Purpose – This study examines the adjusted associations between recorded transaction attributes and realised online printed-book invoice value at PT XYZ during 2023–2025, using the 7P marketing mix as a conceptual organising lens. Design/methodology/approach – A retrospective observational design was applied to a census of 5,989 eligible completed online printed-book invoices. The unit of analysis was an invoice rather than a unique consumer. Ordinary least squares regression was estimated for logged invoice value with HC3 heteroskedasticity-consistent robust standard errors. The primary explanatory variables were academic-book share, destination region, purchase channel, and transaction year. Average selling price and basket size were excluded from the primary model because they are mechanically related to invoice value. Finding/Results – The model explained 28.4% of the variation in logged invoice value, and the predictors were jointly associated with the outcome. A 10-percentage-point increase in academic-book share was associated with an estimated 3.2% higher invoice value. Invoices delivered to Java outside Jabodetabek and outside Java had estimated values approximately 9.0% and 12.9% higher, respectively, than invoices delivered within Jabodetabek. Marketplace and WhatsApp transactions had estimated values approximately 5.3% and 7.7% higher than webstore transactions. Estimated invoice values were also higher in 2024 and 2025 than in 2023. Robustness checks showed that the estimates were stable after excluding non-individual and influential invoices. Originality/Value – This study provides transaction-level evidence from actual completed online book invoices rather than survey-based perceptions. The recorded attributes are interpreted as observable transaction characteristics conceptually related to the 7P framework, not as direct measures of consumers’ perceptions or causal effects of marketing activities.