Research on social media advertising has expanded rapidly in recent years, with much of the existing literature focusing on visual content such as images and videos, as well as platform algorithms that shape advertisement delivery and audience targeting. By contrast, the language used in advertisement copy has received comparatively limited scholarly attention, despite its potential influence on user behaviour and engagement. This gap in knowledge is particularly evident in non-English contexts, including Japan, where linguistic structures, cultural meanings, and communication styles may differ substantially from those examined in English-language studies. Researchers also noted that the same word or phrase may carry different meanings and emotional associations depending on the type of product being advertised, suggesting that the effectiveness of ad copy may vary across product categories.

To explore these issues, the researchers conducted a large-scale analysis of Instagram advertisements posted in Japan between July 2021 and June 2023. The study examined two major product categories commonly promoted on social media platforms: dietary supplements and cosmetics. The final dataset included 12,206 supplement advertisements and 9,486 cosmetic advertisements, providing a substantial sample for examining linguistic trends in digital advertising. By focusing on Instagram, the researchers were able to investigate language use within a highly visual and engagement-driven social media environment where short-form advertising copy plays an important role in attracting user attention and encouraging interaction.

The text from the advertisements was analysed using J-LIWC2015, a psycholinguistic dictionary specifically designed for the Japanese language. J-LIWC2015 categorises words according to psychological, emotional, and semantic characteristics, allowing researchers to quantify linguistic patterns in a systematic manner. Using this approach, the study examined how different language features were statistically associated with click-through rate (CTR), a widely used measure of online advertising performance that reflects the proportion of users who click on an advertisement after viewing it. This methodology enabled the researchers to identify not only which linguistic features were associated with higher engagement, but also how these relationships differed according to product category.

The analysis revealed several notable differences between supplement and cosmetic advertisements. In supplement advertisements, words associated with risk, discrepancies, or problem awareness were positively related to CTR. This finding suggests that supplement advertisements may be more effective when they highlight concerns, deficiencies, or unmet health needs that consumers may wish to address. In contrast, cosmetic advertisements showed a different pattern. Words related to visual perception and appearance, along with language associated with positive emotions and motion, were positively linked to higher CTR. These findings indicate that cosmetic advertising may benefit from language that evokes attractiveness, emotional appeal, and a sense of dynamism or transformation.

Importantly, some linguistic features appeared to function differently depending on the product category being advertised. Motion-related language, for example, was associated with lower CTR in supplement advertisements but higher CTR in cosmetic advertisements. This contrast highlights the importance of considering the broader advertising context when designing social media campaigns. A word or expression that performs effectively in one category may not have the same effect in another. The researchers therefore argue that advertising language should not be treated as universally effective across all products, but rather tailored to the expectations, motivations, and psychological associations linked to specific consumer markets.

Overall, the findings provide large-scale empirical evidence that the relationship between the linguistic features of advertisement copy and CTR varies across product categories in Japanese Instagram advertising. The study contributes to a growing body of research examining how language influences user engagement in digital environments, particularly in non-English contexts that remain underrepresented in social media research. In addition to its academic contribution, the study offers practical insights for marketers and advertisers seeking to improve the effectiveness of social media campaigns. By identifying category-specific linguistic patterns associated with higher engagement, the findings may help organisations develop more targeted and strategically designed advertisement copy for digital audiences.

More information: Kenjiro Inoue et al, The impact of linguistic features on CTR in Instagram ads: A study of supplement and cosmetic products, PLOS One. DOI: 10.1371/journal.pone.0338313

Journal information: PLOS One Provided by University of Tsukuba