way. However, for low-metallicity stars, these conclusions are based on only a very small number of analyzed stars. By inspecting the archival UV spectra of the commonly cited stars, we find that the wind lines of these stars are often variable. This variability appears to be cyclic, and hence resembling the ambiguous variability observed in Galactic O stars on time scales comparable to the rotational periods. This cyclic variability is commonly explained by the presence of large scale co-rotating structures within stellar winds. In low metallicity stars, this type of variability has never yet been studied but is urgently needed. We propose a highly efficient series of long slit UV spectrograph observations of seven late O type stars in NGC 346, the most massive star cluster in the SMC. Taken over 15 orbits, these observations will be equally spaced to monitor the variability of the CIV doublet 1548.2 - 1550.8 wind lines over multiple rotation periods of the targets. This will finally resolve questions about the structure of stellar winds at low metallicity.