When dealing with 2D materials and monolayer families like : transition metal trihalides, TMDCs,...etc. In which cases SOC should be included in our calculations ?
Spin-orbit coupling scales as $Z^4/n^3$ where $Z$ is the atomic number and $n$ is the principal quantum number. When you move down the periodic table and the nucleus becomes heavier and heavier, relativistic effects become more important.
From a practical approach, when should you start incorporating SOC in your DFT calculations? From my experience, that cut-off has been Sulphur (Z=16). It's a good idea to perform a calculation with SOC turned on and investigate the results - if the difference is minimal, you can safely exclude SOC in your calculations. For me, this threshold is similar to the one I adopt for self-consistent calculations when testing different plane-wave cutoffs - Does the total energy of the system change by less than 10 meV/cell between non-SOC and SOC included cases? If so, I would safely ignore SOC. But I must provide a disclaimer here: I'm talking strictly in the context of electronic structure calculations - there could be other properties you could be interested in or exotic behavior such as in the case of Topological insulators.