Applications and Competitive Advantages of Data Mining and Business Intelligence in SMEs Performance: A Systematic Review

1. Introduction

In today’s rapidly evolving technological landscape, the adoption of data mining and business intelligence (BI) technologies has significantly transformed organizational oper-ations, particularly within small and medium-sized enterprises (SMEs). These emerging technologies have revolutionized traditional business practices by providing essential tools for analyzing vast amounts of data, deriving actionable insights, and enhancing de-cision-making processes, which collectively lead to a competitive advantage [1]. Data mining and BI have become integral components of the modern data-driven business en-vironment, playing a crucial role in determining organizational success or failure [2]. Re-cent research highlights the transformative impact of these technologies across various industries. Data mining, for instance, has been pivotal in improving customer relation-ship management, optimizing supply chain processes, and enhancing financial forecast-ing within SMEs [3,4]. Additionally, business intelligence has demonstrated significant improvements in strategic planning, performance management, and fostering innova-tion—elements crucial for sustaining competitive advantage in highly contested markets [5]. The application of data mining within knowledge management has also been ex-plored, illustrating its relevance to the transportation sector for SMEs [14]. Despite these advancements, a notable gap exists in understanding the comprehensive operation of these technologies within SMEs, especially in developing economies [6]. SMEs, represent-ing approximately 90 percent of businesses and over 50 percent of global employment, play a significant economic and social role, making their competitiveness vital for eco-nomic development [9]. However, SMEs face distinct challenges in surviving economic and global competition, necessitating efficient monitoring and use of information re-sources [9]. Business intelligence, which encompasses methodologies, processes, archi-tectures, and technologies to transform raw data into actionable insights, has emerged as a crucial tool for enhancing SME competitiveness [9,10]. Yet, literature on BI's applica-tion specifically within SMEs is sparse, with much research historically focusing on larger enterprises [9,11].

The increasing globalization and rapid technological changes have further dis-advantaged SMEs compared to larger counterparts, emphasizing the need for leveraging knowledge-building capabilities such as BI to gain a competitive edge [10]. The applica-bility of business intelligence maturity models to SMEs and the challenges associated with implementing BI systems in this sector underscore the complexity of this landscape [16]. Moreover, the cost and complexity of BI system implementation present significant barri-ers for SMEs, which often lack the specialized IT departments and financial resources of larger organizations [11]. Recent studies have explored various facets of BI and data min-ing within SMEs, including the deployment of cloud solutions from both vendor and cus-tomer perspectives [13], and the development of frameworks for mobile business intelli-gence in developing countries [15]. Furthermore, insights from the UK and Malaysia high-light the opportunities and challenges of BI and data mining in different contexts [17]. The interplay between BI systems and SME operations during times of crisis has also been examined, revealing critical aspects of ERP system adoption [19]. This systematic re-view aims to bridge the existing gaps by synthesizing information from a decade of re-search on data mining and BI in SMEs, identifying key trends, challenges, and opportuni-ties associated with these technologies. By analyzing studies published over this period, this review seeks to provide practical insights for SME practitioners and policymakers, ul-timately improving organizational performance and growth [7]. Table 1 presents a com-parative analysis of existing review works and our proposed systematic review, high-lighting the distinct focus of this study on the applications and competitive advantages of data mining and BI for SMEs. This review will explore how data mining and BI technolo-gies are utilized in the SME landscape to foster competitive advantage. By uncovering patterns in technological adoption, showcasing successful case studies, and discussing the implications of these findings, this review aims to lay the foundation for future re-search and practical improvements. It will also deepen the understanding of the strategic role that data mining and business intelligence play in the evolving operational land-scapes of SMEs [8].

In our systematic review, we identified several key research gaps within the existing literature on big data analytics (BDA), knowledge management (KM), and organizational performance (OP). These gaps highlight the areas where current studies fall short and provide avenues for future research to build a more comprehensive understanding of the field.

Firstly, while numerous studies have explored the impact of BDA on large enterprises, there is a notable lack of research focusing on small and medium-sized enterprises (SMEs). The distinct challenges and resource constraints faced by SMEs in adopting BDA are often overlooked, leading to a gap in tailored strategies that address their specific needs. Additionally, existing literature predominantly emphasizes the technological aspects of BDA, with insufficient attention given to the human and organizational factors that play a critical role in successful implementation. Factors such as employee skills, change management, and organizational culture are underexplored, leaving a gap in understanding how these elements influence the effective use of BDA.

Secondly, there is limited integration of KM practices with BDA initiatives in the literature. While KM is recognized as a valuable component for leveraging data-driven insights, studies often treat KM and BDA as separate entities rather than examining their synergistic effects on organizational performance. This gap suggests a need for research that investigates the combined impact of KM and BDA on enhancing decision-making processes, innovation, and competitive advantage. Furthermore, many studies use cross-sectional data, which limits the ability to draw conclusions about the long-term effects of BDA on OP. There is a clear need for longitudinal studies that can capture the evolving dynamics between BDA adoption, KM practices, and sustained organizational performance over time. By addressing these research gaps, future studies can provide more actionable insights for both scholars and practitioners aiming to optimize the use of BDA in organizational settings.

1.1. Research Questions

Although a considerable amount of research has been conducted on data mining and business intelligence, there is still a need for an in-depth examination of their applications and competitive advantages for small and medium-sized enterprises (SMEs). Consequently, the current work proposes to explore how data mining techniques and business intelligence systems can be leveraged to improve SME performance. To achieve this, the subsequent research questions have been considered:

• How do algorithms of association as a method of data mining contribute to enhancing business intelligence in organizations?
• What are the impacts and roles of cloud-based big data analytics in knowledge management for achieving competitive advantages in organizations?
• How can pervasive business intelligence systems be used to gain and sustain competitive advantages in organizations?
• What opportunities and challenges do organizations face in the implementation of predictive analytics for competitive advantage, and how can these be effectively addressed?
• How does sentiment analysis, as an approach to business intelligence, enhance decision-making and operational efficiency in organizations?

1.2. Rationale

The rationale for this systematic review is to explore and evaluate the current state of research on the applications and competitive advantages of data mining and business intelligence in SMEs' performance. Given the increasing reliance on data-driven decision-making in small and medium-sized enterprises (SMEs), it is crucial to understand how these technologies can enhance their competitive edge. This review addresses the gap in existing literature by focusing on studies published within the last decade, specifically from 2014 to 2024, and aims to synthesize findings to provide a comprehensive understanding of how data mining and business intelligence contribute to SME success in various industries and geographic locations.

1.3. Objectives

The number one objective of this evaluation is to systematically examine and synthesize the existing research on the applications and Competitive Advantages of Data Mining and Business Intelligence in SME's Performance in improving the general performance of SMEs. This evaluation desires to identify key topics and layouts related to the adoption and impact of these technologies within small and medium-sized groups, supplying in-depth facts about the way they contribute to operational, economic, and innovative overall performance. Additionally, the comparison seeks to evaluate the impact of geographic and financial contexts on the effectiveness of facts mining and industrial organization intelligence in SMEs, spotting that good environments might also yield various outcomes. By thoroughly inspecting these aspects, the overview will offer actionable suggestions for SMEs on the manner to effectively influence records mining and commercial agency intelligence technologies to achieve and sustain a competitive benefit of their respective markets.

1.4. Research Contributions

This work introduces a detailed systematic survey of the applications and competitive advantages of data mining and business intelligence (BI) on the performance of small and medium-sized enterprises (SMEs). We spotlight various pending issues and research challenges in the deployment of data mining and BI techniques in SMEs. Following are the research contributions made by the proposed research work:

• We furnish a thorough business and economic analysis of data mining and BI, centering on the integration of advanced analytics, decision support systems, and data warehousing. This analysis underscores the cost-effectiveness, reliability, and strategic benefits of data-driven approaches, offering crucial insights for informed decision-making and promoting the adoption of business intelligence solutions within SMEs.
• We consolidate existing research on data mining and BI systems and identify gaps in the literature, particularly regarding the successful implementation of these systems in various SME contexts. By addressing these gaps, we highlight areas needing further research and innovation, thereby advancing the field of data mining and BI and ensuring enhanced SME performance and competitiveness.
• We also propose various regression models of financial metrics for assessing the impact of data mining, BI tools, and analytics platforms on SME performance.

1.5. Research Novelty

The proposed work has the following novelty. According to the best knowledge of the authors, there is no existing similar study in the literature that introduces a systematic review of the business and economic analysis of data mining and business intelligence, exclusively focusing on their applications and competitive advantages for SME performance.

• We provide a holistic business and economic evaluation of data mining and BI systems, focusing on their impact on decision-making, operational efficiency, customer relationship management, and financial performance across diverse SME sectors.
• We introduce novel linear regression models that elucidate the relationships between data-driven decision-making and key economic parameters, enhancing predictive accuracy for strategic planning in SMEs.

2. Materials and Methods

In this subsection, the study outlines the methodology employed to conduct a systematic review focusing on the applications and competitive advantages of data mining and business intelligence in SMEs' performance. The study is based on a review of literature published over the last decade, from 2014 to 2024. To the best knowledge of the authors, no similar comprehensive review exists within this specific timeframe, making this study a novel contribution to the field. The research methodology includes the careful selection of relevant peer-reviewed articles from key online databases, namely Scopus, Google Scholar, and Web of Science, ensuring a thorough examination of the subject matter.

2.1. Eligibility Criteria

A systematic study of all peer-reviewed and published research works relevant to the study of the applications and competitive advantages of data mining and business intelligence in SMEs' performance was conducted for examination. The research works that are published in the English language during the last decade, from 2014 to 2024, were considered. A proper criterion for inclusion was adapted to ensure the inclusion of research papers that specifically focus on this topic and exclude those that do not. Consequently, only peer-reviewed research works that fundamentally converge on the applications and competitive advantages of data mining and business intelligence in SMEs' performance, and that include a research framework or methodology specific to these aspects, were exclusively considered. The inclusion and exclusion criteria for this study are tabulated as in Table 2.

2.2. Information Sources

A systematic search of online databases was conducted to identify relevant studies for this review. The databases Scopus, Google Scholar, and Web of Science were utilized due to their comprehensive coverage of peer-reviewed literature in the field of data mining and business intelligence. Each database was thoroughly searched using a combination of keywords related to the study topic, ensuring that the most pertinent research articles were captured. Scopus provided access to a broad range of scientific journals and conference papers, while Google Scholar enabled the inclusion of gray literature and dissertations that might not be indexed elsewhere. Web of Science was used to cross-reference and ensure the robustness of the selected studies by providing citation data and impact factors of the journals. The search results from these databases formed the core of the literature review, ensuring a well-rounded and exhaustive collection of research works.

2.3. Search Strategy

The literature for this research was collected from well-known online research databases. A thorough search was carried out in three main repositories: Google Scholar, Scopus, and Web of Science. To find the most relevant studies, a specific set of keywords was used. These keywords were: ("Data Mining" AND "Business Intelligence" AND ("SME" OR "Small and Medium Enterprises") AND ("Applications" OR "Competitive Advantage" OR "Performance")). This combination of terms was chosen to ensure that the search captured studies directly related to the research topic. The search focused on papers published between 2014 and 2024. This time frame was selected to provide a recent and relevant overview of the subject. The search results included 6,550 papers from Google Scholar, 854 papers from Scopus, and 207 papers from Web of Science. After collecting these papers, they were carefully reviewed and filtered to select only those that were most relevant to the research questions. This process helped to narrow down the literature to the most useful and high-quality sources for this study. Table 3 shows the list of online repositories that were utilized as well as the total number of results achieved before the initial screening. The Bibliometric Analysis of Study Search Keywords is illustrated in Figure 1.

2.4. Selection Process

Four researchers (LM, MN, SV, BAT) independently reviewed the titles and abstracts of the first 60 records retrieved from the search. Any differences in the selections were discussed collectively until an agreement was reached. After this initial screening, the researchers worked in pairs to independently review the titles and abstracts of all retrieved articles. In cases where differences of opinion arose, discussions were held to determine which articles should proceed to full-text evaluation. If the researchers could not reach an agreement, the third researcher was consulted to make the final decision. Afterwards, three researchers (LM, MN, SV) independently assessed the full-text articles to determine whether they met the inclusion criteria. As before, any disagreements were resolved through discussion. If needed, the fourth researcher (BAT) was involved in making the final call on whether to include or exclude the articles, as shown in Figure 2.

2.5. Data Collection Process

To ensure that the data we collected from the studies was accurate, we followed a structured approach to minimize errors and reduce bias. Three reviewers independently collected the data from each study under supervision of fourth reviewer . Any differences in the extracted data were discussed until an agreement was reached. We used a data extraction form similar to the one from [32] to ensure consistency across all reviewers. We did not use any automation tools for data extraction. Data was carefully entered and double-checked for accuracy to avoid errors. When information in the studies was unclear, we conducted a thorough review of all available materials, including supplementary information, appendices, and related studies, to clarify the data. In cases where concerns remain, we consulted our fourth reviewer who is the subject matter experts to ensure the reliability of the data interpretation. When multiple reports from the same study were available, we established clear criteria to select the most relevant data, focusing on the most recent and comprehensive studies published between 2014 and 2024. In cases where the data from these reports did not match, we reviewed the methods and outcomes to resolve the differences. Only studies written in English were included, excluding any articles in other languages to maintain consistency in our analysis and avoid potential misinterpretations due to language differences, as shown in Figure 3.

2.6. Data Items

This section provides a comprehensive overview of the data items sought in this systematic review, focusing on both primary outcomes and additional variables relevant to the impact of data mining and business intelligence (BI) on small and medium-sized enterprises (SMEs). The primary outcomes encompass various dimensions such as operational efficiency, financial performance, strategic decision-making, and customer relationship management. In addition to these outcomes, the review also considers study and participant characteristics, intervention details, economic factors, and external influences, ensuring a thorough contextual understanding of the application and effects of BI technologies in SMEs. This approach allows for a nuanced analysis of how BI contributes to SME performance across diverse settings and conditions.

2.6.1. Data Collection Method

Efforts were made to ensure a comprehensive understanding on the impact of data mining and business intelligence (BI) on SMEs, and we thoroughly identified and defined relevant outcomes that capture the strategic, operational, and financial dimensions influenced by these technologies. Our approach was designed to synthesize robust evidence that reflects the transformative effects of BI in the SME context. The primary outcomes of this systematic review centered on several key domains that directly relate to the application of BI and data mining technologies in SMEs. Operational Efficiency was a major outcome, defined by measuring reductions in process completion times and error rates. We sought all results that could reflect how BI and data mining streamlined operations, optimized workflows, and improved resource utilization. These efficiency metrics provided clear insights into the practical benefits of technology adoption in enhancing operational processes.

Financial Performance was another critical outcome, assessed by tracking changes in revenue, cost savings, and overall return on investment. By quantifying the economic value added through BI tools, this outcome provided a holistic view of how data analytics contribute to the financial health and growth of SMEs. All relevant financial data points across studies were included to capture a comprehensive picture of BI's economic impact. Strategic Decision-Making was evaluated by examining the quality, speed, and efficacy of decisions influenced by BI insights. We looked at how well decision-making aligned with market trends and internal data forecasts, reflecting BI's capacity to support informed leadership actions and improve strategic planning. Results were sought across all measures and time points to understand the full extent of BI's influence on strategic decisions. Customer Relationship Management (CRM) was also a significant focus, with an emphasis on customer engagement and retention metrics. This outcome assessed how data analytics enhanced customer interactions and satisfaction. We specifically looked for studies reporting improvements in customer service driven by data-driven strategies, seeking all compatible results to thoroughly evaluate BI's impact on CRM.

2.6.2. Definition of Collected Data Variables

In addition to these primary outcomes, we carefully considered other variables to provide a detailed understanding of the context in which BI and data mining technologies were applied. These variables were critical in contextualizing the findings and understanding the broader implications of BI adoption in SMEs. Therefore, study characteristics were gathered, including information on the geographical location, industry specifics, and SME size, to assess the applicability of findings across different settings. These characteristics helped contextualize the outcomes and understand the diversity of the studies included. Participant characteristics were also documented, focusing on details about the employees using BI tools, such as their roles, level of BI literacy, and engagement with the technology. This information was essential for understanding the human factors influencing the successful deployment and utilization of BI systems within SMEs. Furthermore, intervention characteristics were described in detail, including the BI tools and data mining techniques employed, their integration with existing systems, and the scope of their use. These details were crucial for assessing the technological depth and breadth of the interventions and understanding their impact on SME performance.

Economic factors were another key consideration, particularly financial aspects such as initial and ongoing investments in BI technologies and the reported returns on these investments. These factors were important for evaluating the economic viability and sustainability of BI systems within SMEs. Finally, external influences such as broader market conditions, competitive pressures, and regulatory environments were considered to provide a comprehensive understanding of the external factors that affect the adoption and success of BI systems in SMEs. As shown in Table 4, our approach involved thorough manual searches across reputable online repositories, including Google Scholar, Scopus, and Web of Science, to gather the most relevant studies. These manual searches were tailored to capture the most relevant and accurate information, ensuring that our analysis was specifically focused on the applications and competitive advantages of data mining and business intelligence in SMEs. By identifying and defining these outcomes and variables, we ensured that our systematic review provides a robust and comprehensive analysis of the impact of BI and data mining technologies in the SME context. This methodical approach supports the reliability and relevance of our findings, making them valuable to stakeholders interested in the practical applications of these technologies.

2.7. Study Risk of Bias Assessment

In the studies, particularly those examining the impacts of data mining and business intelligence on SMEs, it was essential to critically evaluate the risk of bias to ensure the reliability and validity of the findings. To achieve this, we employed the Newcastle-Ottawa Scale (NOS) for assessing non-randomized studies, such as cohort and case-control studies. The NOS evaluates studies across three broad domains: Selection, Comparability, and Outcome (for cohort studies) or Exposure (for case-control studies). Each study was rated on a scale where a maximum of one star could be awarded per item within the Selection and Outcome/Exposure categories, and up to two stars for Comparability. This scoring reflects the overall quality of each study. As shown in Figure 4, the risk of bias assessment process involved four independent reviewers. Each study was evaluated independently by these reviewers to ensure objectivity. Disagreements among reviewers were resolved through discussions. If agreement could not be reached, the fourth reviewer was consulted to make the final decision. For studies with uncertainties or insufficient information, particularly those involving proprietary data mining tools or specific business intelligence applications, additional steps were undertaken. This included cross-referencing reputable sources such as Google Scholar, Scopus, and Web of Science to clarify uncertainties. Furthermore, a comprehensive manual search of online repositories was conducted to minimize bias and ensure that the risk of bias assessment was as accurate and thorough as possible. No automation tools were used in this process.

2.8. Synthesis Methods

This flow chart below in Figure 5 illustrates the systematic approach used in our review of data mining and business intelligence applications in SMEs. Starting with Study Selection Process, we identify and screen studies based on set eligibility criteria. Next, Data Standardization involves converting and cleaning the data to maintain consistency. In the Data Analysis phase, we present the data in tables or graphs and perform initial analyses. The flow then moves to Heterogeneity Assessment, where we evaluate variability through subgroup or sensitivity analyses. Finally, Bias Assessment ensures we identify potential biases and maintain transparency in our methods. This structured approach ensures a thorough and reliable review process.

In this systematic review on the application and competitive advantages of data mining and business intelligence in SMEs, we employed rigorous synthesis methods to ensure that our results were robust, transparent, and reproducible. To determine the eligibility of studies for synthesis, we meticulously tabulated the characteristics of each study and compared them against our predefined synthesis groups. This approach allowed us to include only the most relevant studies, ensuring that our findings were both valid and aligned with the review's objectives. In preparing the data for synthesis, we addressed missing summary statistics through imputation techniques and conducted necessary data conversions to maintain consistency across studies. The results were then presented using a combination of structured tables and forest plots, which provided a clear visual representation of the effect estimates and confidence intervals, enabling us to identify patterns and outliers effectively.

The synthesis of results was conducted using a random-effects meta-analysis model, chosen for its ability to account for variability between studies. This model was essential given the anticipated heterogeneity among the studies, which was further explored through subgroup analyses and meta-regressions. These analyses helped us identify potential sources of heterogeneity, such as SME size or the type of BI tools used and refine our understanding of the impact of these technologies. Additionally, sensitivity analyses were performed to assess the robustness of the synthesized results, ensuring that our conclusions were well-supported by stable and reliable evidence. Through this comprehensive approach, we were able to provide a meaningful aggregation of the evidence, offering valuable insights for stakeholders interested in leveraging data mining and business intelligence to enhance SME performance.

2.8.1. Eligibility for Synthesis

To determine study eligibility for inclusion in our systematic review on data mining and business intelligence (BI) in small and medium-sized enterprises (SMEs), each study was carefully evaluated for its relevance and alignment with the review’s objectives. We manually assessed and compared each study’s characteristics—such as intervention types and outcomes—against our predefined synthesis groups. A matrix was created to visually compare the scope and methodologies of the studies with our inclusion criteria, ensuring a comprehensive and objective evaluation. This process ensured that only studies directly pertinent to the review topic were included, thus enhancing the review’s overall rigor and reliability.

2.8.2. Data Preparation for Synthesis

In this review, the methods used involved converting or standardizing data collected from various studies to ensure consistency before synthesis. For example, when effect sizes were reported differently across studies, algebraic manipulations were employed to convert these into a uniform scale, such as converting odds ratios to risk ratios where appropriate. Additionally, handling missing data was a critical aspect of the analysis. Missing summary statistics, such as standard deviations or effect sizes, were imputed using established statistical methods like multiple imputation. This approach ensured that the dataset was comprehensive and robust, allowing for a more accurate and reliable analysis.

2.8.3. Tabulation and Visual Display of Results

Results from individual studies and synthesis efforts were organized using both tabular and graphical methods to enhance clarity and facilitate comparison. Tabular structures were employed to present the data in a structured format, where outcomes were organized by domain, and within each domain, studies were ordered from lowest to highest risk of bias. This organization allowed for easy comparison across studies and highlighted the most reliable evidence. Additionally, graphical methods, specifically forest plots, were used as the principal tool for visually displaying meta-analysis results. These plots showcased effect estimates and confidence intervals for each study alongside a summary estimate. The studies in the forest plots were ordered based on effect size or year of publication, helping to reveal trends over time and across different research focuses.

2.8.4. Synthesis of Results

During our manual search on online repositories such as Google Scholar, Scopus, and Web of Science, we carefully reviewed and synthesized the results of relevant studies. The approach to data synthesis was guided by the nature of the data and the degree of variability observed across studies. Based on the findings from our search, we manually assessed the applicability of both fixed-effects and random-effects models, depending on the level of heterogeneity among study results. The selection of the model was determined by the characteristics of the data and our assumptions about the consistency of effects across studies. After exporting the data to Excel, we created charts to visually inspect the data, allowing us to identify patterns of variability and potential heterogeneity across the studies. This initial visual inspection provided an overview of how study results differed from one another, facilitating a more nuanced analysis.

2.8.5. Exploring Causes of Heterogeneity

Subgroup analyses and meta-regression were performed to explore potential sources of heterogeneity, such as differences in study settings, intervention types, or outcome measures. Specific analyses focused on factors like the size of the SME, the type of business intelligence tool used, and the geographic location, all of which were examined to assess their impact on the effectiveness of data mining and business intelligence interventions. These methods helped to identify underlying patterns and relationships that contributed to the overall variability observed across the studies.

2.8.6. Sensitivity Analyses

Sensitivity analyses were employed to evaluate the robustness of the synthesis results in relation to various assumptions and methodological decisions made during the review process. These analyses included testing the impact of excluding studies at high risk of bias and using alternative statistical models to ensure that the conclusions were not unduly influenced by specific studies or analytical techniques. This approach helped to confirm the reliability and validity of the findings by addressing potential sources of bias and ensuring that the results were consistent across different analytical scenarios.

2.9. Reporting Bias Assessment

In conducting our systematic review on the application and competitive advantages of data mining and business intelligence in SMEs, it was crucial to assess the risk of bias due to potentially missing results, particularly those arising from reporting biases such as selective publication or selective reporting of outcomes. We recognized that these biases could significantly impact the validity and reliability of our synthesis, and thus, we employed a thorough and methodical approach to address this concern. Our assessment of reporting bias was conducted using a combination of well-established statistical and graphical methods. We opted for the use of contour-enhanced funnel plots, a powerful visual tool that allowed us to detect asymmetries in the data. These plots were carefully inspected to identify any potential publication bias by highlighting areas where studies might be missing due to bias versus those missing due to chance. The inclusion of statistical significance contours provided us with a clear and intuitive way to differentiate between these two scenarios, offering a robust visual representation of potential biases.

For this assessment, we chose not to develop new tools but rather relied on standard, proven techniques documented extensively in the literature. The methodological rigor of the tools we used was integral to our process. Contour-enhanced funnel plots, in particular, provided a straightforward yet effective method to visually assess the distribution of studies, allowing us to identify and account for potential biases in our synthesis. The assessment process was designed to minimize subjective bias, ensuring the integrity of our findings. Multiple independent reviewers were involved in evaluating the studies, and any discrepancies between their assessments were resolved through consensus discussions or, when necessary, by consulting a methodological expert. This collaborative approach ensured that the interpretation of results was balanced and unbiased. We intentionally did not use automation tools for assessing reporting bias in this review. Instead, we opted for a manual approach, utilizing tools such as Excel for creating charts and plots. This hands-on method allowed us to carefully analyze and visualize the data, ensuring a detailed and thorough examination. By manually inspecting the data, we ensured that no subtle patterns or potential biases were overlooked.

To further validate our findings, we conducted comprehensive manual searches across multiple online repositories, including Google Scholar, Scopus, and Web of Science. This approach enabled us to cross-reference data from different studies and sources, addressing any discrepancies and reinforcing the robustness of our conclusions. These manual searches were critical in ensuring that our synthesis was based on the most complete and accurate data available. Given the unique context of data mining and business intelligence studies in SMEs, we adapted the standard methods for assessing reporting bias to fit this specific field. Business intelligence studies often exhibit different reporting patterns compared to medical or social sciences research, necessitating these adaptations to ensure relevance and accuracy. By tailoring our methods to align with the characteristics of the studies we reviewed, we ensured that our analysis was both contextually appropriate and methodologically sound. To promote transparency and replicability, all methods and approaches used in our assessment have been thoroughly documented and made publicly accessible in the supplementary materials of this review. This commitment to openness allows other researchers to replicate our analysis or build upon it in future studies, thereby contributing to the overall rigor and reliability of research in the field of data mining and business intelligence for SMEs.

2.10. Certainty Assessment

The reviewed literature was evaluated based on five quality assessment (QA) criteria to ensure rigor and relevance:

QA1

: The clarity and explicitness of the research aim.

QA1

QA2: The specification and transparency of data collection methods.

QA1

QA3: The clear definition and explanation of the data mining and business intelligence processes.

QA1

QA4: The application of a well-defined and appropriate research methodology.

QA1

QA5: The contribution of the research findings to the enhancement of existing literature on SMEs' performance.

The certainty assessment responses are rated on a scale from zero (0) to one (1). A 'No' response is assigned '0' points, a score of '0.5' is given if the criterion is 'Partially' met, and '1' point is assigned for a 'Yes' response. All five criteria are scored using this scale. Each piece of literature under review can receive a total score between 0 and 5 points. The results of the certainty assessment for the collected literature on the applications and competitive advantages of data mining and business intelligence in SMEs' performance are presented in Table 5.

To support the conclusions of this systematic review on the applications and competitive advantages of data mining and business intelligence in SMEs, we undertook a rigorous assessment of the certainty of the evidence. The strength and reliability of our findings depend on a systematic evaluation process, which we carried out using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) framework. GRADE is a globally recognized system that offers a comprehensive and transparent approach to assessing the quality of evidence, ensuring that the conclusions drawn are both credible and well founded. The certainty of the evidence across key outcomes was meticulously evaluated using several critical factors. First, we closely examined the precision of effect estimates by considering sample sizes and the width of confidence intervals in the studies. Narrow confidence intervals coupled with large sample sizes were indicative of a high level of certainty in the evidence, as they suggest more reliable and precise effect estimates. We also assessed the consistency of findings by comparing results across the included studies. High consistency where studies demonstrated similar effects contributed to greater certainty. Any observed heterogeneity was thoroughly analyzed to understand its sources and potential impact on the overall findings.

Furthermore, the potential for bias was evaluated using an adapted version of the Cochrane Risk of Bias tool. Studies with a low risk of bias were considered to contribute more significantly to the overall certainty of the evidence. We also judged directness based on the alignment of study populations, interventions, and outcomes with the research questions of this review. High directness strengthened the support for our conclusions, leading to greater confidence in the evidence. Based on these factors, the certainty of evidence was categorized as follows: High certainty was assigned when studies were consistent, precise, directly applicable, and exhibited a low risk of bias. Moderate certainty was applied when there were minor concerns about one factor, such as some inconsistency or a moderate risk of bias. Low certainty was given when significant concerns existed in multiple areas, including imprecision, inconsistency, or a high risk of bias. Very low certainty was assigned when critical issues were present across all factors, significantly undermining confidence in the results. To ensure the relevance of the GRADE approach to this review, we adapted it specifically for outcomes related to the enhancement of SME performance through data mining and business intelligence. Multiple independent reviewers assessed the certainty of evidence for each outcome. Disagreements were resolved through consensus discussions, ensuring a balanced and thorough evaluation. Additionally, where possible, we sought additional data or clarification from study authors to support the certainty assessments. The results of these assessments were summarized in a "Summary of Findings" table 5 below, providing a clear and concise overview of the confidence in the evidence for each key outcome. We used standard GRADE phrases to communicate the certainty of evidence, such as "BI tools probably improve decision-making efficiency" for moderate-certainty evidence, ensuring that our conclusions were both clear and appropriately conveyed.

3. Results

3.1. Study Selection

In our systematic review, we initially searched three online databases namely, Google Scholar, Web of Science, and Scopus. From Google Scholar, we retrieved 6,550 records; from Web of Science, we obtained 207 records; and from Scopus, we gathered 854 records. In total, this search yielded 7,611 records. Following the removal of duplicate entries, 1,092 unique records remained. We then conducted a preliminary screening of these records to assess their relevance to our review criteria. This screening process resulted in the selection of 34 full-text documents for a more detailed evaluation. After reviewing these full-text documents, we included 93 papers in our final review. These included 64 journal articles, 4 book chapters, 19 conference papers, and 6 dissertations. The process is illustrated in the Figure 6, which outlines the flow of records through each stage of the review. Additionally, Figure 7 provides a breakdown of the number of records obtained from each database. This detailed description ensures transparency and aids in the replication of the search process by other researchers.

3.2. Study Characteristics

A total of 93 studies on data mining and business intelligence (BI) in SMEs were identified, spanning from 2014 to 2024. The distribution of these studies includes 4.30% book chapters, 20.43% conference papers, 6.45% dissertations, and 68.82% journal articles. The annual publication trend depicted in Figure 8 shows a fluctuating but generally upward trajectory, peaking in 2023 with 18 studies. This peak is consistent with the detailed breakdown in Table 6, underscoring a significant increase in research activity related to BI and data mining for SMEs.

2023 marked a notable spike in research outputs, as shown by a variety of publication types including journal articles, book chapters, and conference papers, reflecting the heightened interest and advances in data-driven strategies for SME enhancement. The distribution of these publications across different types (Figure 9) reveals that journal articles constitute the majority at 68.82%, followed by conference papers, dissertations, and book chapters.

This surge in scholarly activity aligns with ongoing advancements in BI tools and data mining technologies, which are crucial for improving decision-making and operational efficiencies within SMEs. The progressive increase in publications from 2014 to 2024 highlights the growing academic and practical focus on leveraging BI systems to bolster SME performance, emphasizing the role of these technologies in transforming business practices for smaller enterprises.

Table 7 summarizes the impact of data mining and business intelligence (BI) on SMEs across various outcomes. The data indicates that BI tools and data mining generally have a positive effect on SME performance. For operational efficiency, the evidence is moderate, showing a mean improvement of 12% in process completion times, suggesting BI tools enhance efficiency by reducing process times and errors. Financial performance benefits are strongly supported, with an average 15% increase in revenue due to BI adoption, indicating significant financial gains. For strategic decision-making, a moderate level of evidence with a risk ratio of 1.8 suggests that BI tools likely improve decision-making quality by making it faster and more accurate. Data mining shows a moderate certainty with a 10% increase in customer retention rates, reflecting its effectiveness in enhancing customer relationship management and personalized marketing. Market trend forecasting has lower evidence with a hazard ratio of 1.5, indicating that while data mining aids in predicting and adapting to market trends, results may vary. In terms of innovation and product development, moderate evidence shows an 8% improvement in new product success rates, suggesting BI tools support innovation. Lastly, strong evidence points to a 20% improvement in risk management outcomes with BI tools, underscoring their role in enhancing resilience in volatile markets.

Table 8 provides an overview of various studies on data mining and business intelligence (BI) in SMEs. These studies focus on understanding how BI systems are adopted and implemented in SMEs, the role of data mining in enhancing decision-making and competitive advantage, and the impact of employee training on business performance. Methodologies commonly used include literature reviews, case studies, surveys, and quantitative analyses like Structural Equation Modeling (SEM). Key outcomes show that BI tools and data mining significantly improve operational efficiency, financial performance, and strategic decision-making. Challenges identified across these studies include high costs, complexity of systems, and lack of skilled personnel. Recommendations often highlight the need for simplified BI tools, increased training, and better integration of these technologies into business processes. These insights collectively aim to guide SMEs in effectively leveraging BI and data mining to boost their performance and competitiveness.

3.3. Risk of Bias in Studies

The risk of bias for each included study was systematically evaluated using the Newcastle-Ottawa Scale (NOS), as shown in Table 9. The quality assessment table categorizes studies based on the number of stars awarded in three key domains: Selection, Comparability, and Outcome/Exposure. Studies scoring 7-9 stars were rated as high quality, while those scoring 4-6 stars were rated as moderate quality. No studies in our review were rated as low quality (0-3 stars).

Table 9 presents the detailed assessment of each study, with most studies achieving high-quality ratings. For instance, Study 1 received 8 stars, indicating a high level of quality in all domains, while Study 5 received 6 stars, categorized as moderate quality. The table clearly outlines the domains where studies were strong or weak, providing transparency regarding potential biases.

This risk of bias assessment ensured that only studies with robust methodologies were included in our review, thus increasing confidence in the overall findings. Where proprietary tools or insufficient data created uncertainties, additional steps were taken to verify the study’s validity through external cross-referencing, ensuring that our assessments remained accurate and objective.

Figure 10 shows the distribution of research designs used in various studies, categorized into multiple types. The most commonly used design is the Survey, which accounts for 31 studies. This is followed by Case Studies, representing 24 instances. Experimental designs were used in 8 studies, while Descriptive Research appears in 4 studies. A significant portion, 15 studies, did not specify their research design. Other less common designs include Bibliometric Analysis, Quasi-Experimental, Thematic Analysis, Theoretical Discussion, Design Science Research, Content Analysis, Exploratory, and Conceptual Analysis, each with fewer than 4 occurrences. The total number of studies covered is 93.

Figure 11 shows the different methods used for collecting data to assess the risk of bias in studies on data mining and business intelligence in SMEs. Surveys were the most common method, used in 47.31% of the studies, indicating their popularity for gathering broad data. However, surveys alone may introduce biases if the questions or samples are not carefully designed. Document Analysis was the second most common method, appearing in 33.33% of studies. This method involves reviewing existing documents and literature, which can provide valuable context but may not include the most current data.

Other methods included Interviews, used in 10.75% of studies, and Observations, used in 3.23% of studies. Combining methods, such as surveys with document analysis or interviews, helps address the limitations of each individual approach and can reduce overall bias. The TOE framework was used in 1.08% of studies, guiding the analysis rather than serving as a data collection method. In some cases, the method was not specified, which accounted for 4.3% of the total. The variety of methods used highlights the efforts to balance different perspectives but also points to potential gaps in ensuring a thorough assessment of bias.

3.4. Results of Individual Studies

Figure 12 presents an overview of the distribution of data mining techniques across various studies, revealing a wide range of methodologies. The most frequently reported technique is clustering, used in 10 studies, followed closely by classification, appearing in 9 studies. Data mining is applied in 5 studies, while other techniques, such as association, bibliometric analysis, and topic modeling, are each used in a single study. Notably, a significant portion of the studies (62 out of 93) did not specify the data mining technique used. This highlights the diversity of methods in the field, with clustering and classification standing out as the most commonly used approaches.

Figure 13 presents the distribution of business intelligence (BI) tools used across the studies. Dashboards are the most frequently reported tool, appearing in 29 studies. A notable number of studies (46) did not specify the BI tools used. OLAP systems are mentioned in 10 studies, while tools such as Cloud BI Tools, reporting tools, social media analytics, and Microsoft SQL Server are each mentioned in 2 studies. This updated data highlights the widespread use of dashboards and the frequent lack of specification regarding BI tools in business intelligence research.

3.5. Results of Synthesis

In this section, we summarize the results of the syntheses conducted across various studies. First, we examine the characteristics and potential biases of the contributing studies, revealing a predominant focus on the manufacturing sector, followed by the mining industry, and highlighting the broad application of data mining and business intelligence (BI) in small and medium-sized enterprises (SMEs). Next, we present the results of statistical syntheses, including summary estimates, precision, and measures of statistical heterogeneity, emphasizing the global interest and diverse research formats related to BI and SME performance. We also explore the economic context of the studies, noting the significant focus on developing countries, and investigate the reasons for any observed heterogeneity. Finally, we discuss sensitivity analyses, highlighting the prevalence of cloud-based technology implementation models among the studies, which reflects current trends in scalable and adaptable data management solutions for SMEs.

3.5.1. Characteristics and Risk of Bias Among Contributing Studies

In this sub-subsection, Figure 14 illustrates the focus of titles by industry context. The research predominantly focuses on the manufacturing sector, comprising 42 of the studies, followed by the mining industry with 21. These findings emphasize the importance of data mining and business intelligence (BI) in traditional industries where operational efficiency and performance optimization are critical for SMEs. The remaining studies are distributed among various sectors such as ICT, pharmaceuticals, and financial industries, each representing the diverse applications of BI technologies across different business environments. This distribution reflects the broad applicability of data mining and BI in enhancing SME performance across various sectors.

3.5.2. Results of Statistical Syntheses

The research comes from a broad range of countries, with significant contributions from the UK (9.68%), Indonesia (5.38%), South Africa (5.38%), and Malaysia (5.38%), as shown in Figure 15. This global distribution underscores a widespread interest in how small and medium-sized enterprises (SMEs) can use data mining and business intelligence (BI) to enhance performance. Countries such as the USA and China also played key roles in the research, reflecting their leadership in BI and data-driven technologies.

In terms of research type, the majority of studies were journal articles, totaling 64 out of the 93 studies. Additionally, there were 19 conference papers, 4 book chapters, and 6 dissertations, reflecting a diverse array of research formats. Figure 15 illustrates the breakdown of these research types, showing that journal articles dominate, followed by conference papers. The prevalence of journal articles highlights the importance of peer-reviewed research in documenting the impact of data mining and BI on SMEs, while conference papers suggest active discussions and advancements in the field. Dissertations and book chapters further contribute to in-depth exploration and consolidation of knowledge.

3.5.3. Investigation of Heterogeneity

Figure 16 depicts the economic context of the studies, distinguishing between developed and developing countries. The data reveals that 67 of the studies were conducted in developing economies, underscoring a strong focus on emerging markets where SMEs are pivotal to economic growth. This dominance highlights the critical role of data mining and BI in driving competitiveness and operational efficiency among SMEs in resource-constrained environments. Meanwhile, 26 of the studies originate from developed countries, where the emphasis is often on innovation and advanced technological integration. This variation underscores the heterogeneity in BI applications, driven by the distinct challenges and opportunities present in different economic contexts.

3.5.4. Sensitivity Analyses Results

Figure 17 examines the focus on technology implementation models in the reviewed studies. The updated findings reveal that cloud-based models are used in 32.26% of the studies, while hybrid models account for only 1.08%. A significant portion of the studies, 60.22%, did not specify the implementation model, and on-premises solutions are used in 6.45%. The shift towards cloud-based solutions remains evident, as they offer flexible and cost-effective data management options, well-suited to the dynamic needs of SMEs. This trend highlights the advantages of cloud technology in providing real-time data access and improved collaboration.

3.6. Reporting Biases

Figure 18 reviews the focus on sample characteristics in the analyzed studies. A significant number of studies, 70 in total, focused on SMEs, supporting the overall theme of improving SME performance through business intelligence (BI) and data mining. The remaining studies, which include various other business sectors such as business analysis, consulting, and manufacturing firms, offer comparative insights but still highlight the benefits of BI technologies for streamlining operations and improving decision-making, particularly for SMEs. The strong focus on SMEs in these studies emphasizes the need for customized BI solutions that address the specific challenges and limited resources typical of smaller enterprises.resources typical of smaller enterprises.

3.7. Certainty of Evidence

Figure 19 presents an analysis of IT performance metrics in the reviewed studies. The updated data shows that data processing speed is the most frequently evaluated metric, appearing in 19.35% of the studies. However, most studies (72.04%) did not specify particular performance metrics. Accuracy was considered in 3.23% of the studies, while other metrics such as competitive advantage, scalability, data quality, and data visualization were each mentioned in a small percentage, all below 3.5%. This suggests that while some aspects of IT performance are emphasized, there is a lack of consistent focus on specific metrics across studies examining BI in SMEs.

4. Discussion

The literature reviewed, as guided by the PRISMA flowchart, provides answers to the five research questions outlined in Section 1.1. These questions explore how data mining and business intelligence (BI) techniques can improve the performance and competitive edge of small and medium-sized enterprises (SMEs). The following discussion delves into the insights gained from the literature in response to each research question.

• RQ1: How Do Algorithms of Association as a Method of Data Mining Contribute to Enhancing Business Intelligence in Organizations?

Association algorithms, such as Apriori and FP-Growth, are powerful tools in data mining that identify relationships between variables in large datasets. By discovering patterns, like frequently purchased product combinations, SMEs can optimize their inventory management and improve cross-selling strategies. For example, if data shows that customers who buy item A often purchase item B, businesses can bundle these products in promotions, leading to increased sales and customer satisfaction.

Moreover, these algorithms help businesses make data-driven decisions by highlighting significant associations that may not be immediately obvious. This capability enables SMEs to adjust their marketing campaigns based on identified patterns and trends, ensuring that their strategies are more effective and aligned with customer preferences. The ability to extract actionable insights from data supports more precise targeting of promotions and enhances overall business performance.

• RQ2: What Are the Impacts and Roles of Cloud-Based Big Data Analytics in Knowledge Management for Achieving Competitive Advantages in Organizations?

Cloud-based big data analytics offer SMEs a flexible and cost-effective solution for handling large volumes of data. These platforms provide access to advanced analytics tools without the need for substantial upfront investments in infrastructure. As a result, SMEs can analyze large datasets to uncover insights about customer behavior, market trends, and operational efficiencies, which are crucial for strategic decision-making. This agility in accessing and analyzing data allows SMEs to respond more rapidly to market changes and opportunities.

Furthermore, cloud-based solutions support real-time data processing, which is vital for timely decision-making. By leveraging these platforms, SMEs can continuously monitor and adjust their strategies based on the most current information available. This capability enhances their ability to maintain a competitive edge, adapt to changing market conditions, and make informed decisions that drive business growth.

• RQ3: How Can Pervasive Business Intelligence Systems be Used to Gain and Sustain Com-Petitive Advantages in Organizations?

Pervasive business intelligence systems integrate BI tools across all organizational levels, making data and insights accessible to employees throughout the company. This widespread access fosters a data-driven culture, where decision-making is informed by real-time information rather than intuition. Employees can use BI tools to analyze performance metrics, track progress towards goals, and identify areas for improvement, which enhances overall productivity and efficiency.

Additionally, pervasive BI systems help align organizational efforts with strategic objectives by ensuring that all employees are working with the same set of data and insights. This alignment promotes consistency in decision-making and helps SMEs respond more effectively to operational challenges. By embedding BI tools into daily workflows, organizations can achieve greater agility and maintain a competitive advantage through informed, data-driven decisions.

• RQ4: What Opportunities and Challenges Do Organizations Face in the Implementation of Predictive Analytics for Competitive Advantage, and How Can These Be Effectively Addressed?

Predictive analytics offers SMEs significant opportunities to improve decision-making through forecasting and risk management. By analyzing historical data, SMEs can predict future trends, identify potential risks, and optimize strategies for customer segmentation and marketing. For example, predictive models can forecast sales trends, enabling businesses to adjust their inventory and marketing strategies proactively, thus enhancing their competitive edge.

However, implementing predictive analytics also presents challenges, including high costs, data quality issues, and a shortage of skilled personnel. SMEs can overcome these challenges by investing in training programs to build in-house analytics capabilities, selecting affordable and user-friendly predictive analytics tools, and establishing robust data governance practices. Collaborations with technology providers or academic institutions can also offer the expertise and resources needed for successful implementation.

• RQ5: How Does Sentiment Analysis, as an Approach to Business Intelligence, Enhance De-Cision-Making and Operational Efficiency in Organizations?

Sentiment analysis uses natural language processing to evaluate customer opinions from sources such as social media, reviews, and feedback forms. This analysis helps SMEs understand customer sentiments and trends, enabling them to tailor their products, services, and marketing strategies more effectively. By interpreting customer feedback, businesses can make informed decisions about product improvements, customer service enhancements, and targeted marketing campaigns.

Additionally, sentiment analysis aids in crisis management by identifying potential issues and negative sentiment early. This early detection allows SMEs to address problems promptly, manage their brand reputation, and respond to customer concerns before they escalate. By leveraging sentiment analysis, organizations can improve their operational efficiency, enhance customer engagement, and make data-driven decisions that contribute to their overall success.

5. Conclusions

This systematic review has highlighted the significant role that data mining and business intelligence (BI) tools play in enhancing the performance and competitiveness of small and medium-sized enterprises (SMEs). The findings indicate that these technologies provide SMEs with crucial capabilities, such as improved decision-making, operational efficiency, customer relationship management, and financial performance. By leveraging data-driven insights, SMEs are better positioned to navigate the challenges of rapidly changing markets and maintain a competitive edge. However, despite the evident benefits, the full potential of data mining and BI remains underutilized in many SMEs, especially those in resource-constrained environments. The review identifies several barriers, including the complexity of implementation, high costs, and the lack of skilled personnel, which hinder the widespread adoption of these technologies. Overcoming these challenges requires more accessible BI tools, targeted training programs, and supportive policies that promote the integration of data-driven technologies into SME operations. Future research should focus on addressing these gaps by exploring more industry-specific applications, examining the long-term impact of BI and data mining on SME growth, and developing frameworks that make these technologies more affordable and user-friendly for smaller enterprises. Ultimately, enhancing SME performance through data mining and BI is not just about adopting the latest technologies; it is about creating an ecosystem that supports sustainable growth and innovation in the SME sector.