In order to investigate the thermo mechanical fatigue in SAC-Bi solder balls, a multi-scale correlative workflow is developed, starting at 3D X-ray tomography, continuing with cross-sectional FE-SEM analyses, target preparation of grain boundaries (GBs), studying those GBs with FE-SEM transmission techniques and finally measuring the elemental GB composition with atom probe tomography (APT). Moreover, the experimental results are correlated with finite element method (FEM), density functional theory (DFT) and phase field method (PFM) calculations. The developed experimental workflow is set up to correlatively investigate GB segregation: Starting at the millimetre scale, non-destructively and statistically analysing fatigue in the entire ball grid array utilising X-ray tomography, continuing with cross-sectional microstructure analyses on the micrometre scale with electron microscopy techniques and finally studying the elemental composition of single GBs on near-atomic scale using APT. The experimental results are subsequently verified with various modelling techniques: FEM is applied for simulations of the stress distribution in the solder balls on the millimetre scale, PFM for the microstructural simulation on the micrometre scale and DFT for the modelling of elemental segregation to GBs on the atomic scale. In summary, the developed workflow bridges the gap between big data, statistical analysis utilising machine learning and the underlying micro-  and nano-structural mechanisms during thermo-mechanical fatigue of solder balls. The insights gained from this study can be used to engineer stress distributions in the solder and the elemental composition of recrystallised GBs in order to prolong the fatigue life of solder balls in microelectronic devices.