The term “cloud computing” was first used in 1996, but it wasn’t until the early 2000s when Google and Amazon began offering cloud services that it gained widespread popularity ([45]). Since then, its use has steadily grown, with a significant increase during the COVID-19 pandemic. According to recent surveys, over 90% of organizations now use cloud computing in their operations ([32]; [53]). The global cloud computing market is projected to continuously grow and is expected to reach $1.5 trillion by 2030 ([53]). Cloud computing generally falls into three categories: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service (SaaS). SaaS is the most commonly used model, as corporations grow more comfortable with running portions of their business in a controlled cloud environment ([36]). A well-known SaaS application is Cloud Enterprise Resource Planning (CERP).

Organizations can streamline their business processes and data management by shifting their mission-critical ERP to the cloud. However, the benefits of this approach have been difficult to achieve. In fact, a recent PwC survey revealed that only 10 percent of respondents noted significant business benefits in their workplaces ([44]). As a result, some organizations are contemplating returning some of their data and applications to traditional systems ([31]). Similar concerns were raised by other researchers. For example, [4] point out the significant impact, both positive and negative, of shifting traditional ERP systems to cloud-based solutions. [41] explores current trends of cloud repatriation, providing insight into why some organizations revert to traditional ERP solutions despite potential cloud benefits. The search for successful factors in utilizing and benefiting from CERP persists.

The purpose of this research is to create and evaluate a model for CERP success. Whereas many CERP success models and guidelines have been proposed ([3]; [6]), little empirical evidence exists that offer convincing arguments for migration to CERP. The focus of this research is to quantitatively analyze available empirical evidence. The purpose is to generate evidence-based insights to aid companies to materialize tangible business benefits from their investment in cloud computing in general and CERP in particular. Specifically, the model aims to identify and validate the relationship between a set of independent variables (the critical success factors) and a set of dependent variables (CERP success) that can guide future development of CERP in practice and inspire further research on this critical topic.

A systematic review of the literature is conducted to provide answers to the two research questions. Unlike past reviews, this research focuses on a quantitative analysis of the variables involved using meta-analysis. Meta-analysis is a set of tools used for knowledge synthesis, theory development, and testing ([10]; [29]; [47]). It has been applied to topics in various disciplines including education, business, psychology, medicine, and information systems. [9] conducted a meta-analysis of ERP implementation studies. However, their study is qualitative and does not generate any effect size measures related to CERP success. The current research is the first attempt to gain a better understanding of the relationship between success factors and CERP success through meta-analysis.

To collect relevant studies for the meta-analysis, the Web of Science database was searched with the phrase: “business benefits of cloud ERP”, which returned 39 articles. A second search was conducted using the phrase: “critical success factors for cloud ERP,” which returned 24 articles. Similar searches using Scopus returned 83 and 45 articles, respectively. Additional articles were identified through searches using Science Direct, Semantic Scholar, Google Scholar and AI research tool Elicit (https://elicit.com/). Finally, citing mining was used to locate additional studies from the initial lists of articles. In all, close to 200 articles were identified.

The titles and abstracts of the sample articles were next scanned to assess their relevance to the research questions. Studies that specifically focused on the applications, benefits, and challenges of CERP were shortlisted for further review. Full-text articles of shortlisted studies were then scrutinized in detail to further assess their relevance. Studies were excluded for the following reasons:

Conceptual papers (e.g., [4])Qualitative studies (e.g., [38])Review papers (e.g., [8]; [9])Empirical studies without usable effect sizes (e.g., [5]; [7]; [15]; [18]; [25]; [39]; [48]; [50]; [54])

Figure A1in the Appendix depicts a PRISMA-style flow diagram of study identification and selection. The CSF dataset consisted of 65 effect sizes representing relationships between critical success factors and cloud ERP success extracted from 10 independent primary studies. The Success dataset consisted of 12 effect sizes measuring relationships between cloud ERP success and organizational outcomes extracted from 8 independent primary studies. In this article, k denotes the number of extracted effect sizes, not the number of statistically independent studies or samples. The number of independent studies is reported separately in the “Studies” column of the results tables. Because several primary studies contributed multiple correlations from the same sample, summing the per-effect sample sizes would overstate the number of unique participants. Therefore, sample size information is used only for computing sampling variances and inverse-variance weights, whereas independent-study counts are used to describe clustering and dependence in the robust-variance models. For each study, the reported statistics are coded and converted into correlations, if necessary, using commonly accepted formulas ([10]), as shown in AppendixTable A1. The coding sheet is available at https://figshare.com/articles/dataset/CERP_dataset/33954475?file=69183988.

Because multiple effect sizes were frequently extracted from the same primary study, the assumption of statistical independence required by conventional meta-analysis was invalid. To address this issue, the study identifier was treated as the clustering variable and Robust Variance Estimation (RVE) was employed, so that standard errors and tests were adjusted for dependence among effect sizes originating from the same study. Correlations were transformed into Fisher’s z values and weighted by the inverse of their sampling variances. Thus, the analyses use all available effect sizes while avoiding the interpretation that 65 CSF effects or 12 success effects represent 65 or 12 independent samples. In addition to estimating overall effects, four moderator analyses were performed as outlined in Table 3 below.

Table 3. Moderator models.

The CSF correlations were tested for moderator effect based on the Technology-Organization-Environment (TOE) framework. The success correlations were tested for moderator effect based on three levels: financial/market, user/service, or organizational performance. Construct-level analyses were also conducted for determinants that appeared in at least two independent studies. Finally, study-level moderators were tested based on three study characteristics: developed vs. developing countries, SME vs. non-SME organizations, and study dated pre-Covid vs. post-Covid. The rationale is that, compared with developing countries, developed countries tend to have better technology infrastructure and hence more likely to reap the benefits of CERP. SMEs are the main targets of CERP vendors and hence are more likely to report larger benefits than non-SMEs. Finally, cloud usage jumped during Covid and should result in more CERP benefits than pre-Covid.

Robustness was assessed using Leave-One-Out diagnostics, while publication bias was evaluated with Egger’s p-value and funnel plots. All analyses were conducted in R.

The intercept-only RVE model revealed a significant positive relationship between critical success factors and cloud ERP success. The pooled mean effect was 0.33 (p < 0.01), based on 65 effect sizes clustered within 10 independent primary studies. This finding indicates that critical success factors collectively exert a small-to-moderate positive influence on cloud ERP implementation success. Similarly, the meta-analysis of organizational outcomes indicated that cloud ERP success is positively associated with organizational performance measures such as financial performance, operational performance, market performance, service quality, goal accomplishment, and net benefits. The pooled effect for the outcome model was 0.42 (p < 0.01), based on 12 effect sizes clustered within 8 independent primary studies. Table 4 displays the pooled effects and heterogeneity statistics; the k column reports the number of extracted effect sizes, whereas the Studies column reports the number of independent study clusters used in the RVE models.

Table 4. Overall pooled effects (Fisher’s z, back-transformed r).

Note: k refers to the number of extracted effect sizes. “Studies” refers to the number of independent primary-study clusters used for robust variance estimation. Sample sizes were used to compute sampling variances and weights but were not summed as unique participant counts because several studies contributed multiple dependent effect sizes from the same sample.

These findings indicate that cloud ERP success is influenced by a broad set of critical success factors which, in turn, generates meaningful business value. High heterogeneity indices suggest that moderator analyses are warranted. Results are presented in Table 5 and Table 6.

Table 5. CSF moderators.

Table 6. Success moderators.

Table 5 shows that the pooled effects within each TOE category were positive. However, the omnibus small-sample-corrected HTZ test was nonsignificant (F = 1.01, p = .615), indicating that TOE category did not statistically moderate the CSF relationship. Therefore, the category-level differences should be interpreted as descriptive patterns rather than statistically confirmed differences. The omnibus tests for the other moderators in Table 5 were also nonsignificant.

Table 6 showed that organizational and user/service performance had a larger effect size than that of financial performance (0.57 and 0.515 vs. 0.197). However, the omnibus test was nonsignificant (F = 2.34, p = 0.356), suggesting that outcome type did not moderate the success relationship. None of the other moderators in Table 6 had a significant omnibus test result either. These findings may be due to low statistical power from applying the test to a small sample. Thus, the absence of significant moderation is better interpreted as inconclusive rather than as evidence of no moderation. As more data from future primary research become available, the relationships can become clearer.

To further explore the effects of individual CSFs, a construct-level RVE moderator analysis was performed on constructs investigated in two or more studies. Several observations can be derived from the results shown in Table 7:

Innovation constructs dominate the top (Switching Benefits, Relative Advantage) but also include the only negative one (Switching Costs).Technology constructs are mixed (Compatibility high, Complexity null).Organization constructs are consistently moderate (Top Management Support, Organizational Culture).Environment constructs are weak (Vendor Support moderate, Competitive Pressure weak).

These findings are descriptive and suggest promising hypotheses that can be tested in future research.

As cloud computing becomes increasingly popular, companies have migrated their mission-critical ERP systems to the cloud by leaps and bounds in the last decade. Enticed by lower costs and higher efficiency, CERP has found widespread adoption especially among small and medium businesses. While academic literature often highlights the theoretical advantages of CERP, industry surveys present a more nuanced perspective. For instance, a study by Epicor revealed that although 96% of organizations felt confident during the initial stages of CERP adoption, 48% expressed a need for greater support from venders during implementation and beyond ([14]). Lack of vendor support and other challenges can result in disillusionment with materializing real business benefits of CERP ([44]). It is critical to examine empirical evidence to gain a better understanding of CERP’s accrued business value.

The purpose of this research is twofold. The first is to quantify the business benefits of CERP and the second is to quantify the effect of factors that contribute to the successful implementation of CERP. After a systematic review of the literature, empirical data from 15 primary studies were meta-analyzed. Results showed that the adoption of CERP is, on average, likely to yield moderate positive business benefits. This is true despite inconsistent results reported in different studies. For example, [17] reported CERP has a positive effect on market performance (r = 0.70); however, [16] reported a negative effect (r = -0.13). The significance test result suggests that any negative effect is probably a result of random chance rather than a manifestation of the true relationship.

Whereas the evidence suggests that CERP’s business benefits are likely significant statistically, it does not mean that the benefits are substantial in real terms. Table 6 reveals that largest success measures examined in prior empirical studies are organization or user-oriented variables, e.g., strategic constituencies satisfaction or goal accomplishment ([24]), net benefit ([40]). As a result, managers should broaden their evaluation criteria beyond traditional financial metrics when assessing cloud ERP investments. As shown in Table 1, other categories of CERP success exist, and they should be studied as well to demonstrate the value of CERP. Prior work on cloud-based ERP business value similarly emphasizes that SMEs may realize value through process integration, accessibility, and operational improvements that extend beyond immediate financial indicators ([21]). Future research can collect empirical data on the people aspects (communication and collaboration), the system aspects (ease of use, ease of maintenance, etc.) or customer aspects (customer satisfaction). This type of research can lead to discovery of larger effect sizes or business benefits and hence provide greater justification for continued investment in CERP.

In addition, researchers should statistically test the effect of any measures they investigate. It is not useful to measure some variables but report only raw scores or the means and standard deviations as was done in several studies (e.g., [15]; [48]). A mean of 3.5 (out of 5 points) is not an objective measure of, say, increased market performance. An effect size, e.g., a correlation coefficient of 0.60, offers a concrete measure of the underlying relationship between CERP and its benefits.

The second objective of this research is to test the effect of factors contributing to successful implementation of CERP. Our results confirm the significant effects of many variables proposed in prior studies. Similar to the effect of CERP success, the size of the effect of the critical success factors is modest. More research is needed to have a clearer understanding of the effect of individual factors and their interactions ([38]). It would be insightful to compare the effects of various categories—such as organizational, people, and technology factors—on implementation success to determine which has the greatest impact and understand the underlying reasons ([39]). A popular way to rank the importance of the CSFs of CERP is to count the number of times each factor has been investigated (e.g., see [7]; [20]; [51]).

The present study formally tested category differences using a TOE-based robust variance estimation meta-regression. Innovation-related factors showed the strongest descriptive average relationship with cloud ERP success, followed by organizational factors, technical factors, and environmental factors ([2]; [11]). However, these differences were not statistically significant after cluster-robust adjustment and should therefore be interpreted as exploratory patterns rather than confirmed category effects. At the construct level, innovation factors such as switching benefits and relative advantage were strongly associated with cloud ERP success, while traditional organization factors including top management support and organizational culture also showed moderate positive associations. Compatibility and vendor support also emerged as notable descriptive predictors. These exploratory findings provide directions for future researchers in developing and testing theories.

For practitioners, these findings indicate that successful cloud ERP implementations depend not only on internal managerial support but also on external ecosystem conditions and the perceived strategic advantages of cloud-based ERP systems. At the same time, while one of the benefits of cloud computing is streamlined in-house IT operations, managers should obtain vendor support without abdicating strategic planning and cloud controlling. The cloud-powered companies develop an enterprise-wide data strategy, optimize data governance, risk, and control framework, and work closely with service providers to integrate their CERP with other systems ([44]). These are the hallmarks of innovative organizations that have achieved measurable business value.

This research systematically examined empirical evidence regarding the accrued business value in Cloud-based Enterprise Resource Planning (CERP) systems. While the adoption of cloud ERP has increased rapidly, driven by potential cost savings, scalability, and operational improvements, achieving consistent and tangible business benefits remains challenging. The meta-analysis confirms significant positive outcomes from the use of CERP, although the magnitude of these benefits varies widely. Table 8 summarizes the major findings.

Critical success factors—categorized as organizational, people-related, technological, and environmental—have a similar moderate impact on successful CERP implementation. Factors such as top management support, relative advantage, robust data security, and vendor support have consistently been touted as critical. However, other technological factors such as technology compatibility may be equally important in materializing the full potential of CERP investments ([52]). Practitioners should pay careful attention to these confirmed critical

Table 8.Research findings summary.

success factors to maximize the likelihood of successful outcomes. More empirical research along this line is encouraged. Researchers are also urged to provide comprehensive statistical reporting in their studies to facilitate knowledge accumulation and enable effective synthesis through meta-analysis. Additionally, future empirical studies should broaden their focus beyond firm-level financial performance to include more diverse success measures, such as employee satisfaction, customer retention, and system usability. Such research can provide more comprehensive insights and assist businesses in harnessing the full value of their cloud ERP investments.

Figure A1.PRISMA-style study identification and selection flow.

Table A1.Meta-analysis coding.

Figure A2.Funnel plot: CSFs → CERP success.

Figure A3. Funnel plot: CERP success → outcomes.