User stickiness to facial recognition payment technology: insights from Sako's trust typology, privacy concerns, and a cross-cultural context

The study addresses growing adoption of facial recognition payments (FRP) amid persistent concerns over biometric data privacy and security. Prior research largely examines intention to use, leaving a notable intention–behavior gap for actual continuous usage. The research question is: What factors can affect and determine individuals’ stickiness (continuous usage behavior) to FRP? Grounded in the theory of planned behavior (TPB), the study posits that attitudes and perceived behavioral control (PBC), alongside subjective norms, influence stickiness. To strengthen explanatory power, Sako’s (1992) trust typology—competence, contractual, and goodwill trust—is integrated as antecedents to attitudes, PBC, and privacy concerns. Privacy concerns are expected to negatively affect attitudes and stickiness. A cross-cultural comparison between China and the USA is incorporated due to differing cultural values, regulatory environments, and acceptance of FRP: China tends to prioritize convenience and collectivist benefits, while the USA emphasizes individual privacy and data protection. The study aims to clarify mechanisms through which trust and privacy concerns shape FRP stickiness across these contexts.

The review outlines FRP’s growing deployment across commerce (vending, retail, hospitality) and distinguishes continuous use behavior (stickiness) from intentions. Stickiness originates from website loyalty literature and reflects repeated, habitual use driven by satisfied experiences and perceived value. In FRP, higher stickiness indicates choosing FRP over alternatives and more frequent use. Trust is critical for ongoing technology use given FRP’s inherent risks (biometric data, leakage). Prior FRP studies often treat trust as monolithic (e.g., trust in technology or provider) and focus on intentions. Sako’s (1992) trust typology provides a nuanced framework: competence trust (belief in capability and quality), contractual trust (belief in adherence to agreements/promises and ethical standards), and goodwill trust (belief the partner will act benevolently beyond formal obligations). In FRP, competence relates to reliable, efficient payments; contractual trust to security guarantees and privacy protection; goodwill trust to friendly, supportive service beyond expectations. Cross-cultural literature shows marked differences: Chinese users (collectivist context, supportive regulations) exhibit higher acceptance and willingness to trade privacy for efficiency; American users (individualist context, stricter/fragmented regulation) express stronger privacy concerns and skepticism about corporate use of biometrics. These differences likely influence continuous FRP behavior, motivating a comparative investigation.

Design: Cross-sectional online survey of FRP users in China and the USA via Credamo. Total responses: 1356; valid: 1278 (China 648; USA 630); response rate 91.29%. Sampling ensured single-response per IP and random selection from a large panel. Translation followed back-translation; expert review and pretest ensured content validity. Several attention and quality checks confirmed FRP experience. Nonresponse bias was checked using early–late comparison; no significant differences found. Measures: 7-point Likert scales. Sako’s trust typology—competence trust (2 items), contractual trust (2), goodwill trust (3) adapted from Fatima & Razzaque (2014). TPB constructs: attitudes (4), PBC (3), subjective norms (3) from Nasri & Charfeddine (2012). Privacy concerns (3) from Cheng et al. (2024). FRP stickiness (3) from Chiang & Hsiao (2015). Controls: trust propensity (3) from Wanner et al. (2022), technological literacy (3) from Gonzales II & Gonzales (2024), Kim & Jeon (2020), individual innovativeness (4) from Lee (2024), and prior FRP usage experience (tenure categories). Common method bias (CMB): Ex ante procedural remedies (scale/anchor variation, anonymity, construct masking); ex post tests included Harman’s single-factor (no factor >50%), full collinearity (all VIFs <3.3), and marker variable (profession) showing no significant correlations with model variables. Analysis: Partial least squares structural equation modeling (PLS-SEM) via SmartPLS 4 with 10,000 bootstrap resamples. Reliability acceptable (loadings 0.731–0.875; CR 0.794–0.901; alpha 0.766–0.889); convergent validity (AVE 0.562–0.728); discriminant validity via HTMT (<0.85). Structural model assessed with path coefficients, R², and Q². Multi-group analysis compared Chinese and USA samples.

- Model explanatory power: R² for attitude = 0.626; PBC = 0.418; privacy concerns = 0.457; FRP stickiness = 0.779. Predictive relevance Q²: attitude = 0.578; PBC = 0.392; privacy concerns = 0.413; stickiness = 0.743. - TPB effects on stickiness (full sample, n=1278): - Attitudes → stickiness: β = 0.233, p<0.001 (H1 supported). - PBC → stickiness: β = 0.251, p<0.001 (H2 supported). - Subjective norms → stickiness: β = 0.084, ns (H3 not supported). - Trust typology effects: - Competence trust → attitude: β = 0.358, p<0.001 (H4); → PBC: β = 0.387, p<0.001 (H5). - Contractual trust → attitude: β = 0.178, p<0.05 (H6); → PBC: β = 0.181, p<0.05 (H7). - Goodwill trust → attitude: β = 0.321, p<0.001 (H8); → PBC: β = 0.298, p<0.001 (H9). - Privacy concerns effects: - Privacy concerns → attitude: β = -0.162, p<0.05 (H10 supported). - Privacy concerns → stickiness: β = -0.153, p<0.05 (H11 supported). - Trust reducing privacy concerns: - Competence trust → privacy concerns: β = -0.196, p<0.01 (H12 supported). - Contractual trust → privacy concerns: β = -0.187, p<0.01 (H13 supported). - Goodwill trust → privacy concerns: β = -0.179, p<0.01 (H14 supported). - Control variables on stickiness: trust propensity (β = 0.009), technological literacy (β = 0.022), individual innovativeness (β = 0.017), prior FRP experience (β = 0.011) all nonsignificant (H15–H18 not supported). - Mediation (complementary) observed for: attitudes and PBC mediating effects between each trust type and stickiness; privacy concerns mediating between each trust type and stickiness; and attitudes mediating between privacy concerns and stickiness. - Cross-cultural differences (MGA): In China, privacy concerns had nonsignificant effects on attitudes and stickiness; in the USA, privacy concerns significantly and negatively affected attitudes and stickiness. Effects of competence, contractual, and goodwill trust on attitudes, PBC, and privacy concerns were positive/negative as hypothesized in both countries. Competence trust emerged as the strongest antecedent overall.

The findings answer the research question by showing that user stickiness to FRP is driven primarily by positive attitudes and higher perceived behavioral control, which are themselves strengthened by distinct forms of trust per Sako’s typology. Competence, contractual, and goodwill trust improve attitudes and PBC and reduce privacy concerns, thereby increasing stickiness. Privacy concerns undermine both attitudes and stickiness, highlighting the centrality of perceived data security in sustaining usage, especially in privacy-sensitive contexts. The study clarifies mechanisms beyond intention by focusing on actual continuous use behavior (stickiness), demonstrating that trust operates through TPB paths (attitudes, PBC) and via reduced privacy concerns. Cross-cultural analysis reveals that privacy concerns influence behavior differently across contexts: they are salient deterrents in the USA but less consequential in China, where convenience may outweigh privacy worries. These insights advance theory by integrating trust typology and privacy concerns into TPB for sustained technology use and inform practice on how to enhance FRP’s long-term adoption across cultures.

This paper extends FRP research from intentions to continuous use behavior by operationalizing stickiness and integrating TPB with Sako’s trust typology and privacy concerns. It demonstrates that attitudes and PBC drive stickiness, subjective norms do not, and that competence, contractual, and goodwill trust bolster attitudes and PBC and reduce privacy concerns. Privacy concerns directly dampen attitudes and stickiness, with competency trust emerging as the most influential antecedent. Cross-cultural analysis shows privacy concerns are decisive in the USA but not in China, emphasizing the need for context-specific strategies. Future research should: (1) employ longitudinal designs to test causality among trust, TPB constructs, privacy concerns, and stickiness; (2) examine alternative trust frameworks (e.g., initial vs. experiential trust) to unpack dynamics over time; and (3) investigate scenario-specific stickiness (e.g., public transportation vs. private retail) where perceived stewardship (government vs. private) may alter privacy risk perceptions and sustained use.

- Cross-sectional design limits causal inference; longitudinal studies are recommended to establish temporal ordering among constructs. - Trust conceptualization limited to Sako’s typology; future work could compare other trust forms (e.g., initial vs. experiential trust) and their dynamics. - General stickiness measured across contexts; stickiness may vary by payment scenarios (e.g., public transit vs. private retail) due to differing perceptions of data stewardship and risk. - Self-reported survey data, although CMB checks were satisfactory, may still contain measurement biases.