Nutritional therapy is a cornerstone of the clinical management of chronic kidney disease (CKD). Nevertheless, randomized controlled trials often have failed to show a relevant benefit of low-protein diets in nonselected CKD populations in terms of slowing the progression of kidney disease and need for dialysis. The more the target population is selected, the less the results can be generalizable to implement in clinical practice. On the contrary, observational studies, especially if performed with patient-centered, flexible approaches, point toward an extensive implementation of dietary protein restriction in different and unselected CKD populations. The observational evidence cannot be disregarded anymore. The most recent guidelines advise implementing low-protein diets or even very-low-protein diets in all CKD patients as early as stage 3. However, the lack of data from large randomized controlled trials on unselected CKD populations as well as on specific subpopulations, such as diabetic or obese patients, which nowadays comprise the majority of CKD subjects, reduces the generalizability of the recommendations. For some patient populations, such as those encompassing very old, nephrotic, or pregnant patients, the literature is even more limited because of the lower prevalence of these conditions and diffused prejudices against reducing protein intake. This pragmatic review discusses the need for integrating information derived from randomized trials with evidence derived from observational studies to guide feasible strategies for more successful implementation of low-protein diets in the treatment of all segments of the CKD population.