Drawing on heuristic-systematic model, this study develops a research model to examine how content factors (i.e., persuasive and informative cues) and source factors (i.e., brand popularity and reputation) affect consumers’ liking behavior toward advertisements in microblogs, specifically on Weibo. Brand type is regarded as a moderator in the research model, which is empirically evaluated using over 240,000 tweets across approximately 70 auto brands collected from Weibo. Manual coding and machine learning algorithms are integrated to develop a classification model that tags tweets. Results show that content factors (i.e., persuasive and informative cues) and source factors (i.e., brand popularity and reputation) have significant influence on consumers’ liking behavior toward advertisement tweets in microblogs. Source factors exert stronger effects on tweet liking than content factors. Particularly, brand popularity is more powerful in increasing the number of likes than brand reputation. In addition, we find that these relationships vary significantly depending on brand type. For functional brands, persuasive cues tend to result in more tweet likes, whereas source factors are more powerful for prestige brands in driving consumers to like their advertisement tweets. Our findings enhance the current understanding of consumers’ liking behavior on social media and provide managerial insights for brands seeking to facilitate consumer engagement.