2.1. Digital transformation and Industry 4.0

In the field of digital transformation related to Industry 4.0, many scientific studies and examples of practical applications can be found. The presented review refers only to the most relevant works. It also includes publications related to the assessment of digital maturity of enterprises.

The concept of Industry 4.0 was launched in Germany in 2011 [18], and the term refers to the dynamically occurring changes due to the practical application of modern technologies in social and economic life.

In the United States, on the other hand, the concept of Industry 4.0 is commonly known as the Internet Industry of Things, advanced manufacturing or smart manufacturing. Although the term "smart industry" was originally used only in the United States, it is now used worldwide both in industry and scientific units [19].

According to Dilberoglu et al. [20] and Mosterman i Zander [21], Industry 4.0 is an integrated set of intelligent manufacturing systems and advanced information technologies based on integrated sets of software systems. However, according to Guoping et al. [22], Industry 4.0 is a set of technologies based on the digitization and interconnection of all production units present in an economic system. Therefore, it can be assumed that Industry 4.0 is the integration of various systems through the use of digital resources, including intelligent ones, which by communicating with one another can support and make certain decisions as well as perform various operations with much less human involvement than before [23]. The basic technologies of Industry 4.0 along with literature references are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Industry 4.0 technologies.

https://doi.org/10.1371/journal.pone.0253965.t001

In terms of publications on assessing the level of maturity and/or readiness for the Industry 4.0 technologies, it should be noted that their number has been growing recently.

One study [33] proposes a novel and innovative model to assess the degree of readiness of a manufacturing company to implement the Industry 4.0 technologies. Based on this model, a tool was developed to identify actions needed to increase the degree of readiness of companies to implement the Industry 4.0 principles and practices. The proposed approach was inspired by the framework developed by SAE to identify and measure best practices for implementing lean manufacturing in companies [34].

Leineweber et al. [35] presented a method for assessing the digital maturity of enterprises using more than 40 evaluation criteria from the areas of technology, human resources and organization. Sarvari et al. [36] showed a technology development roadmap for the Industry 4.0 transformation in an enterprise to facilitate the planning and implementation process of this concept. The map developed helps to understand each move and what decisions need to be made, as well as who should make them and when during the digital transformation in an enterprise. Schumacher et al. [15] proposed a model to assess the maturity of Industry 4.0 among industrial enterprises in the domain of discrete manufacturing. This model includes 9 dimensions to which 62 items were assigned to evaluate the maturity of Industry 4.0. The dimensions "Products", "Customers", "Operations" and "Technology" were created to assess the basic factors. In addition, the dimensions "Strategy", "Leadership", "Management", "Culture" and "People" allowed for the inclusion of organizational aspects in this assessment.

Branco et al. [37] examined factors characterizing readiness for Industry 4.0. in the manufacturing companies of the EU countries. The analysis proved that the existence of digital infrastructure combined with analytical capabilities to handle large data sets makes it possible to achieve a high level of readiness of manufacturing companies for Industry 4.0. Basl and Doucek [38] developed a metamodel to assess an organization’s digital readiness for Industry 4.0. It ranks selected maturity models and readiness indicators against each other, while identifying areas with potential for further research. Bibby and Dehe [39] developed an assessment framework for maturity for Industry 4.0. in a central company to later compare it with 12 organizations in its supply network. Enterprise assessment focuses on three dimensions: smart factory, people and culture, and strategy. Akdil et al. [40] proposed an Industry 4.0 maturity assessment model and a questionnaire to assess this level. The model covers different application areas of Industry 4.0 such as smart finance, smart marketing and human resources. In order to determine a company’s maturity level for Industry 4.0, four stages were included: "None", "Existence", "Survival" and "Maturity". The questions in each criterion had a 4-stage scale to assess the level of digital maturity–from 0:"None" to 3:"Maturity". Kuruczleki et al. [41] determined the readiness of the EU-28 countries for the fourth industrial revolution by creating a readiness index of Industry 4.0 consisting of eight indicators: total intramural R&D expenditure, gross domestic expenditure on R&D, community trade mark applications, community design applications, total R&D personnel and researchers, tertiary educational attainment, ICT specialists, digital single market–promoting e-commerce for businesses, enterprises selling online. Lizarralde et al. [42] developed a maturity model for machine tool companies. This maturity model is evaluation and identification tool for the areas of the organization where specific development is required. Liebrecht et al. [43] developed a methodology to support decision making of industrial companies for the Industry 4.0 method application in a production environment. Tortora et al. [44] conducted a survey on the level of readiness of Italian small and medium-sized enterprises to introduce the idea of Industry 4.0. The results showed that these companies have limited and insufficient knowledge of Industry 4.0 technologies. More than 50% of the respondents demonstrated a basic or low level of knowledge of the technologies in question. Saad et al. [45] adapted and used the Smart SME Technology Readiness Assessment (SSTRA) methodology to investigate the level of technology readiness of these companies to implement Industry 4.0 based on smart production planning and control. De Carolis et al. [46] proposed a model to study the digital maturity of a company. The assessment was carried out in 5 areas: design and engineering as well as production, quality, maintenance and logistics management. This approach considers all relevant aspects related to the idea of Industry 4.0. Lin et al. [47] used a model based on the Singapore Smart Industry Readiness Index by the Economic Development Board to assess digital maturity. This model enables companies to conduct self-assessment in order to systematically and comprehensively adapt to Industry 4.0. The assessment is conducted in three Building Blocks: Process, Technology and Organization, and 16 dimensions in total. Jung et al. [48] introduced the Smart Manufacturing System Readiness Level (SMSRL) to measure a company’s digital readiness in four dimensions: organizational maturity, IT maturity, performance measurement, and information connectivity. The model enables small and medium-sized manufacturing companies to assess their readiness to implement the necessary technologies that can help them in their digital transformation.

A summary of selected models for assessing digital maturity and readiness, along with their brief description, is presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Maturity and readiness models.

https://doi.org/10.1371/journal.pone.0253965.t002

2.2. Application of MCDM methods

MCDM methods are used to solve complex decision-making issues in various areas. They are also adopted in issues related to Industry 4.0 and the digitization of economies and societies [56–62] (Table 3). However, so far these methods have not been used to assess the digital maturity of enterprises.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of applications of MCDM methods to solve different problems.

https://doi.org/10.1371/journal.pone.0253965.t003

2.3. Research gaps

The literature search shows that the topic of digitalization is current and refers to various aspects of this process. Research on the effects of its implementation seems to be fully justified both for the current assessment and for the prospects of development.

Its results also indicate that in the vast majority of the presented studies, on the assessment of digital readiness and/or maturity, no importance weighting was taken into account for the adopted criteria or dimensions (factors) [33,39,63–65].

Only a small number of these works include weights for the adopted factors for assessment [15,50,66].

The variety of factors taken into account in these studies, and their different impact on the processes studied, shows the relevance of including weights in multi-criteria analyses regarding digitization.

The search also shows that the multiple criteria-decision making methods have not been used so far to assess digital readiness and/or maturity of enterprises. Also, the presented assessments failed to include developing countries, which are undoubtedly the countries of Central and Eastern Europe.

Therefore, there is a research gap in terms of studying digital maturity in the CEE companies, which is due to the lack of this type of studies. That is why, the conducted literature search fully rationalizes examining such a research subject. Also, the application of computational methods from the area of multi-criteria decision-making methodologies such as TOPSIS, MOORA and VIKOR methods, makes this study a new approach to this type of analysis. It allows for a broader look at the research and increases its both reliability and credibility. In turn, the obtained knowledge creates wide possibilities of interpretation, inference and formulation of recommendations.

3.1. Area of research

The European Union is a community of 27 member states, 11 of which belong to the countries of Central and Eastern Europe (CEE) (Fig 2). These countries include Poland, the Czech Republic, Slovakia, Slovenia, Bulgaria, Romania, Croatia, Lithuania, Latvia, Estonia, and Hungary (Fig 2, Table 4). All of them have common cultural and historical roots. They also share a common past as socialist buffer states under the USSR or as its constituent parts, namely the Soviet republics. The term Central and Eastern European countries is a composite of two terms for this part of Europe–geographical (Central) and political (Eastern). Basic information about these countries is presented in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Countries of Central and Eastern Europe belonging to the European Union.

https://doi.org/10.1371/journal.pone.0253965.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Basic information about the CEE countries belonging to the EU (own elaboration based on [67]).

https://doi.org/10.1371/journal.pone.0253965.t004

3.2. Materials

In order to assess the level of digital maturity of enterprises, including manufacturing enterprises in the CEE countries, data from the Eurostat database were used [67]. They concern the use of information and communication technologies (ICT) in enterprises and digital skills. On the basis of this data, a set of determinants was determined to rank and assess the level of digital maturity of enterprises in the ECC countries and to analyze their similarities.

In order to assess the level of digital readiness of all enterprises and manufacturing enterprises in the CEE countries, it was necessary to select appropriate determinants. Based on the literature review, 9 areas that have a decisive influence on the level of digital readiness of manufacturing companies were selected. Eight areas are related to the Industry 4.0 technologies and one to digital skills of employees.

For these areas, 14 determinants of digital maturity were adopted, the description of which is presented in Table 5. The selected determinants of digital maturity meet the postulate of relevance to the topic related to Industry 4.0 (they refer to the main pillars of Industry 4.0). They are characterized by simplicity in construction, ease of interpretation of results, availability and high quality of data, which are updated on an ongoing basis, as well as they are comparable at an international level.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. The adopted determinants characterizing the digitalization processes of enterprises in the CEE countries.

https://doi.org/10.1371/journal.pone.0253965.t005

3.3. Methods

A set of multi-criteria decision-making methods was used to assess the level of digital maturity of enterprises in the CEE countries. These methods include: TOPSIS method, VIKOR method and MOORA method. Brief characteristics of these methods are presented in this section.

3.3.1 The VIKOR method.

The VIKOR method belongs to the methods of multi-criteria decision optimization and is based on the concept of measuring the distance of the studied variant from the ideal scenario [68]. This method introduces the so-called ranking index, based on the measurement of distance from the ideal solution. This index is an extension of the aggregate function theory in the compromise programming method [69,70]. Individual variants belonging to set A (a₁, a₂,… a_m) and evaluated by n criteria are described by the factor fij, which is the weight of the variant a_j with respect to the criterion n_i.

The output parameter for the analysis conducted by the VIKOR method is the Lp-metric distance, determined from Eq (1): (1) where: and are the best and worst values of all criterion functions for all alternatives from set A (a₁, a₂,… a_m); n is the number of criteria.

The algorithm of the VIKOR method to determine the compromise ranking is as follows:

• to construct the decision matrix, according to Eq (2):

(2) where: x_ij∈ℝ

• to construct the normalized decision matrix, according to Eq (3):

(3)

• to determine the best and worst values for all criterion functions i = 1, 2,…n. If the i-th criterion represents profit (the higher the better), then and ; if the i-th criterion represents cost (the lower the better), then and ;
• to calculate S_j and R_j values forming the ranking measure, from the following relationships:

(4) (5) where: w_i weight of the i-th criterion; S_i and R_i represent the utylity measure and the regret measure.

The solution obtained by min_jS_j is with the maximum group utility (’majority’ rule), and the solution obtained by min_jR_j is with the minimum individual regret of the ’opponent’

• To calculate the VIKOR–Q_j index value (6):

(6) where: (7) where: v is the measure of strategy weight (takes value from 0 to 1); S_j and R_j are calculated in Step 3 and are introduced as the weight of strategy of ’the majority of criteria’ (or ’the maximum group utility’). In the present study, v = 0.5.

• to create the ranking of alternatives according to the value of Q_j.

The alternative with the smallest VIKOR value is referred to as the best (Q _minimum).

3.3.2 The TOPSIS method.

The TOPSIS method is one of the most popular methods used for solving multicriteria discrete tasks [71,72]. The considered decision variants are compared with abstract weighted reference solutions: ideal and anti-ideal. What distinguishes this method is the use of the measure of relative distance to the best solution, representing the pattern (ideal) and the worst solution, representing the anti-pattern (anti-ideal).

Stages of the research procedure in the TOPSIS method are as follows:

• to determine the decision matrix, according to Eq (2).
• to determine the normalized matrix, according to Eq (3).
• to determine the weighted normalized decision matrix:

(8) where: w_i is the weight of the i-th criterion

• to determine the ideal solution S⁺ and the non-ideal solution S^-:

(9)(10)

• to determine the Euclidean distance of the object from the ideal variant S⁺ and the non-ideal variant S^-:

(11)(12)

• to determine the coefficient of relative closeness of the decision variants S_i to the ideal solution S⁺ (14):

(13)

The values of the TOPSIS synthetic measure belong to the range <0.1>, yet the higher the value reached by the synthetic measure, the higher position the object achieves in the ranking. The synthetic index (P_i) calculated for each country orders the units linearly and allows the classification from the highest level to the lowest level.

3.3.3 The MOORA method.

The MOORA method by Brauers and Zavadskas [73] method belongs to the family of multi-criteria decision optimization methods. The application of this method makes it possible to select the best alternative, evaluated in terms of favorable and/or unfavorable criteria [74].

The main steps in applying the MOORA method are as follows:

• to create a decision matrix X with m number of alternatives and n number of criteria according to Eq (2).
• to create the normalized decision matrix according to Eq (3).
• to calculate the normalized score value for each alternative, taking into account all alternatives. In fact, the final score of each alternative is obtained by means of the following equation:

(14)

In this equation, represents the MOORA score for the i-th alternative. j = 1, 2, 3……t and j = t+1, t+2……n refer to the objectives that must be appropriately maximized (favorable criteria) and minimized (unfavorable criteria).

• If weights are included in the evaluations of alternatives with respect to specific criteria, then the MOORA score for each alternative is measured by the following equation:

(15) where: w_i is the weight of the i-th criterion.

In the final ranking, alternatives with higher scores are desired. In other words, the alternative with the highest score is considered the best option and the alternative with the lowest score is seen as the worst option.

3.4. Method for determining the digital maturity level

All of the discussed methods allow solution ranking, from the best to the worst. However, none of them enable the determination of the level of digital maturity. Therefore, the authors decided to determine the level of digital maturity by using the mean value and standard deviation. The level classes of digital readiness in the MOORA and TOPSIS methods are defined as follows:

Class I—high level of digital maturity: (16)

Class II–average high level of digital maturity: (17)

Class III–average low level of digital maturity: (18)

Class IV–low level of digital maturity: (19)

On the other hand, in the VIKOR method, where the ideal solution is closest to or equals 0 the above ranges must be reversed, i.e., Eq (16), which determines the low level in the TOPIS and MOORA methods, allows for the high level to be determined in the VIKOR method, and so on:

Class I—high level of digital maturity: (20)

Class II–average high level of digital maturity: (21)

Class III–average low level of digital maturity: (22)

Class IV–low level of digital maturity: (23) where: and are the average values of P_i, Q_j and and are the standard deviation of P_i, Q_j and .

3.5. The Shannon’s entropy method–to determine weights of determinants

The Shannon’s entropy method was used to determine the weights of the assumed determinants. The algorithm for determining the weights in this method is as follows:

• to construct the decision matrix according to Eq (2).
• to construct the normalized decision matrix:

(24)

• to determine entropy:

(25)

(26)

• to determine the variation level of entropy for each criterion (the degree of intrinsic divergence of scores from subsequent criteria) from Eq (27):

(27)

• to determine the weights (degree of importance) of the criteria according to Eq (28):

(28)

3.6.Multidimensional scaling

In order to graphically represent the structure of similarity (or dissimilarity) between the studied objects based on the selected set of variables (features), multidimensional scaling (MDS) was used. The graphical presentation of results takes the form of a scatter plot of objects on a multidimensional scaling map. The smaller the distance between the analyzed objects, the more similar they are to one another. The graphical representation of results takes the form of a 2- or 3-dimensional map [75].

What distinguishes MDS from other similar techniques (e.g., factor analysis, cluster analysis) is that in the MDS method, there are no biases as to which factors may influence which dimensions [76].

Under the assumption that there are n points, denoted as x1, x2, ⋅⋅⋅⋅, xn, the vector form is expressed as: (29) then the algorithm for the multidimensional scaling process consists of the following steps:

• to determine the distance or dissimilarity between the number of n points using a certain algorithm, and then to obtain the inner product matrix B_n × n according to equation:

(30)

• to calculate the eigenvalues r₁, r₂, ⋅⋅⋅ r_s and eigenvectors v of the B_n × n dimensional matrix, which makes it possible to obtain the coordinates of all points in the dimensional matrix. In this way, a perceptual map can be created.

The quality of matching the output (reconstructed data) to the input data is measured by the so-called STRESS function. The smaller the value of this function, the better the match of the reconstructed distance matrix to the observed distance matrix. The STRESS function is defined as the root of the standardized sum of squares of the residuals between the input distances and the distances reconstructed by multidimensional scaling. It takes the following form [77]: (31) where: d_ij is the distance or dissimilarity between points i; is a function on the input data, and it depends on whether one is dealing with metric or non-metric multidimensional scaling.

In metric multidimensional scaling, , while in non-metric scaling, the function is a monotonic transformation of the observed input data.

The criteria for evaluating the matching results are shown in Table 6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. STRESS evaluation criteria [own elaboration based on [75]].

https://doi.org/10.1371/journal.pone.0253965.t006

Multidimensional scaling uses the Euclidean distance as a measure of distance between points. By denoting the coordinates of objects in dimension r (r-dimension) as X_i = (X_i1, X_i2,…….X_ir), the Euclidean distance between X_i and X_j can be defined as: (32)

4.1. The statistical analysis of the determinants of digital maturity among enterprises in the CEE countries, including manufacturing enterprises

The indicators used for the study, characterizing the technological and social dimensions of digital maturity of enterprises in the CEE countries, were pre-processed and their basic statistical parameters were determined, which is summarized in Tables 7 and 8. These indicators were delineated for all and manufacturing enterprises (according to NACE rev. 2 activity).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Selected statistical parameters of the indicators determining the level of digital maturity among all enterprises in the CEE countries.

https://doi.org/10.1371/journal.pone.0253965.t007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Selected statistical parameters of the indicators determining the level of digital maturity among manufacturing enterprises in the CEE countries.

https://doi.org/10.1371/journal.pone.0253965.t008

Based on the results, it can be concluded that the determinants of digital maturity selected for analysis (treated as variables), meet the condition of diagnostic features, which must be marked with significant variation (above 10%). For these determinants, the value of the coefficient of variation was found to be characterized by significant spread. The highest value of the coefficient of variation for manufacturing enterprises was found for the determinants "use of sensors or RFID tags to monitor or automate production processes (to manage logistics, to track the movement of products)"– 108.11% and for "analysis of big data internally using machine learning"– 89.44%. For all enterprises, the highest value of the coefficient of variation within the studied group was reported for the determinant "enterprises using software solutions like Customer Relationship Management (CRM)"– 69,53%. The lowest value of the coefficient of variation for all enterprises was found for the determinant “purchase of high CC services (25.64%), and for manufacturing enterprises, it was “analysis of big data from the geolocation of portable devices” (31.01%).

The difference between the median and mean values of the determinants shows distribution asymmetry. The positive sign (median greater than the mean value) characterizes the left-sided asymmetry (left-skewed distributions), indicating the predominance of countries with high determinant values. The negative sign is associated with the right-sided asymmetry (right-skewed distributions), indicating the predominance of countries with low determinant values.

The left-skewed asymmetric distributions in the case of all enterprises were found for the determinants “analysis of big data from smart devices or sensors”, “purchase of high CC services, use of smart meters, smart lamps, smart thermostats to optimize energy consumption in the enterprise’s premises”, and “use of 3D printing”.

In the case of manufacturing enterprises, the left-skewed asymmetric distributions were found for the determinants “analysis of big data from smart devices or sensors”, “analysis of big data from the geolocation of portable devices”, “use of smart meters, smart lamps, smart thermostats to optimize energy consumption in the enterprise’s premises”, and “enterprises that provide training to develop/upgrade ICT skills of their personnel”. For the remaining determinants of all and manufacturing enterprises, the determinants of digital maturity adopted the right-skewed asymmetric distributions.

Since all determinants are stimulants, it is reasonable to assume that only for determinants with the left-skewed asymmetry distributions, a favorable situation in terms of their level was noted. The predominance of countries with high values was observed.

With regard to the comparison of the CEE countries to the entire European Union (EU-27), Fig 3 presents the distribution of values of the digitization determinants for enterprises in these countries.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Summary of the values of studied determinants with regard to digital maturity among enterprises in the CEE countries compared to the EU countries (a–all enterprises, b–manufacturing enterprises).

https://doi.org/10.1371/journal.pone.0253965.g003

The results clearly show that the digital maturity determinants in the CEE countries are low versus the average values across the EU. The highest results among the CEE countries were obtained by the Czech Republic, Slovenia, Estonia, Croatia, and the lowest results by Bulgaria and Romania.

4.2. The assessment and ranking of digital maturity among enterprises in the CEE countries

The use of the presented MCDM methods (TOPSIS, MOORA and VIKOR) allowed for the ranking of studied CEE countries in terms of the level of digital maturity among their enterprises. All diagnostic variables adopted for the study were stimulants.

The values of weights for these determinants, calculated by the entropy method (Eqs 24–28), are presented in Fig 4, while the total values of weights for each dimension are summarized in Table 9.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Values of weights for the determinants of digital maturity for all and manufacturing enterprises in the CEE countries.

https://doi.org/10.1371/journal.pone.0253965.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. Total values of weights for individual dimensions of digital maturity of enterprises.

https://doi.org/10.1371/journal.pone.0253965.t009

The weights calculated for individual dimensions of digital maturity indicate that the dimensions Internet of things, big data analytics and artificial intelligence are the most important criteria for assessing the digital maturity of enterprises in the CEE countries.

The obtained values demonstrate that the highest weights should be adopted for the determinant “analysis of big data internally using machine learning” (artificial intelligence dimension) and “analysis of big data from smart devices or sensors” (big data dimension), and the lowest weights for “enterprises which have ERP software package to share information between different functional areas” (integration of internal processes dimension) and “purchase of cloud computing services used over the Internet” (cloud computing dimension). The weights determined by the entropy method for individual determinants of digital maturity define the degree of their disorder as a set of data, namely the degree of its uniqueness within the examined set of countries.

In the next stage of the analysis, ordering indices were determined for each method, i.e.: the index in the MOORA method, the Pi index in the TOPSIS method and the Qi index in the VIKOR method. The results are summarized in Table 10. Based on the values of these indices, a classification (ranking of countries) was made in terms of the level of digital maturity in all and manufacturing enterprises in the CEE countries belonging to the EU. The results of this classification are shown in Figs 5 and 6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Classification (ranking) of the CEE countries in terms of digital maturity among all enterprises using the MOORA, TOPSIS and VIKOR methods.

https://doi.org/10.1371/journal.pone.0253965.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Classification (ranking) of the CEE countries in terms of digital maturity among manufacturing companies using the MOORA, TOPSIS and VIKOR methods.

https://doi.org/10.1371/journal.pone.0253965.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 10. Values of indices classifying the digital maturity among all and manufacturing enterprises by the MOORA, TOPSIS and VIKOR methods for the CEE countries.

https://doi.org/10.1371/journal.pone.0253965.t010

The results of the classification (ranking) of the CEE countries show that for some countries (alternatives) the findings are the same or very similar, and for others–different. For example, Slovakia in terms of digital maturity among all enterprises in the ranking for all methods was found to be on the 6^th place, while the Czech Republic—using the MOORA method–on the 5^th place, using the TOPSIS method–on the 4^th place, and using the VIKOR method–on the 3^rd place. The differences between the rankings for individual methods for all enterprises are shown in Fig 7A and for manufacturing enterprises in Fig 7B.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Absolute differences in the rankings between the methods MOORA-TOPSIS, MOORA-VIKOR, TOPSIS-VIKOR (a–all enterprises, b–manufacturing enterprises).

https://doi.org/10.1371/journal.pone.0253965.g007

The analysis of differences showed that complete correspondence for the ranking of digital maturity among enterprises between the applied methods was observed for Slovakia, and for manufacturing enterprises–for Romania and Hungary. In general, the magnitude of differences in the rankings was found to be insignificant (difference in one or two positions). The difference in 3 positions in the ranking was reported for Estonia (Fig 7) and Croatia, yet only for manufacturing companies (Fig 7B).

The results of the ranking of the CEE countries in terms of the digital maturity among all and manufacturing enterprises showed similar results for most of the countries, and only for several were they different. In order to obtain a single reliable ranking of studied countries, the mean-rank method was utilized to combine the results obtained using the MOORA, TOPSIS and VIKOR methods. The mean-rank method seems to be, in this case, most simple and objective, as it can provide an unambiguous result [78,79]. The ranking for 11 countries in terms of digital maturity among all and manufacturing enterprises is shown in Fig 8.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. The ranking of the CEE countries in terms of digital maturity among their enterprises by the mean-rank method.

https://doi.org/10.1371/journal.pone.0253965.g008

Thus, the final ranking of the CEE countries in terms of digital maturity is as follows:

• for all enterprises: Estonia > Lithuania > Croatia > Czechia > Slovenia > Slovakia > Latvia > Poland > Hungary > Bulgaria > Romania,
• for manufacturing companies: Czechia > Slovenia > Estonia > Lithuania > Slovakia > Croatia > Poland > Latvia > Hungary > Romania > Bulgaria.

The results showed that in terms of the level of digital maturity among all enterprises of the CEE countries, Estonia and Lithuania were found to perform best, and Bulgaria and Romania were found to perform worst. In addition, Bulgaria and Romania were also observed to have the worst results in terms of the level of digital maturity among manufacturing enterprises. Within this group of companies, the best performers were the Czech Republic, Slovenia and Estonia.

Based on the mean values and standard deviation (12–15), the digital maturity classes of the CEE countries were also determined. These classes for individual countries were determined for each method separately (Fig 9), and then using the mean-rank method, the levels were evaluated (Fig 10).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Assessment of the level of digital maturity among enterprises in the CEE countries (a–all enterprises, b–manufacturing enterprises) (1 –high level, 2 –average high level, 3 –average low level, 4 –low level).

https://doi.org/10.1371/journal.pone.0253965.g009

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Assessment of the level of digital maturity among enterprises in the CEE countries by the mean- rank method (1 –high level, 2 –average high level, 3 –average low level, 4 –low level).

https://doi.org/10.1371/journal.pone.0253965.g010

Countries that were found to have a high level of digital maturity (according to the mean-rank method), for all enterprises (over 10 employees, excluding the financial sector), included Estonia and Lithuania. However, when analyzing the level of digital readiness of enterprises only from the manufacturing sector, it turned out that the Czech Republic was shown to have a high level.

An average high level of digital maturity among all enterprises was reported for the Czech Republic, Croatia and Slovenia, and among manufacturing enterprises for Estonia, Lithuania and Slovenia.

An average low level among all enterprises was found for Latvia, Slovakia, Poland and Hungary, and among manufacturing enterprises for Hungary, Poland, Slovakia, Croatia and Latvia. On the other hand, a low level of digital maturity among all and manufacturing enterprises was reported for Romania and Bulgaria.

For most CEE countries, the level of digital maturity among all and manufacturing enterprises was found to be the same. The exception was the Czech Republic, being the leader in terms of the level of digital maturity among manufacturing enterprises, and in the case of all enterprises, the level of this maturity was defined as average high. The opposite situation was reported for Estonia and Lithuania, being the leaders in the level of digital maturity for all enterprises, and for manufacturing enterprises, the level of digital maturity was assessed as average high.

In order to illustrate similarities between the CEE countries in terms of digital maturity characterized by the set of 14 digitization determinants, using the multidimensional scaling method, scatter plots in a two-dimensional space were prepared (Fig 11). The quality of matching the output to the input data was measured using the STRESS function, the results of which are presented in Table 11. Also, these graphs include the value for all EU countries, which represents the average level of digital maturity among enterprises across the EU.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 11.

Similarities between the CEE countries in terms of digital maturity characterized by the set of 14 determinants in a two-dimensional space (a–all enterprises, b–manufacturing enterprises).

https://doi.org/10.1371/journal.pone.0253965.g011

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 11. The value of STRESS function.

https://doi.org/10.1371/journal.pone.0253965.t011

The results of the analysis presented in Fig 11 indicate that the distances between the CEE countries in terms of digital maturity among all and manufacturing enterprises vary significantly. In general, the results show that they have little similarity in terms of digital maturity. However, in the case of digital maturity among manufacturing companies, a slightly greater narrowing of the borders between these countries can be seen, which means that the disparity between them in this regard is smaller. Changes in the distance between countries in terms of digital maturity among all enterprises versus manufacturing enterprises can be interpreted as changes in digital transformation occurring faster in manufacturing enterprises than in other sectors of the economy. In general, the closer countries are in a two-dimensional space to each other, the more similar they are (the more distant they are, the more dissimilar they are).

It can also be observed that Poland and Slovakia show high similarity in terms of digital maturity both among all and manufacturing enterprises. In terms of digital maturity among all enterprises, Croatia, Estonia, Slovakia, the Czech Republic and Lithuania were found to exceed the EU27 average for most determinants, while Estonia, Slovakia and the Czech Republic were found to exceed the EU27 average for manufacturing enterprises. The worst results in terms of digital maturity versus the EU average were reported for Hungary, Romania and Bulgaria.