How short video marketing influences purchase intention in social commerce: the role of users’ persona perception, shared values, and individual-level factors

The study examines how short video marketing influences consumers’ purchase intention in China’s rapidly expanding short video ecosystem (over 1.0 billion users by Dec 2022). Short videos, with rich multisensory cues and fast transmission, can heighten engagement, authenticity, and interactivity, potentially affecting purchasing decisions. Prior work explored brand equity, social learning, influencers, and creation capabilities; however, the role of persona perception—the perception and evaluation of personalities and attributes displayed by influencers—has been overlooked. This paper seeks to understand whether consumers generate purchase intention while watching short videos, which persona perception factors foster shared value with platforms, and what factors influence shared value and purchase intention. Guided by the stimulus-organism-response (S-O-R) paradigm, the study posits persona perception as the stimulus, shared value creation as the organismal state, and purchase intention as the response, aiming to reveal direct, mediating, and moderating mechanisms relevant to precision marketing and consumer behavior on short video platforms.

The paper adopts the S-O-R framework to model how environmental stimuli (persona perceptions) influence internal states (shared value creation) and behavioral responses (purchase intention), applicable to collectivist cultural contexts like China. Hypotheses are developed drawing on prior research: (1) Credibility of influencers fosters trust and engagement and is expected to positively affect shared value creation (CSV). (2) Willingness to use the platform/influencer relates to continued engagement and economic contributions, predicting higher CSV. (3–5) Information quality characteristics—consistency, completeness, clarity—should enhance CSV given their roles in satisfaction, decision quality, and brand success. (6–8) Emotional/persona-related factors—likability, empathy, similarity—should increase CSV by building attachment, trust, and value co-creation. (9) CSV should positively influence purchase intention due to improved economic and social value creation and interactive selling. (10) Regulatory focus (promotion vs prevention) moderates the CSV–purchase intention link: promotion-focused individuals respond more to positive outcomes; prevention-focused to avoiding negative outcomes. (11) Social presence moderates the CSV–purchase intention link by shaping trust, warmth, and perceived human touch in mediated interactions. Control variables include age, gender, income, and brand awareness due to their known impacts on purchasing.

Design: Cross-sectional survey with structural equation modeling (SEM) to test an S-O-R-based model. Measures: Established scales adapted from prior studies. Credibility (4 items: Weismueller et al., 2020), Consistency (4: Shamala et al., 2017), Completeness (4: Kim et al., 2021), Clarity (3: Chu et al., 2018), Likability (4: Bornet & Brangier, 2016), Empathy (3: Mangus et al., 2020), Similarity (4: Xiao et al., 2020), Willingness to use (4: Aparicio et al., 2021), Shared value creation (8: Ham et al., 2020; economic and social dimensions), Purchase intention (3: Kim et al., 2012). Scale: 7-point Likert (1 = strongly disagree to 7 = strongly agree). Pretest: 20 short video users; interviews to refine items; problematic indicators removed; final questionnaire had 41 items. Translation: Back-translation (English–Chinese–English) with expert panel review to ensure content and surface validity. Sampling and data collection: Conducted in Chinese universities (summer 2022). Inclusion: participants familiar with short video platforms. Administration: in-class, face-to-face structured interviews for clarity; average completion time ≈ 18 minutes. Responses: 400 collected over 4 weeks; 350 valid after removing 50 incomplete/invalid. Sample: 28% male, 72% female; age mainly 21–29 (53.7%) and <20 (45.4%); 94.9% monthly income ≤ RMB 3000; daily viewing ≤1 h (46.6%), 1–2 h (36.3%). Common method bias: Procedural remedies (anonymity, item order, antisense items); Harman’s single-factor test: first factor 28.587% (<40%), suggesting CMV controlled. Analysis: Confirmatory factor analysis, reliability (Cronbach’s α and composite reliability >0.7), convergent validity (AVE >0.5; loadings >0.7), discriminant validity (Fornell–Larcker). Model fit: χ2 = 2173; CFI = 0.911; TLI = 0.897; IFI = 0.913; RMSEA = 0.058. Hypotheses testing via AMOS SEM. Mediation: bootstrapping (95% CI) for indirect effects. Moderation: mean-centering and interaction terms for regulatory focus and social presence; multi-group comparisons (promotion vs prevention; high vs low social presence). Robustness: PROCESS (SPSS 24.0) Model 14 for moderated mediation.

- Measurement quality: All constructs showed acceptable reliability (Cronbach’s α and CR > 0.7) and convergent validity (AVE > 0.5); discriminant validity supported. Model fit indices acceptable (CFI 0.911; RMSEA 0.058). - Direct effects on shared value creation (CSV): • Credibility → CSV: β = 0.122, p < 0.01 (H1 supported) • Consistency → CSV: β = 0.152, p < 0.001 (H2 supported) • Completeness → CSV: β = 0.121, p < 0.05 (H3 supported) • Clarity → CSV: β = 0.213, p < 0.001 (H4 supported) • Likability → CSV: β = 0.105, p < 0.05 (H5 supported) • Empathy → CSV: β = 0.113, p < 0.05 (H6 supported) • Similarity → CSV: β = 0.072, p < 0.05 (H7 supported) • Willingness to use → CSV: β = 0.293, p < 0.001 (largest effect among persona dimensions; H8 supported) - CSV → Purchase intention: β = 0.878, p < 0.001 (H9 supported), indicating a strong positive effect. - Mediation: Bootstrapped indirect effects from persona perception dimensions to purchase intention via CSV were significant; 95% CIs excluded zero (Table 7), supporting CSV’s mediating role. - Moderation (Regulatory focus): Interaction CSV × Regulatory focus significant (β = 18.717, p < 0.01). Group analysis: CSV → Purchase intention significant for promotion-focused (β = 0.808, p < 0.001) and prevention-focused (β = 0.665, p < 0.001) users, with stronger effect for promotion-focused (H10 supported; consistent with H10a/H10b logic). - Moderation (Social presence): Interaction CSV × Social presence significant (β = 11.470, p < 0.05). High social presence group: CSV → Purchase intention β = 0.616, p < 0.001; Low social presence group: β = 0.320, p = 0.129 (n.s.), indicating a positive moderating effect primarily at higher social presence (H11 supported). - Robustness: PROCESS Model 14 confirmed significant mediation (CSV) and moderation effects; moderated mediation confidence intervals indicated reliable results.

Findings support the S-O-R mechanism in short video marketing: persona perception (stimulus) shapes shared value creation (organism), which in turn strongly drives purchase intention (response). All examined persona dimensions positively influence shared value creation, with willingness to use and clarity among the strongest contributors, highlighting the importance of both content presentation quality and audience inclination to engage with influencers. Shared value creation effectively explains how persona perceptions translate into intention to purchase, reinforcing the importance of economic and social value co-creation between influencers/platforms and viewers. Individual differences shape these effects: promotion-focused consumers respond more to positive shared value outcomes, while prevention-focused consumers are more sensitive to messages framed around avoiding negative outcomes; higher social presence amplifies the CSV–intention linkage by enhancing warmth, trust, and perceived human touch. These results inform theory by extending S-O-R with persona perception antecedents and dual moderators (regulatory focus and social presence), and inform practice by emphasizing persona design, information quality, and social presence enhancement to increase purchase intentions on short video platforms.

This study extends the S-O-R framework to short video marketing by identifying persona perception dimensions as key antecedents of shared value creation, demonstrating CSV’s mediating role, and showing that regulatory focus and social presence moderate the CSV–purchase intention link. Empirical evidence from 350 Chinese short video users indicates that credibility, information quality (consistency, completeness, clarity), likability, empathy, similarity, and willingness to use all enhance shared value creation, which in turn strongly predicts purchase intention. The study contributes to consumer behavior theory in social commerce and offers practical guidance for brands and merchants to optimize influencer personas, content quality, and platform/social presence features for precision marketing. Future research should adopt longitudinal designs, expand samples across regions and cultures, and incorporate additional variables (e.g., product involvement) to further generalize and refine the model.

- Cross-sectional design limits causal inference and does not capture temporal dynamics; longitudinal studies are recommended. - Cultural and regional specificity: data from Chinese university contexts may limit generalizability to other cultures and age/income segments; cross-cultural and broader demographic samples are needed. - Model scope: while key antecedents, mediator, and moderators are included, other relevant variables (e.g., product category characteristics, product involvement, platform algorithms) may further explain purchase intention in short video contexts.