2.1 I4.0 Adoption

Tanzania manufacturing industries (TMIs) structure, including assembling (4%), processing (43%), and MIs (53%), faces challenges like complex legal structures, lack of modern technology, skilled personnel shortage, and inadequate research and design centers [21]. Despite embracing Industry 1.0, 2.0, and 3.0 stages, opportunities exist for I4.0, as suggested by Ngowi [20], Ntamulyango [22], and Nhelekwa et al. [23]. The fourth IR is defined as the connection between autonomous, independent devices that can communicate in real-time and possibly cooperate in a smart environment to make decisions and take actions based on the data gathered, following the success of the third IR [24, 25]. There are 11 fundamental components of I4.0, according to Khin and Kee [15] and Haiston [26], including BDA, collaborative robots, machine-to-machine communication, cybersecurity, VR, cloud computing, cyber-physical security, the IoT, simulation, horizontal and vertical integration, and supply chain. The integration of humans and machines has significantly improved productivity, efficiency, and human life quality, enhancing product quality, time efficiency, and stakeholder cooperation, supported by scientific studies [27].

I4.0 is a relatively recent idea that has become a hot topic; there is not a single conception of its adoption in the literature [20]. Nonetheless, the term “I4.0”, which refers to the fusion of cutting-edge industrial technology and information and communication technologies (ICTs), is commonly used in academic contexts. The IoT, cloud computing, and cyber-physical systems are just a few of the technologies included in this area of I4.0. Also, one of I4.0's key characteristics is the capacity to integrate equipment with production systems in a smart and autonomous environment. This makes it possible for the components to communicate in real-time and for the manufacturing system to receive feedback [25].

2.2 Technology Adoption Theories

2.3 I4.0 Models

The I4.0 Model builds on the outline of the I4.0 Reference Architecture Model (RAMI 4.0) model. The creation of the RAMI 4.0 was part of Germany's Platform I4.0 program (RAMI 4.0) [32]. As I4.0 became more well-known throughout Europe and beyond, the need for a clear and uniform nomenclature increased. The goal of RAMI 4.0 was to create a common framework for communications and ideas among nations. The structure of RAMI 4.0 is an idea map that depicts manufacturing processes and production products in a straightforward and organized manner to foster a shared knowledge of I4.0. RAMI 4.0 ensures that participants participating in I4.0 conversations can communicate effectively [33]. This model defines each production object and includes the data and functions that go with it. The output is a full virtual description of the object. A three-dimensional framework known as the value stream and life cycle, hierarchy levels, and layers serves as the foundation for RAMI 4.0.

2.4 Empirical Literature Review

2.5 Research Gap

MIs are baffled by the I4.0 concept. Khin and Kee [15] and Nimawat and Das Gidwani [34] investigated the adoption of important I4.0 technologies. These technologies include organizational learning, virtual reality, and radio-frequency identification (RFID). The authors examined the market structure, the supply chain, management leadership, and implementation challenges for implementing RFID as organizational, environmental, and organizational learning capacity for implementing additive manufacturing. Studies have been done to determine the essential elements of implementing I4.0 for the Manufacturing Industry, especially in emerging nations. However, studies conducted by academics in industrialized countries to assist the quick spread of the idea among MIs were highly accepted and judged significant [25]. Adoption of I4.0 in a developing economy is still difficult. I4.0 is necessary for Tanzania's MIs to compete in the international market. There have been few studies on the adoption of I4.0, which is primarily aimed at MIs in developing nations like Tanzania. Given this obvious information gap, the research aims to scrutinize the challenges and strategies influencing negatively and positively, respectively, I4.0 adoption in Tanzanian MIs, particularly in the FBMIs.

2.6 Conceptual Framework

A conceptual framework depicts the conceptual status and connections between the items under study (Figure 1). It is an interrelated system of assumptions, expectations, and beliefs [37]. It depicts the theoretical connection between the variables under investigation by connecting the study's notions of branding practices, identifying gaps in the state of knowledge regarding the phenomena or subject, and outlining the methodological underpinnings of the research project.

3.1 Research Design and Study Area

This study deployed a cross-sectional research approach. A cross-sectional research design is an empirical study that examines a sensation within its real-life context using multiple sources of evidence and examining the boundaries between phenomena [38]. This form of study design is justified because it permits the use of various sources of evidence and, occasionally, triangulation techniques. This study was conducted in medium and large MIs producing food and beverages in Dar es Salaam and Pwani regions. This study was conducted in FBMIs in Dar es Salaam and Pwani, Tanzania's most industrialized region. Forty-two percent of Tanzania's manufacturing comprises food, tobacco, and beverages [39]. Therefore, there are many FBMIs in Tanzania, resulting in a focus on those industries. Micro and small FBMIs have low technology and digitalization, making assessing their awareness of I4.0 difficult, as most are not yet in Industry 3.0 [40, 41]. Compared to large industries, small and micro industries lack advanced manufacturing technologies [40]. They focus on standard processes and low technology levels and operate with standard machines [40]. The food industry encompasses many processes, including farming, harvesting, processing, packaging, transportation, and distribution. Each process involves different technologies, equipment, and production methods, making generalizing I4.0 adoption across the entire industry challenging.

3.2 Study Population and Sample Size

The targeted population was medium and large FBMIs that manufacture food and beverages based in Dar es Salaam and Pwani regions. From CTI and DI [42] and Lwesya et al. [43], medium and large industries have at least 50–99 employees and at least 100 employees, respectively. They should also have above $85,000–$340,000 and above $340,000 for medium and large enterprises, respectively. Likewise, the government of Tanzania categorizes the business sizes as shown in Figure 2. Thus, those criteria were used to determine the industry size. Automation and technological levels were also considered. The study population was 129; however, upon considering the selection criteria, such as the level of automation for some basic machinery, the capital invested for machinery, and the nature of products, 103 industries were included in the study. The sample size was obtained through the purposive sampling procedures.

3.3 Data Collection Methods

A closed-ended questionnaire was developed with four sections: Part A (Demographic information and industry information), Part B (Awareness level questions), Part C (Challenges for implementing I4.0) and Part D (Strategies for implementing I4.0). Some questions were adapted from previous studies, including the study by Maganga [37], Özavcı [45], Ntamulyango [22], and Nimawat and Gidwani [46].

3.4 Validity and Reliability

The validity of the questionnaire was tested, first, through face validation as suggested by Yusoff et al. [47] and Roebianto et al. [48], and second, by using Pearson Product Moment Correlations using SPSS 23 from the Pearson product–moment correlation. If the significance value < 0.05, then the questionnaire should be declared valid; otherwise, it is not [49]. Reliability is the degree to which research findings are valid across time and fairly represent the studied population. Internal consistency must be established to guarantee validity before a test is used for study or examination [48]. So, the reliability test was performed by determining Cronbach's Alpha to measure the internal consistency of the responses provided to the supplied questionnaires, as per Leung [50] and Taherdoost [51]. The test was carried out in SPSS 23.

3.5 Data Analysis

The awareness level of the I4.0 technologies and their applicability in FBMIs was compared with the awareness scale of the model developed by Paguigan and Jacinto [52]. The normality of the data was tested using Shapiro–Wilk and Kolmogorov–Smirnov tests, and then a Measure of Sampling Adequacy (MSA) was conducted in SPSS 23 through Kaiser-Meyer Olkin (KMO) and Bartlett's test of sphericity. In assessing the reliable sampling adequacy, the KMO value had to be over 0.5 and a significance level for Bartlett's test below 0.05, thus suggesting that there is a substantial correlation in the collected data [53].

Both inferential and descriptive analyses were deployed to assess the gathered data regarding I4.0 awareness, challenges hindering I4.0 implementations, and strategies for implementing I4.0 in FBMIs. The awareness of I4.0 within the FBMIs was assessed by setting up a questionnaire regarding the operations or practices of I4.0 for the FBMIs. Figure 3 assisted in categorizing the awareness levels, as Paguigan and Jacinto [52] suggested.

SPSS 23 software was used to analyze the suggested factors. Such factor analysis approaches included Exploratory Factor Analysis (EFA), which analyzed challenges based on their categories (financial, lack of technical expertise, infrastructure, leadership and management, organizational culture, and regulatory compliance issues). EFA is an approach used to determine the underlying structure of a relatively large set of variables [54]. In assessing I4.0 challenges in FBMIs, an EFA assisted in determining the fundamental associations between observed variables. The decisive factor was based on the loading, which is suggested to be at least 0.5 [55]. The Kruskal-Wallis Test was conducted to ascertain whether the challenges were rated differently by industry categories (food manufacturing, beverage manufacturing, and food-beverage manufacturing).

3.6 Strategy Development and Validation Process

The strategy development process involves reviewing or gathering opinions on deploying I4.0 technologies within Tanzania's FBMIs. Strategies were compiled and included in the closed-ended questionnaire based on the literature review. All 103 FBMIS ranked strategies based on relative importance to each other. Moreover, optimal strategies were rated early and top-rated. The validation was carried out by conducting focus group discussions (FGDs). FGDs consisted of different experts from FBMIs. Three FBMIs involved members who did not rank strategies during the data collection. Three experts came from the food industry, three from the beverage industry, and three from the industry that produces both food and beverages. They included engineers and production managers for particular industries. The working experience of experts who validated strategies revealed that three experts had 5 to 10 years of experience, four experts had experience between 11 and 20 years, and two experts had experience above 20 years. The focus group was carried out at each purposively selected industry.

4.1 Education and Experience Level of Respondents

The target sample for the study was 103, and all FBMIs participated fully. Obtaining the exact sample size is regarded as an accurate population representation, but obtaining more data than the sample size is regarded as good [56]. The analysis of the educational level of respondents revealed that a whopping number of respondents, 61, had a bachelor's degree, which is equivalent to 59.2%. Among these, 17 respondents (16.5%) had masters, and 11 respondents (10.7%) had other levels of education that were not specified in the questionnaire. The study also found that 10 respondents (9.7%) had a diploma level of education, and four respondents (3.9%) had a PhD. The general finding on the respondent's education is that 79.2% of respondents had at least a bachelor's degree, indicating a higher level of education for the respondents.

The experience of respondent results revealed that five respondents (4.85%) had experience of less than 1 year, 16 respondents (15.5%) had 1–5 years of experience, 39 respondents (37.86%) had experience between 6 and 10 years, 23 respondents (22.3%) had experience between 11 and 20 years, and 20 respondents (19.4%) had experience above 20 years. The general insight on the respondents' experience revealed that 82 respondents, which amounted to 79.6% of total respondents, had worked in the FBMIs for over 6 years.

4.2 FBMIs Details

This part analyzed the categories of industries in which data were collected. The categories were divided into three categories: food manufacturing industry, beverage industry, and food and beverage industry. The result of the analysis showed that 59 companies (57.3%) were beverage manufacturing companies, 27 industries (26.2%) were food manufacturing companies, and 17 (16.5%) companies produced both foods and beverages. The number of employees in the FBMIs was analyzed to understand how big each company was. The analysis revealed that only medium and large MIs were available in the sample. Twenty-two companies, equivalent to 21.4%, had employees between 50 and 99, which are medium-sized, while 81 industries (78.4%) had more than 100 employees, which fall under the large company category. The financial capacity of industries was assessed in this section, and the annual turnover categories were divided into four categories: less than 10 million, 10 to 50 million, 51 to 100 million, and above 100 million. After analyzing the responses, it was revealed that 22 industries (21.4%) had a turnover between 51 and 100 million, and 81 industries (78.6) had a turnover above 100 million.

The manufacturing process of the sampled industries was assessed to understand the level of automation in the manufacturing process. The categories included in the questionnaire were manufacturing production processes, manual production lines, semi-automated production lines, automated production lines, and job shop production. The analysis of responses revealed that only automated and semi-automated production lines were selected, with 57.3% and 42.7%, respectively. The level of digitalization in the sampled industries was assessed, and the questionnaire asked respondents to select the level to which their industries were digitalized. The selection scale included a purely physical production system, a low-digitalized production system, a medium-digitalized production system, a well-digitalized production system, and a fully digitalized production system. The analysis of responses showed that 66 industries (64.1%) had a medium digitalized production system, 27 industries (26.2%) had a well-digitalized production system, and 10 industries (9.7%) had a fully digitalized production system.

4.3 Validity Test

The validity of the questionnaire was tested, first, through face validation as suggested by Yusoff et al. [47] and Roebianto et al. [48], and second, by using Pearson Product Moment Correlations using SPSS 23. The face validation was conducted through a representative sample by sending sample questionnaires to the target population and letting them feed the questionnaires. The questionnaires' responses were checked to find if the questionnaires were filled out accordingly. The returned questionnaire showed it was well prepared, and respondents understood the wording, enabling them to collect the needed data. The questionnaire items were valid based on the Pearson product–moment correlation since the significance values were less than 0.05 [49]. For example, p-values for general I4.0 awareness questions were 0.03, while awareness for I4.0 technologies was 0.009. Likewise, the p-values for challenges were 0.001 and 0.004 for the questionnaire items. Hence, the questionnaire was declared valid to collect data.

4.4 Reliability Test

Reliability is the ability of the research method to produce consistent results on repeated measurements under similar conditions [51]. The most commonly used internal consistency measure is CA. The reliability test was conducted in SPSS 23 on items according to their categories, that is, awareness of I4.0 technologies, challenges for implementation of I4.0, and strategies for implementation of I4.0. The reliability results were excellent, with CA above 0.90 for all items (based on standardized items) [51]. Table 2 shows the reliability results.

4.5 Awareness of the Applicability of I4.0 in FBMIs

The first part of this study's findings is intended to assess the awareness of I4.0 by the employees within the FBMIs through the practices of their companies, that is, the study asked questions regarding the operations or practices of I4.0 to the FBMIs. The purpose was to know whether employees understand the I4.0 technologies being used in their companies. The questionnaire asked respondents to rank each statement using the Likert scale of five points according to the level to which they agreed to a particular statement. The scale was (“1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree and 5—strongly agree”). The mean score for each statement was calculated, and then the awareness level was determined from the awareness ranges developed by Eski et al. [57] and Paguigan and Jacinto [52].

Paguigan and Jacinto [52] developed the scale for determining the awareness level of the tenants on campus policy on energy conservation and implementation using the mean scores, as Eski et al. [57] categorized their levels in 5-point; it was the same for Paguigan and Jacinto [52] as shown in Figure 3. The study adopted the approach used by Paguigan and Jacinto [52]; the mean scores for the ratings on the applicability of I4.0 were calculated, and the results were subjected to the ranking in the scale developed by Paguigan and Jacinto [52]. The findings showed that the overall mean score was 3.75, corresponding to awareness on the five-point Likert scale of awareness. Figure 4 shows the overall mean score.

The score for each factor revealed that employees were more informed about the benefits of adopting I4.0 (AW₅), with a mean score of 4.78 ranked first. The prospect of I4.0 to improve industry's competitiveness on the international stage was ranked second in terms of awareness by employees, with a mean score of 4.65, corresponding to full awareness. The other aspect employees were fully aware of was the possibility of the company improving market share through implementing I4.0 technologies (AW₁₄) with a mean score of 4.30. Respondents were unaware of only one aspect, that is, the framework or roadmap for planning the activities for implementing I4.0 (AW₉), with a mean score of 2.44. This generalizes that most FBMIs have not fully established a clear roadmap to use I4.0 technologies. Table 3 shows the results for the awareness of the applicability of I4.0 in FBMIs.

The independent t-test was conducted to compare awareness level on the applicability of I4.0 in FBMIs for food and beverage, there were no significant difference on the awareness level between food and beverages for six items with t-values greater than 1.96 [58]; AW₃—“Have you attended any workshop or training or seminar regarding to I4.0 or digital transformation technologies in the last 2 years?”, AW₄—“Is your organisation aware of the current technologies associated with I4.0?”, AW₇—“Are there any financial arrangements for the implementation of I4.0 technologies?”, AW₈—“Has an I4.0 strategy been defined in your company for the next five years?”, AW₉—“Is there a framework/roadmap for planning the activities for the implementation of I4.0 solutions in your company?”, AW₁₀—“Is there any involvement, support, or experience of top management toward the implementation of I4.0?”, and AW₁₃—“Are you aware that the digital transformation to I4.0 could improve your industry's competitiveness at the international stage?” The remaining items displayed a significant difference in the awareness level with a t-value of less than 1.96 [59]. Table 3 shows the results for the independent t-test and awareness level on the applicability of I4.0 in FBMIs. The t-values can be negative or positive in Table 3; they mean that they lie to the left of the mean [60]. The t-distribution, like the standard normal, has a mean of 0, and all values to the left are negative, while values to the right are positive [61].

4.6 I4.0 Technologies Awareness Level by FBMIs

The awareness of I4.0 technologies by FBMIs was assessed to determine how well I4.0 technologies were understood among the FBMIs. The approach used by Paguigan and Jacinto [52] was maintained in this section to assess the awareness of the applicability of I4.0 technologies in FBMIs. The analysis showed that the overall mean score was 3.50, corresponding to awareness on the awareness scale. This means that, in general, FBMIs were aware of I4.0 technologies. Figure 5 shows the overall mean score.

The individual technology assessment showed that artificial intelligence (TW₁) was a very popular technology with a mean score of 4.8, which corresponds to fully aware on the five-point awareness scale. The analysis found that seven technologies were well known as respondents were fully aware of them, together with TW₁, others including TW₄ ($$$ \overline{x}=4.54 $$$) autonomous robotics, TW₅ ($$$ \overline{x}=4.50 $$$) cyber security, (TW₁₆, $$$ \overline{x}=4.43 $$$) “smart sensors,” (TW₁₅, $$$ \overline{x}=4.34 $$$) “IoT,” (TW₆, $$$ \overline{x}=4.33 $$$) “simulation” and (TW₂₃, $$$ \overline{x}=4.20 $$$) “RFID chips (Radio Frequency Identification for object tracking).” On the other hand, the analysis revealed that FBMIs were not aware of five technologies namely; (TW₂₂, $$$ \overline{x}=2.50 $$$) $$$ \mathrm{digital}\ \mathrm{twin}, $$$ (TW₁₈, $$$ \overline{x}= $$$ 2.22) “dark manufacturing,” (TW₂₁, $$$ \overline{x}=2.17 $$$) “virtual factory,” (TW₁₃, $$$ \overline{x}= $$$ 2.16) “block chain technology” and (TW₁₄, $$$ \overline{x}=2.07 $$$) “nanotechnology.” Table 4 depicts the awareness of I4.0 technologies by FBMIs.

Also, an independent t-test (Table 4) was conducted to compare the I4.0 technologies awareness level between FBMIs. The result of independent t-test statistics showed that there was no significant difference on the awareness level on nine I4.0 technologies: TW₁—“artificial intelligence (AI),” TW₂—“cyber-physical systems,” TW₃—“big data analytics,” TW₅—“Cyber security,” TW7—“the industrial IoT,” TW₈—“cloud computing,” TW10—“augmented reality,” TW₁₁—“cyber manufacturing” and TW₁₇—“cyber-physical systems.” These nine items had t-values greater than 1.96. The remaining items showed a significant difference in the level of awareness among FBMIs with t-values less than 1.96 [59].

4.7 Challenges for Implementation of I4.0 Technologies in FBMIs

The study assessed challenges hindering the implementation of I4.0 in FBMIs in Dar es Salaam and Pwani regions. The implementation challenges were categorized into six groups (financial, lack of technical expertise, infrastructure, leadership and management, organizational culture, regulatory compliance, and legal issues). The first part of the analysis on the challenges of I4.0 technologies implementation included descriptive analysis followed by a normality test to check data distribution using the Kolmogorov–Smirnov test and Shapiro–Wilk.

The second part of the analysis of the challenges involved the use of EFA to assess the theoretical construct of the variables. Before factor analysis, MSA was conducted through KMO to ensure the possibility of conducting factor analysis [55]. Then, the Kruskal Wallis test was used to assess the responses based on categories of FBMIs, followed by developing strategies for implementing I4.0 in FBMIs.

4.7.3 EFA on the Challenges Hindering the Implementation of I4.0 in FBMIs

The MSA results showed that the collected data were suitable for conducting factor analysis as the value of KMO was 0.785. For conducting factor analysis, the KMO value must be greater than 0.5, as suggested by several studies (Abdol [54, 63]). EFA results supported the six-dimension structure of the study through the extraction of components with Eigen values greater than one [64], as shown in Figure 6 and Table 7.

TABLE 7. EFA results for challenges of implementing I4.0 in FBMIs.

		Component
Codes	Statement	1	2	3	4	5	6
FC1	The investment costs associated with implementing I4.0 technologies				0.776
FC2	The return on investment (ROI) and cost–benefit analysis for implementing I4.0 technologies				0.693
FC3	Identifying and securing funds for implementing I4.0 technologies				0.635
FC4	Investing in new equipment, sensors, and data analytics systems can be financially burdensome				0.848
FC5	Cost for maintenance and repair of I4.0-related technologies				0.726
FC6	Costs to hire/recruit technical personnel on advanced technologies				0.759
FC7	Costs associated with training employees, consultancy services and integrating these technologies				0.706
TC1	Adopting and implementing I4.0 technologies requires specialized technical knowledge						0.661
TC2	The industry may lack the in-house expertise (or skilled workforce) or resources to navigate and utilize these technologies effectively						0.810
TC3	Difficult to understand which technologies are most relevant to specific operations and how to integrate them into their existing systems						0.863
TC4	Poor research and development on I4.0 implementation						0.697
IC1	Incorporating new I4.0 technologies into existing infrastructure					0.738
IC2	Unreliable electricity supply to the industry					0.792
IC3	Unreliable internet connectivity to the industry					0.515
IC4	Inadequate advanced software to facilitate I4.0 implementation					0.767
IC5	Absence/inadequate of additional resources for computational power and robust hardware to facilitate I4.0 implementation					0.649
IC6	Upgrading or replacing existing systems can be disruptive and time-consuming, requiring careful planning and investment					0.690
IC7	Limited resources to invest in robust cybersecurity measures, making them more vulnerable to potential breaches and data theft					0.829
MC1	Having top management which does not support I4.0 implementation	0.787
MC2	Having leaders and top management who are not committed enough to put I4.0 into practice	0.914
MC3	Having the organizationorganisation's leadership/top management who have less awareness/knowledge about I4.0	0.801
MC4	The organizationorganisation's leadership/top management who have no vision and strategy for I4.0 implementations	0.831
MC5	Having organization which has no roadmap or implementation framework for I4.0	0.855
OC1	Resistance to change within organizations can hinder the adoption of I4.0 technologies		0.855
OC2	Challenges in terms of employee mindset and willingness to embrace new technologies and adapt to new ways of working		0.876
OC3	Collaboration and communication challenges between various departments and teams within a company		0.756
OC4	Building a culture of innovation is a cultural challenge that companies need to overcome when implementing I4.0		0.859
OC5	Fear of job displacement due to I4.0 implementation		0.839
OC6	The integration of connected devices and increased data collection in I4.0 raises concerns about data security and privacy		0.874
RC1	Legal issues and complications of your industry			0.866
RC2	Adopting I4.0 technologies may require industry to ensure that these technologies comply with industry-specific regulations			0.830
RC3	Meeting compliance requirements and ensuring traceability throughout the supply chain can be complex and resource-intensive			0.813
RC4	I4.0 lacks global standards and protocols, resulting in interoperability issues between different systems and technologies. This hinders the seamless integration needed for a truly connected industry.			0.907
RC5	Lack of governmental support and policies			0.763

The criteria for approving study constructs was through factor loading above 0.5 of each factor on their respective variables [55]. All factors presenting the challenges for implementing I4.0 technologies in FBMIs were retained in the study as they had loadings above the minimum value of 0.5, as Razali et al. [55] explained. One factor of RC5 ‘lack of governmental support and policies was dropped as it did not load in any of the six extracted components. The extracted components had a cumulative variance explanation of 76.92%.

4.8 Strategies for Implementation of I4.0 Technologies in FBMIs

The responses to the proposed strategies (Table 9) for implementing I4.0 in FBMIs were analyzed in this section. The mean score and RII for each strategy were calculated. The scores showed that ST₉ “Embrace data-driven decision making” was the strategy with the highest mean score of 4.8 and RII = 0.959, which indicates that respondents consider this strategy to be most important toward the implementation of I4.0 technologies in FBMIs, the lowest ranked strategy was ST₂₁ “Training leaders and top management regarding the economic benefits of I4.0 implementation” with a mean score of 2.98 and RII = 0.596. all strategies were retained due to EFA since their extraction values were above 0.5.

Therefore, all strategies had mean scores above 4.0, corresponding to “agree” in the Likert scale measurement, which implies that respondents agree to the proposed strategies except ST21 “Training leaders and top management regarding the economic benefits of I4.0 implementation” which had a mean score of 2.98.

Focus group discussions consisted of a different group of experts from FBMIs. Three FBMIs involved members who did not rank strategies during the data collection. Three experts came from the food industry, three from the beverage industry, and three from the industry that produces both food and beverages. They included engineers and production managers for particular industries. The focus group was carried out at each purposively selected industry.

4.9 Discussion

The study's main objective is to assess challenges and propose strategies for implementing I4.0 in FBMIs. Proposing strategies and assessing challenges were first preceded by assessing I4.0 awareness levels for FBMIs in Dar es Salaam and Pwani Regions. The results indicate a high level of awareness from 103 FBMIs. Also, the findings demonstrated no significant differences in the awareness level among categories of FBMIs. Such results were also observed by Ntamulyango [22], who assessed awareness among beverage industries in Dar es Salaam. However, there is a difference between the results of the current study and Ntamulyango's [22] study. Ntamulyango [22] only involved beverage industries, while the current study involved food and beverages. Therefore, the awareness level of implementing I4.0 in FBMIs can be generalized. Likewise, when Maganga [37] developed the transition framework for the selected MIs in Dar es Salaam and Pwani regions, his study concluded that the awareness level for the manufacturing mainly on I4.0 for the quality issues was found to be high.

The awareness level was also confirmed through the performed t-test. For example, the independent t-test was conducted to compare awareness levels on the applicability of I4.0 in FBMIs for food and beverage. Results concluded that there were no significant differences in the awareness level between food and beverages for six items with t-values greater than 1.96 [58]; AW₃—“Have you attended any workshop/training or seminar regarding I4.0 or digital transformation technologies in the last 2 years?”, AW₄—“Is your organisation aware of the current technologies associated with I4.0?”, AW₇—“Are there any financial arrangements for the implementation of I4.0 technologies?”, AW₈—“Has an I4.0 strategy been defined in your company for the next five years?”, AW₉—“Is there a framework/roadmap for planning the activities for the implementation of I4.0 solutions in your company?”, AW₁₀—“Is there any involvement, support, or experience of top management toward the implementation of I4.0?” and AW₁₃—“Are you aware that the digital transformation to I4.0 could improve your industry's competitiveness at the international stage?” The remaining items displayed a significant difference in the awareness level with a t-value of less than 1.96 [59]. Table 3 shows the results for the independent t-test and awareness level on the applicability of I4.0 in FBMIs. The _t-values can be negative or positive in Table 3; they mean that they lie to the left of the mean [60].

The next stage was to investigate the variations in the difficulties associated with implementing I4.0 for FBMIs; null hypothesis testing was required, as suggested by Tay et al. [65]. The test was conducted to ascertain whether the challenges were rated differently by industry categories (food manufacturing, beverage manufacturing, and food-beverage manufacturing). So, the following are the null hypothesis and alternative hypothesis based on FBMI categories: Ho: There have been no statistical differences in the challenges of implementing I4.0 for FBMIs and Ha: There have been statistical differences in implementing I4.0 for FBMIs.

Thus, the study adopted the Kruskal Wallis test for grouping variables as per Nimawat and Das Gidwani [34]. Results concluded that there were no significant differences for most challenges of implementing I4.0 in FBMIs between food manufacturing, beverage manufacturing and those that produce both food and beverage. There is a significant difference between FBMIs on TC₁, TC₂, TC₃ and MC₅. When comparing these challenges with other studies, the majority are similar, except for some that seem unique for the FBMIs, specifically in Tanzania. For example, even though I4.0 concepts emerged in Germany in 2011, scant studies in Tanzania focus on assessing challenges and conclude that “Poor research and development on I4.0 implementation” was one of them. Existing studies, including Ntamulyango [22] and Maganga [37], did not conclude that the absence of research and development focusing on I4.0 is one of them.

Having assessed challenges affecting the implementation of I4.0 in FBMIs, the next stage was looking for the potential solution toward those challenges. For example, Santos and Martinho [66] emphasized the need for implementable strategies to address challenges. Thus, the study proposed strategies for implementing I4.0 in the FBMIs. Each strategy's mean score and RII were calculated. The scores showed that ST₉ “Embrace data-driven decision making” was the strategy with the highest mean score of 4.8 and RII = 0.959, which indicates that respondents consider this strategy to be most important toward the implementation of I4.0 technologies in FBMIs, the lowest ranked strategy was ST₂₁ “Training leaders and top management regarding the economic benefits of I4.0 implementation” with mean score 0f 2.98 and RII = 0.596. all strategies were retained due to EFA since their extraction values were above 0.5. The obtained strategies align with various I4.0 strategies proposed by previous studies, including Vidanagamachchi and Lanka [35], Mpangule [36], Maganga and Taifa [27], Khin and Kee [15], Nguyen and Luu [24], Bettiol et al. [12], Khin and Kee [15] and Oria et al. [11].

4.10 Significance of the Study

This research has academic and practical implications. Since the research focused on understanding the level of I4.0 adoption in FBMIs, it is evident that the literature contributes to the existing one. Moreover, practitioners must realize the potential of I4.0, particularly how digital processes are going. Various challenges and strategies for I4.0 adoption by FBMIs are addressed, which can also serve as a yardstick for FBMIs, policymakers, and the government in Tanzania to establish foundations for I4.0 logistics to be employed in FBMIs to maintain the production of goods sustainably. Consequently, implementing the strategies is anticipated to improve the FBMIs in adopting I4.0, thus accruing associated benefits. The benefits would generally lead to optimized costs, quality, flexibility, dependability, speed, connectivity, innovation, transparency, responsiveness, and real-time inventory management [67, 68].

5.1 Recommendation

Although the FBMIs have demonstrated a high level of I4.0 awareness of technologies and the general I4.0 conceptualization, it is recommended that stakeholders keep fostering the emphasis on implementing all pertinent technologies fully. This would enable industries to accrue benefits from the full implementation of I4.0 in such industries. The study recommends implementing the proposed strategies, which should be fostered by the top management commitment and fully supported by allocating a sufficient budget. Since the research focused on understanding the level of I4.0 awareness in FBMIs, practitioners must realize the potential of I4.0, particularly how I4.0-leveraged technologies can impact industries. Various strategies for I4.0 adoption by FBMIs are addressed, which can also serve as a yardstick for FBMIs and other pertinent stakeholders to establish foundations for I4.0 logistics to be employed in FBMIs to produce goods sustainably.

5.2 Recommendations for Future Studies

The study explored awareness, challenges, and proposed strategies by engaging FBMIs within the Dar es Salaam and Pwani regions. Thus, other studies can be conducted in the rest of the regions. Also, only the FBMIs were the focused industries; other manufacturing industries, including tobacco, metal, plastic, textiles, leather, and mining, should be investigated to establish awareness levels and propose vital strategies. Moreover, only large and medium industries were involved, even for food and beverages; there is room for investigating challenges that hinder micro and small enterprises in implementing some I4.0-related technologies. It might not be easy for micro and small enterprises to implement all I4.0 technologies, but some technologies should assist such industries in accruing the benefits of deploying I4.0.