When Algorithmic Persuasion Embraces You: Collaborative Relationship Styles Outperform Paternalistic Styles—The Role of Perceived Emotional Closeness and Self- Goal Strength

Abstract

As an increasing number of consumer interactions are empowered by algorithms and artificial intelligence, algorithmic persuasion has become an important tool for shaping consumer attitudes and behaviours. However, it faces a fundamental paradox. On the one hand, algorithms can use consumers’ personal data to provide unprecedentedly personalised content which is highly relevant to them and matches their preferences. Such personalisation has the potential to enhance consumers’ information processing fluency, generate a sense of “feeling right”, and ultimately increase persuasiveness. On the other hand, personalisation can also serve as a cue that activates consumers’ persuasion knowledge, making them feel manipulated, question the genuine motives behind algorithmic persuasion, and even perceive a threat to their autonomy, which may trigger psychological reactance and consequently undermine persuasion effectiveness.
Drawing on an abductive sequential mixed-method design, this thesis reveals that the relationship style of algorithmic persuasion is a potential key to navigating this tension between personalisation and persuasion knowledge, while maintaining consumers’ psychological autonomy and critical judgement. Specifically, collaborative relationship style algorithmic persuasion aligns more closely with consumers’ cognitive and affective psychological states than paternalistic relationship style persuasion, thereby achieving better persuasive outcomes. Furthermore, this thesis identifies emotional closeness as the mediating mechanism through which relationship style exerts its influence on algorithmic persuasion, interpreted through the lens of construal level theory. Finally, this thesis identifies self-goal strength as a critical moderating factor that shapes the effect of relationship style on persuasiveness. When consumers possess strong and clearly defined self-goals, the advantage of collaborative relationship style algorithmic persuasion over paternalistic relationship style one becomes less pronounced, as intrinsically driven consumers are less susceptible to external interventions.
Overall, this thesis theorises the tension between personalisation and persuasion knowledge in algorithmic persuasion. It fills the gap in the algorithmic persuasion literature, which has largely focused on cognitive mechanisms, by demonstrating the affective role of emotional closeness within empirical evidence. It theoretically contributes to the literature on algorithmic persuasion, particularly in relation to the matching effect and construal level theory. Practically, it provides guidance for practitioners seeking to embed algorithms to enhance consumer experience and wellbeing. Finally, this thesis critically discusses its limitations and offers directions for future research to build upon these findings.