The goal of the present work was to provide some guidelines for method development in hydrophilic interaction chromatography (HILIC). For this purpose, a training set of 82 representative pharmaceutical compounds possessing diverse polarity and including acidic, basic and neutral properties was analyzed. All these drugs were injected on five short HILIC columns packed with sub-2 µm particles and dedicated for UHPLC (ultra-high performance liquid chromatography) operation. Four different pH conditions ranging from pH 3 to 6 were tested at two ionic strengths (10 and 50 mmol/L) and finally, the reference organic modifier in HILIC, namely acetonitrile was modified with small amounts of methanol or isopropanol. From these experiments and using multivariate data analysis, it is clear that the stationary phase was the most relevant parameters for tuning selectivity in HILIC, since the types of interactions (i.e. dipole–dipole, hydrogen bonding and ion exchange) with analytes strongly vary between columns. Among the selected phases, the diol phase was the less interesting one, in terms of selectivity and peak shape. The zwitterionic phase was attractive, as it allowed a better retention of acidic compounds. Finally, the bare silica phase was the most versatile HILIC column packed with sub-2 µm particles in terms of retention, peak shape and selectivity. Mobile phase pH was the other important parameter to achieve an appropriate selectivity and retention, even if it remains always difficult to assess precisely the mobile phase pH, analyte pKa and silanols pKa, when working with more 70% acetonitrile. Finally, buffer ionic strength and organic modifier nature could be considered as secondary parameters for HILIC method development. In conclusion, screening four different columns packed with sub-2 µm particles at two mobile phase pH, using a fast gradient seems to be a good generic approach for initial HILIC method development. The total time for such a screening was estimated at ∼1 h, including reequilibration.