The research is motivated by the massive data generation resulting from digital transformation and the need to leverage this data for improved organizational performance. Data-driven decision-making (DDDM) is highlighted as a crucial strategy for harnessing the value of data. The study focuses on Pakistan's banking sector, a significant contributor to the country's GDP, but one with a relatively low rate of bank account ownership. Digitalization in this sector began in 2008, and while some banks are investing in data analytics, the application of these analytics in decision-making remains under-researched. The study aims to address this gap by examining the relationship between DDDM, data analytics capability, and bank productivity. The research contributes by empirically investigating the impact of DDDM on firm productivity, exploring the moderating role of data analytics investment, and employing the Instrumental Variable Two-Stage Least Square (2SLS-IV) methodology to address endogeneity concerns common in studies of this nature. It also explores the organizational adoption of digital transformation and its impact on decision-making, a previously under-researched area in Pakistan's banking sector.

The literature review establishes the critical role of effective decision-making in organizational success. Big data analytics (BDA) is presented as a tool for facilitating faster, more precise, and efficient decision-making, reducing risks and extracting valuable insights. However, effective BDA requires reliable data and a data-driven organizational culture. The study draws on the diffusion of innovation theory (DOI) to understand how data analytics and DDDM can boost productivity. Existing research indicates a strong positive correlation between DDDM practices and firm performance, with studies showing productivity gains of 5-6% for DDDM-adopting firms. The literature also highlights the importance of investment in big data and analytics for enhancing organizational performance. However, the impact of data analytics investment on the relationship between DDDM and firm performance is less explored, forming a key focus of this study.

The study includes all commercial and microfinance banks registered with the State Bank of Pakistan (excluding foreign and specialist institutions), resulting in a sample of 36 banks. Data was collected using a structured survey administered to chief information officers, data analysts, IT heads, and senior bank management. The survey, based on a previously validated questionnaire, gathered data on data analytics investment, IT utilization, and adaptation to organizational change associated with DDDM adoption. Secondary data, including output measures (loans and deposits), capital, and employees, were collected from publicly available sources. To address endogeneity issues, the study employs the Instrumental Variable Two-Stage Least Square (2SLS-IV) method. For the first model examining the impact of DDDM on productivity, lagged output and adjustment cost were used as instruments. For the second model, which also includes data analytics investment as a moderator, lagged output and exploration (a measure of innovation and investment in new technologies) were used as instruments. Principal component analysis (PCA) was used to create composite indices for DDDM, exploration, adjustment cost, and human capital, ensuring the validity of the PCA was tested using the Bartlett test of sphericity and the Kaiser-Meyer-Olkin measure of sampling adequacy. The Cobb-Douglas production function was used to model bank productivity, with loans and deposits as the output variable. Two models were tested. The first tests the impact of DDDM on productivity, while the second incorporates data analytics as a moderator.

The descriptive statistics indicate that the sample banks are relatively large and stable, with a low mean non-performing loan ratio. The Cronbach's alpha values for the composite indices (DDDM, Adjustment Cost, Exploration, Human Capital) suggest acceptable internal consistency reliability. The IV-2SLS regression results show a statistically significant and positive impact of DDDM on bank output (a 10.5% increase). The Sargan test indicates that the instruments were valid and did not influence the results. The model including data analytics as a moderator shows that data analytics investment also has a significant positive impact on bank output (6.7% increase). Importantly, the interaction term (DA*DDDM) is also statistically significant and positive, indicating that the positive impact of DDDM on output is further enhanced (by an additional 4.43%) when banks invest in data analytics. Control variables, including Z-Score and bank type (commercial vs. microfinance), were also significant, with commercial banks showing better performance.

The findings support the hypothesis that DDDM positively impacts bank productivity and that this impact is amplified by investment in data analytics. These results align with the diffusion of innovation theory and existing literature, emphasizing the importance of both technological investment and organizational change for realizing the benefits of DDDM. The use of 2SLS-IV effectively addresses endogeneity concerns, enhancing the reliability of the causal inferences. The significant and positive interaction term between DDDM and data analytics highlights the complementarity of these factors. The results suggest that simply investing in data analytics is insufficient; a culture of data-driven decision-making across the organization is necessary to fully reap the benefits.

This study provides strong empirical evidence for the positive impact of DDDM on bank productivity in Pakistan, further enhanced by investment in data analytics. The findings highlight the importance of a holistic approach that combines technological investment with organizational change to successfully implement DDDM. Future research could explore the impact of DDDM on other performance metrics (risk management, profitability, etc.) and investigate the generalizability of the findings to other industries and contexts.

The study focuses solely on the Pakistani banking sector, limiting the generalizability of the findings to other contexts. The study also uses productivity as the primary performance measure, potentially overlooking other important outcomes. Future research should explore broader performance indicators and extend the analysis to other sectors and countries to test for robustness.