Observational studies play an increasingly important role in estimating causal effects of a treatment or an exposure, especially with the growing availability of routinely collected real-world data. To facilitate drawing causal inference from observational data, we introduce a conceptual framework centered around "four targets"-target estimand, target population, target trial, and target validity. We illustrate the utility of our proposed "four targets" framework with the example of buprenorphine dosing for treating opioid use disorder, explaining the rationale and process for employing the framework to guide causal thinking from observational data. The "four targets" framework is beneficial for those new to epidemiologic research, enabling them to grasp fundamental concepts and acquire the skills necessary for drawing reliable causal inferences from observational data.