The cognitive construct of data: Visibility bias at data genesis in influencer-driven business intelligence

Research on business intelligence (BI) discusses bias downstream, in interpretation, assuming that metrics themselves are neutral representations of behavior. This study questions the epistemological integrity of the central database. It views bias as a cognitive, upstream conduct rooted in the data-generation stage. It uses a dual-method qualitative design to explore how visibility-based measures, such as likes and shares, become high-value metrics in influencer-driven BI systems. Starting with an epistemological analysis, a measurement selection test demonstrated that influencer metrics such as likes, shares, views, and followers capture only superficial, visible behaviors rather than underlying habits such as trust, reflection, and genuine engagement. Second, a reflexive thematic analysis of participants’ responses from a mock dashboard experiment revealed a dissociation in decision-making. Although high-visibility measures invoke familiarity and dominance effects, establishing an illusion of ubiquity, they were not associated with authenticity or value for participants. Instead, participants tend to see visibility as performative rather than meaningful. Decision-makers often confuse visibility with value despite understanding its constraints. The article explores process innovation in BI, analytically distinguishes between measurement bias and interpretive bias, and analyzes the significance of visibility-based metrics for acquiring epistemic authority.