Instagram Advertising Dataset

Abstract (our paper) This study analyzes linguistic features impacting click-through rate (CTR) in Japanese Instagram ads (21,692 ads; July 2021-June 2023, Meta's Marketing API). CTR was computed as link clicks/impressions from Meta's Ads Manager. Using J-LIWC2015, we quantified psycholinguistic dimensions, predominantly in Japanese. Multivariate regression models, controlling for caption length, log-transformed impressions, and product-level fixed effects, identified distinct linguistic patterns predicting CTR by product category. For supplement ads, ``risk'' $(\beta=0.108$ and ``discrepancy'' $(\beta=0.051)$ positively impacted CTR; ``motion'' $(\beta=-0.090)$ and ``negative emotion'' $(\beta=-0.076)$ decreased it. For cosmetic ads, ``see'' $(\beta=0.134)$, ``positive emotion'' $(\beta=0.088)$, and ``motion'' $(\beta=0.046)$ were positive predictors, while ``body'' $(\beta=-0.103)$ and ``negative emotion'' $(\beta=-0.053)$ decreased it. These findings underscore the critical role of linguistic features in enhancing advertising impact when aligned with the psychological needs of target audiences. By leveraging these insights, marketers can develop data-driven communication strategies to optimize engagement on Instagram. Data This dataset contains anonymized caption texts, product category labels (supplement or cosmetic), impressions, clicks, and computed CTR values used in the analyses. Brand identifiers and sensitive information have been masked to comply with privacy requirements. Publication This data set was created for our study. If you make use of this data set, please cite:Kenjiro Inoue, Mitsuo Yoshida. The impact of linguistic features on CTR in Instagram ads: A study of supplement and cosmetic products. PLOS One. vol.21, no.4, e0338313, 2026. https://doi.org/10.1371/journal.pone.0338313