Six of Eight Factors Drive Big Data Analytics Adoption in Sudanese Telecoms, Study Finds

Big data analytics has become one of the most closely watched technologies in the telecommunications industry, promising better decision-making, sharper customer insight, and a durable competitive edge. Yet a growing number of organizations report that their analytics investments have not delivered the business value they expected. A new study from Sudan, published in Frontiers in Big Data, offers one of the first integrated empirical tests of what actually drives big data analytics adoption in a developing country’s telecom sector — and what happens to performance once the technology is in place.

The research team, led by Ibrahim Magboul of Ahfad University for Women in Omdurman and Diaeldin Osman of Clark Atlanta University, surveyed 201 employees across Sudan’s three major telecommunication operators — Zain, MTN, and Sudani — all headquartered in the capital, Khartoum. Using a purposive sampling technique, the researchers distributed 291 questionnaires between July and August 2023 and achieved a 69 percent valid response rate. Respondents ranged from junior staff to top-level executives, with 60 percent identifying as middle-level employees. Most had substantial hands-on experience with analytics tools: 60 percent had used big data analytics for one to three years, and 90 percent reported generating reports with these tools twice a day.

What makes the study methodologically distinctive is its integration of four established theoretical frameworks into a single validated model. The Technology Acceptance Model (TAM) captures individual user perceptions at the micro level; the Technology-Organization-Environment (TOE) framework captures organizational context at the meso level; the IS Success Model focuses on information and system characteristics; and the Resource-Based View (RBV) explains how analytics capabilities become strategic resources. The authors argue that big data adoption is a multi-level phenomenon that no single theory can adequately explain, and that the causal mechanisms identified in developed economies may not universally apply in resource-constrained environments.

To test their model, the researchers deployed partial least squares structural equation modeling (PLS-SEM), a statistical technique that estimates relationships between latent constructs while accounting for measurement error. The analysis used 5,000 bootstrap subsamples with bias-corrected 95 percent confidence intervals. Before testing the structural paths, the team rigorously validated the measurement model: all factor loadings exceeded the 0.7 threshold recommended by Hair and colleagues, all Cronbach’s alpha values surpassed 0.7, composite reliability exceeded 0.7 for every construct, and average variance extracted exceeded 0.50, confirming convergent validity. Discriminant validity was confirmed using both the Fornell-Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio, with all values below the 0.9 threshold. Variance inflation factors stayed well below 3.3, ruling out pathological collinearity and common method bias.

The structural results were striking. Six of the eight hypothesized antecedents significantly predicted big data analytics adoption. Information quality emerged as the strongest driver, with a path coefficient of 0.397 (t = 7.048, p < 0.001), confirming that accurate, complete, consistent, and timely data is the single most powerful enabler of analytics success. Information center support followed at 0.239 (t = 3.511), vendor support at 0.201 (t = 3.173), business-IT alignment at 0.195 (t = 3.703), big data culture at 0.154 (t = 2.891), and training at 0.149 (t = 2.955). All six effects were statistically significant at the 5 percent level.

But two findings defied expectations. Big data competency — the organization’s collective ability to manage, analyze, and derive insights from big data — showed essentially no effect on adoption (β = 0.001, t = 0.020, p = 0.984). Perceived usefulness, the core TAM construct capturing an employee’s belief that analytics tools enhance job performance, was similarly non-significant (β = 0.011, t = 0.267, p = 0.790). The authors interpret these null results not as measurement failures but as theoretically generative anomalies that reveal something fundamental about technology adoption in extreme resource-constrained environments.

For big data competency, the researchers propose a “capability-use gap”: organizations may possess competent personnel but lack the complementary infrastructure, managerial prioritization, or resource alignment necessary to translate that competency into actual adoption. Sudan’s telecom sector faces persistent challenges including electricity outages, limited bandwidth, and outdated computing equipment — and since April 2023, armed conflict has disrupted telecommunications infrastructure across the country. Three mechanisms may explain the finding: infrastructural dependency, where skills alone cannot compensate for inadequate hardware and networks; strategic misalignment, where competent employees work in departments where analytics is not prioritized by top management; and a measurement-reality gap, where survey-based self-reports overstate actual capabilities or distribute them unevenly across departments.

The non-significant perceived usefulness result challenges the universal applicability of TAM itself. The authors suggest a “feasibility-before-perception” pathway: in Sudan’s telecom sector, adoption decisions may be constrained by external feasibility barriers — electricity outages, internet instability, hardware unavailability, and conflict-related infrastructure destruction — that render perceptions of usefulness irrelevant if the technology cannot physically be used. Mandatory adoption policies and “symbolic adoption,” where organizations claim analytics usage for legitimacy without substantive implementation, may further disconnect employee perceptions from actual behavior. The authors propose that future research test a “reverse TAM” model in which objective infrastructural conditions predict perceived usefulness and ease of use, rather than the traditional path where perceptions predict adoption intentions.

Crucially, the study confirms that big data analytics adoption has a strong positive effect on perceived performance (β = 0.647, t = 15.871, p < 0.001). The model explained 66.1 percent of the variance in analytics adoption and 41.9 percent in perceived performance — both classified as substantial predictive ability under Cohen’s criteria. This finding aligns with prior research showing that analytics-enabled sensing capabilities and analytics-driven cultures improve organizational outcomes, and it assures managers that the technology can deliver measurable business value even in difficult operating environments.

The practical implications are clear. Managers should carefully select analytics vendors with deep telecommunications experience, invest in continuous training in data literacy and visualization, institute robust data governance policies to safeguard information quality, and strengthen business-IT alignment so that analytics initiatives connect to strategic objectives. The authors recommend that governments in developing countries initiate technology interventions specifically aimed at helping companies adopt analytics tools that promote better performance. The study’s limitations — a sample drawn from only three operators in one country, a cross-sectional design, and the extraordinary circumstances of data collection during an active armed conflict — suggest future research should expand to larger samples, other sectors, and longitudinal designs. But the core message stands: in resource-constrained environments, feasibility comes before perception, and infrastructure, data quality, and external support are the true engines of analytics success.

Subject of Research: Big data analytics adoption and perceived performance in Sudanese telecommunication companies

Article Title: An empirical test of antecedents and perceived performance of big data analytics in telecommunication companies

Article References: Magboul, I., Osman, D., Herzallah, F., Osman, A. N., & Gittens, D. (2026). An empirical test of antecedents and perceived performance of big data analytics in telecommunication companies. Frontiers in Big Data, 9, Article 1763520. https://doi.org/10.3389/fdata.2026.1763520

Image Credits: AI Generated

DOI: 10.3389/fdata.2026.1763520

Keywords: big data analytics, telecommunications, Sudan, structural equation modeling, technology acceptance model, developing countries, empirical, test, antecedents, perceived, performance, data