Digital business transformation forces firms to develop foundational capabilities to remain competitive. However, despite considerable academic and managerial interest, the nature of a digital business capability (DBC) that creates value by effectively managing digital business transformation remains unclear. Drawing on a mixed-methods approach, we conceptualize and operationalize the DBC construct. In Study 1, we examine the effects of DBC on firm performance using a cross-industry, multisource dataset. In Study 2, we assess the effects of DBC on customer performance using a unique multisource, multilevel dataset collected at two points in time. The results reveal that DBC contributes to performance, even beyond the effects of established constructs. Importantly, DBC increasingly drives firm performance after reaching a critical level of internal dynamism (i.e., U-shaped moderation). By contrast, DBC particularly pays off at an optimal level of external dynamism (i.e., inverse U-shaped moderation). DBC is more valuable for business-to-consumer than for business-to-business firms.
We follow Marsh et al. (2013) and take the recommended thresholds of Hu and Bentler (1999) as criteria for good model fit. Accordingly, fit indices should ideally be close to the following thresholds: CFI ≥ .95, TLI ≥ .95, SRMR ≤ .08, and RMSEA ≤ .06.
This CFA included DBC, structural flux, technological dynamism, IT capabilities, digital marketing capabilities, market orientation, entrepreneurial orientation, interaction orientation, and technology orientation. We used item parceling to develop aggregated scales for all second-order constructs (Bagozzi and Edwards 1998). Before running this overall CFA, we conducted separate CFAs and validated the second-order constructs market, entrepreneurial, and interaction orientation (e.g., Vorhies et al. 2011).
The regression model is specified as follows: ROS = β0 + β1 DBC + β2 B2B/C + β3 TD + β4 SF + β5 DBC × B2B/C + β6 DBC × TD + β7 TD2 + β8 DBC × SF + β9 SF2 + β10 DBC × TD2 + β11 DBC × SF2 + β12λ + βcontrols CONTROLS + ε, where ROS is the industry-adjusted return on sales, B2B/C is B2B versus B2C, TD is technological dynamism, SF is structural flux, λ is the inverse Mills ratio, and CONTROLS includes the variables firm size, firm age, product versus service, and relevance of digital business transformation.
We calculated the slopes at the low and high end of technological dynamism using unstandardized regression coefficients (not the standardized regression coefficients reported in Table 6), by deriving the following equation (Jaccard and Turrisi 2003): \( \frac{\partial^2\mathrm{ROS}}{\partial \mathrm{DBC}\partial \mathrm{TD}} \) = β6 + 2 β10 TD. Thus, the turning point is \( -\frac{\upbeta_6}{2\ {\upbeta}_{10}} \) .
Chi-square is a direct function of the sample size, so it cannot be used to meaningfully judge model fit given our large sample of 3212 customers (cf. Homburg et al. 2015).
Direct effects model: digital strategy ➔ firm performance (b = −.01, p > .10), digital strategy ➔ innovation performance (b = .26, p < .05), digital strategy ➔ customer relationship performance (b = −.11, p > .10); digital integration ➔ firm performance (b = .18, p < .10), digital integration ➔ innovation performance (b = .11, p > .10), digital integration ➔ customer relationship performance (b = .20, p < .10); digital control ➔ firm performance (b = .19, p < .05), digital control ➔ innovation performance (b = .08, p > .10), digital control ➔ customer relationship performance (b = .13, p > .10).