Although researchers have demonstrated that studying multiple examples is more effective than studying one example to facilitate learning, the principles found in the literature for designing multiple examples remain ambiguous. This paper reviews variation theory research on example design which sheds light on unclear issues regarding the effects of example variability. First, the distinction of surface/structural should be replaced by critical/uncritical in example study. Aspects and features that are critical to students' understanding should be identified and compared in example design. Second, variation as well as similarity among examples should be taken into consideration in example design. Certain patterns of variation and invariance should be adopted to systematically determine the variability of examples. Third, students with different levels of prior knowledge perceive different aspects of examples that are critical for their learning. Examples should be designed according to aspects that are critical to specific students.