<p>This study focuses on identifying an optimised synchronisation word based on multiple practical criteria for wireless internet-of-things systems, where reliable and energy-efficient synchronisation is essential to extend battery life by reducing unnecessary receiver activations (wake-ups) and false detections. To address the limitations of purely theoretical designs, a weighted multi-criteria evaluation framework is presented specifically tailored for resource-constrained receivers (e.g., IQRF). A mathematical receiver model utilising a correlation detector operating in additive white Gaussian noise is formulated, and the effects of synchronisation-word selection on detectability and the rate of false detections are analytically determined for both fixed alignment and sliding-window search methods. The methodology is augmented by laboratory measurements conducted on a pair of parallel Texas Instruments receivers. The resulting data are compared with theoretical expectations, and both agreements and discrepancies are analysed in terms of their underlying causes and system-level implications. The results yield practical design recommendations for low-power wireless internet-of-things devices, including recommendations on bit balance, suppression of aperiodic autocorrelation sidelobes, robustness to cyclic shifts, and diminished cross-correlation with recurring traffic patterns, demonstrating that the proposed framework identifies synchronisation words that reduce false alarms by orders of magnitude compared to standard baselines.</p>