Master protocol is gaining increasing attention in recent years, especially in oncology drugs and biologics. The master protocol are usually classified into three categories based on trial design, namely, umbrella, basket, and platform trials. Among these, basket designs are probably the most widely used, especially in oncology early phase trials, after the determination of RP2D. The basket design investigates a single targeted therapy in the multiple disease or disease subtypes with the goal to identify a disease indication where the drug might demonstrate early signal of efficacy. The basket design can be quite complex where a range of design elements can be implemented, such as Bayesian, or adaptive, or both. The complexity and the flexibility bring not only advantages but also many challenges in statistical design, conduct, analysis and interpretation of the results. This talk will give an overview of basket designs and also illustrates the design consideration and operational challenges with Bayesian adaptive basket design examples in early phase oncology.