Advertising and keyword selection strategies on Baidu: an empirical analysis

Purpose As the search engine platform, Baidu has already developed keyword advertising as one of its main business scopes, while in-feed advertising is emerging as another intelligent choice for the company. Our purpose is to validate the effectiveness of keyword and retargeted in-feed advertising on offline sales and whether the effectiveness of these two advertising strategies relies on keyword attributes work. Design/methodology/approach We utilize data from the ad campaigns of a prominent manufacturer within the machinery and equipment (hereinafter referred to as “the company”) on Baidu. To scrutinize the research hypotheses, we have employed linear regression models. Subsequently, we address potential endogeneity issues and use various techniques to ascertain the reliability of the results. Findings Empirical evidence indicates that both keyword and in-feed advertising enhance offline sales. Upon examining the moderating role of keyword attributes (specificity and length), we observe that specific keywords (price and word-of-mouth (WOM)) accelerate the boosting effect of advertising on sales; similarly, the longer the keywords, the more obvious the enhanced impact of advertising on sales. Moreover, the positive influence of specific keywords (price and WOM) on advertising effectiveness is more outstanding when the keywords are longer. Originality/value To our knowledge, no empirical investigation has yet to analyze keyword and retargeted in-feed advertising concurrently within the search engine context. Our research is the inaugural work to reveal that they serve as mutual substitutes regarding their impact on sales. Furthermore, this paper pioneers examining the moderating effects exerted by keyword attributes (specificity and length) on the effectiveness of these two ad types. The findings presented herein offer valuable insights into the harmonious coexistence and collaboration among companies, advertisers, users and search engine platforms.