Informatics and robotics are the workhorses of a technological revolution in drug research. On them are based combinatorial chemistry, which yields compounds by the many thousands, and high-throughput bioassays, which screen them for activity. The results are avalanches of 'hits', which invade the databases like swarms of locusts. But far from being a plague, these innumerable compounds become a blessing if properly screened for 'drugability', i.e., for 'drug-like' properties such as good pharmacokinetic (PK) behavior. Pharmacokinetic profiling of bioactive compounds has, thus, become a sine qua non condition for cherry-picking the most promising hits. Just as important, but less visible, are the structure-property and structure-ADME relations, which emerge from PK profiling and provide useful feedback when designing new synthetic series. As a result, the screening, design, and optimization of pharmacokinetic properties has become the bottleneck and a major challenge in drug research. To shorten the time-consuming development and high rate of attrition of active compounds ultimately doomed by hidden pharmacokinetic defects, powerful biological, physicochemical, and computational approaches are being developed, whose objectives are to increase the clinical relevance of drug design and to eliminate as soon as possible compounds with unfavorable physicochemical properties and pharmacokinetic profiles.