Mass spectrometry (MS)-based phosphoproteomics has achieved extraordinary success in qualitative and quantitative analysis of cellular protein phosphorylation. Considering that an estimated level of phosphorylation in a cell is placed at well above 100,000 sites, there is still much room for improvement. Here, we attempt to extend the depth of phosphoproteome coverage while maintaining realistic aspirations in terms of available material, robustness, and instrument running time. We developed three strategies, where each provided a different balance between these three key parameters. The first strategy simply used enrichment by Ti(4+)-IMAC followed by reversed chromatography LC-MS (termed 1D). The second strategy incorporated an additional fractionation step through the use of HILIC (2D). Finally, a third strategy was designed employing first an SCX fractionation, followed by Ti(4+)-IMAC enrichment and additional fractionation by HILIC (3D). A preliminary evaluation was performed on the HeLa cell line. Detecting 3700 phosphopeptides in about 2 h, the 1D strategy was found to be the most sensitive but limited in comprehensivity, mainly due to issues with complexity and dynamic range. Overall, the best balance was achieved using the 2D based strategy, identifying close to 17,000 phosphopeptides with less than 1 mg of material in about 48 h. Subsequently, we confirmed the findings with the K562 cell sample. When sufficient material was available, the 3D strategy increased phosphoproteome allowing over 22,000 unique phosphopeptides to be identified. Unfortunately, the 3D strategy required more time and over 1 mg of material before it started to outperform 2D. Ultimately, combining all strategies, we were able to identify over 16,000 and nearly 24,000 unique phosphorylation sites from the cancer cell lines HeLa and K562, respectively. In summary, we demonstrate the need to carry out extensive fractionation for deep mining of the phosphoproteome and provide a guide for appropriate strategies depending on sample amount and/or analysis time.