Keywords

1 Introduction

In the global economy, apart from factors such as raw materials, equipment, machines, finance, energy, land and cheap labour, which determine the competitive capabilities of a company in the traditional economy, additional resources play an important role. They include an intellectual capital, i.e. employees, their knowledge, competences, attitudes and ideas. The creation of this capital requires sophisticated tools and specialists with high competences. Hence, the gathering of knowledge and the ability to use and maintain it seem to be one of the greatest challenges faced by the economy today. In some industries, such as IT, telecommunication, pharmacy, consulting, education, this resource turns out to be so important that managers are looking for methods to determine its value [11, p. 126]. This means that a company has to develop systematic rules and models for knowledge management, which in the modern global economy are the main determinants for the space of freedom for their survival and development as well as an important source of competitive advantage [14]. The previous findings of practitioners and theoreticians of management sciences, which make the knowledge management the main discipline responsible for creating the wealth and employment across all industries, support this thesis [4, p. 140]. Knowledge management, especially in recent years, has been becoming one of the most developed research disciplines, mainly for practical reasons. Knowledge is a goal and a mean to reach the goal. It must be skilfully acquired, explored, exploited and gathered in knowledge bases and employees’ minds. Moreover, companies must also demonstrate their capabilities to manage knowledge, including the ability to retain it [2, p. 71].

2 Knowledge Management in a Company

Depending on the potential of a company, its current position on the market and strategic intentions, it creates and mobilizes knowledge and competence resources of various degree of significance for the functioning of a company.

The view that the condition for an effective operation of a company is its ability to survive, adapt and develop [6, p. 68] and that the ability to identify the environment and shape the company’s activity in accordance with the requirements imposed by it provides a basis for functioning of each company and is the main determinant of its successes [16] has solid grounds in the literature and organizational practice. If we accept this view as correct, it can be concluded that the company should reap benefits from various categories of knowledge. These categories include basic knowledge for the needs of rudimental operations, advanced knowledge that allows gaining a competitive advantage, and innovative knowledge which ensures the position of a leader and is full of solutions that other market participants do not have.

The extent to which companies are able to use their knowledge is determined on the one hand by the abilities to enrich the knowledge possessed and to acquire new knowledge, while on the other hand by the ability to create conditions for the needs of appropriate rationing (distribution) and further maintenance of such knowledge or/and its transfer. This requires both overcoming the barriers in appreciating the importance of knowledge management and creating the conditions for enhancing the ability to bridge the gap between the creation of knowledge resources and their use in the company. Based on a literature query, a list of barriers in knowledge management can be created. These are:

  • incoherence of the management staff, who in fact thinks only about its position, territory and influences [13], [3, p. 33],

  • insufficient use of employees’ knowledge [22, p. 33], [15, p. 32],

  • information problems and shortage of knowledge resulting from an uneven distribution of knowledge in terms of needs and requirements of all the company’s units and employees, which in turn generates problems of an informational natureFootnote 1,

  • difficulties in acquiring informal knowledge [18, p. 46],

  • inadequate organizational culture and a lack of an atmosphere to share the knowledge openly Footnote 2,

  • overestimation or underestimation of the role of technique and technology, especially of the IT infrastructure [16, p. 29],

  • useless databases, which should be eliminated basing on the criterion of practical usability [19, p. 29],

  • stereotypical thinking that knowledge means power [19, p. 30],

  • a gap between the concept and the action [21, pp. 11–15],

  • rejecting (not taking into account) negative knowledgeFootnote 3.

  • limiting contacts and informal conversationsFootnote 4,

  • limiting to cognitive knowledgeFootnote 5,

  • transferring knowledge through a single medium [21, p. 30].

With regard to the factors conducive to proper knowledge management, apart from acquiring, accumulating, processing and creating knowledge to a possible broad extent, the company must create conditions for its exchange and transfer or retention with the intent to multiply its value (quality) [10].

To meet these requirements, the company’s management board should:

  • put a considerable emphasis on education and training of employees [9, pp. 99–169],

  • continuously develop the intellectual curiosity of employeesFootnote 6,

  • avoid stereotypical thinkingFootnote 7,

  • consciously create mechanisms for identifying and developing the factors conducive to appearance of systemic effects of knowledgeFootnote 8,

  • create conditions for collective learningFootnote 9,

  • create conditions for developing the creativity and innovation as it takes place at RANK XEROX [20, pp. 20–22],

  • develop knowledge systems (databases and networks) and rules of their use [20, p. 20],

  • manage competences in the field of the company’s human resource function as it is done in NUCOR STEEL CORPORATION – an American metallurgical company [23, pp. 11–17].

The K’NETIX network in a French Buckman Laboratory is an example that petrifies the findings of this part of the considerations, illustrating a manner of knowledge management. These are perfectly organized knowledge resources, which can be widely used by employees [1, pp. 415–416].

3 Verification of Findings

Knowledge management in a company can be examined in four dimensions determining the space of freedom for this processFootnote 10:

  • degree of knowledge organization (codification),

  • intensity of spreading the knowledge and its renewal,

  • ability to absorb (use) the knowledge,

  • skill to retain (maintain) it.

As it appears from the above, a useful instrument that effectively supports the use of the intellectual capital in a company is the development and use of a knowledge system, taking into account those dimensions, in which a requirement to build the loyalty is permanently embedded and without which effective knowledge management will not be possible.

A correctly functioning knowledge system supported by a consistent strengthening of loyal behaviours and attitudes of employees requires identification of the premises guiding the employees who decide to stay with the company and work for it for a long period of time. Then, these premises should be used to develop an incentive (reward) system that will allow not only satisfying the expectations of employees and make them stay with the company and work for it with engagement, but will also allow building their loyalty towards the employer and thus preserving the knowledge resources and the intellectual capital. In order to substantiate the thesis formulated in this way, there was presented an experiment carried out by the authors of this study, which allowed identifying the premises that guide the employees when they decide whether to stay with the company.

3.1 The Subject of the Analysis and Presentation of Data

When analysing the literature on the subject, components of the intellectual capital were identified [1] as well as important factors which can significantly affect the development and retention of knowledge in the organization [2]. Based on these factors, the criteria – components of the model of retention and accumulation of knowledge in the company were established. In order to verify the model, data on the American economy from the period of 1995–2015 were collected and used. They are presented in Table 1. The principal component analysis method (PCA) was used in the study, while the calculations were carried out using GRETL statistical software. The criteria were formulated in the following form:

  • X1 – expenditures on human capital, expressed in average annual remuneration (per capita),

  • X2 – expenditures on R&D on an annual basis (per capita),

  • X3 – expenditures on marketing activities (annually, per capita),

  • X4 – expenditures on PR activities (ratio of annual volume to the total number of business entities registered in the USA),

  • X5 – annual value of payments of retirement benefits resulting from Social Security (per capita, excluding the 401(k) plan).

Table 1. Input data of the model
Full size table

All numerical values presented in Table 1 are expressed in thousands of US dollars and in constant prices of 2015.

3.2 Description of Research Tool

Principal component analysis (PCA) is used to determine new variables, a possibly small subset of which will provide as much information as possible about the whole variability in the data set. The new set of variables forms an orthogonal base in the space of features. Variables are selected in such a way that the first variable maps as much variation in the data as possible. After determining the first variable, another one is determined so that it is orthogonal to the first one and explains the remaining variability as much as possible. The next variable is selected in such a way that it is orthogonal to the first two ones etc. The set of vectors obtained in this way forms an orthogonal base in the space of features. The purpose of the principal component analysis method is therefore to find a transformation of the coordinate system that will describe the variability between observations in the best way [12].

PCA method maximizes the variance of the first coordinate, then the variance of the second coordinate and the next ones. The coordinate values transformed in such a way are called loadings from generated factors (principal components). In this way, a new observation space is built, in which the initial factors explain the most of the variability. So, the aforementioned operation can be used to understand the structure of the population studied and the nature of the data used. PCA method may be based on a correlation matrix or a covariance matrix created from the input set.

The algorithm in both versions is identical, but the results are different. When a covariance matrix is used, the input set variables with the largest variation in have the greatest impact on the result, which may be advisable, if the variables represent comparable, relatively uniform values. In turn, the use of the correlation matrix corresponds to the initial normalization of the input set, so that each variable has an identical variance at the input (without weights), which may be advisable, if we are unable to ensure the comparability of values of the variables tested.

The PCA algorithm consists of the following steps based on [5]:

  • Determination of means for rows

    $$ {\text{u}}[{\text{m}}] = \frac{1}{N}\sum\limits_{n = 1}^{N} {X[m,n]} $$
    (1)

  • Calculation of the deviation matrix

    $$ X^{\prime}[{\text{i}},{\text{j}}] = - {\text{X}}[{\text{i}},{\text{j}}] - {\text{u[i]}} $$
    (2)

  • Determination of the covariance/correlation matrix

    $$ {\mathbf{C}} = {\mathbb{E}}\left[ {{\mathbf{B}} \otimes {\mathbf{B}}} \right] = {\mathbb{E}}\left[ {{\mathbf{B}} \cdot {\mathbf{B}}^{\varvec{*}} } \right] = \frac{1}{N}{\mathbf{B}} \cdot {\mathbf{B}}^{*} $$
    (3)

  • Calculation of eigenvalues of the covariance/correlation matrix

    $$ {\mathbf{V}}^{{ - {\mathbf{1}}}} {\mathbf{CV}} = {\mathbf{D}} $$
    (4)

  • Selection of eigenvalues (in order to minimize losses in information, the ones with the highest value are selected)

  • Determination of eigenvectors

    $$ \left[ {\begin{array}{*{20}c} {a_{11} - \lambda } & {a_{12} } & \cdots & {a_{1n} } \\ {a_{21} } & {a_{22} - \lambda } & \cdots & {a_{2n} } \\ \vdots & \vdots & \ddots & \vdots \\ {a_{n1} } & {a_{n2} } & \cdots & {a_{nn} - \lambda } \\ \end{array} } \right] \cdot \left[ {\begin{array}{*{20}c} {x_{1} } \\ {x_{2} } \\ \vdots \\ {x_{n} } \\ \end{array} } \right] = 0 $$
    (5)

  • Projection onto eigenvectors

    $$ y = \left[ {\begin{array}{*{20}c} {y_{0} } \\ {y_{1} } \\ \vdots \\ {y_{n - 1} } \\ \end{array} } \right] = \varvec{V}^{T} \cdot x = \left[ {\begin{array}{*{20}c} {v_{0}^{T} } \\ {v_{1}^{T} } \\ \vdots \\ {v_{n - 1}^{T} } \\ \end{array} } \right] \cdot x $$
    (6)

where:

  • V – matrix of eigenvectors,

  • x – projected vector,

  • y – vector in the new space,

  • N – number of eigenvector.

Each principal component is therefore described by:

  • eigenvalue,

  • eigenvector,

  • factor loadings,

  • contributions of variables,

  • communalities.

3.3 Presentation and Evaluation of Quality of the Results Obtained

The results obtained with the GRETL statistical processor are shown in Tables 2 and 3. Since there is no single universal criterion for selecting the number of principal components, it is justified to use multiple criteria for this purpose [7, pp. 84–89]:

Table 2. Results generated by the GRETL statistical processor (1)
Full size table
Table 3. Results generated by the GRETL statistical processor (2)
Full size table

Percentage of Explained Variance.

The number of principal components, which a researcher should adopt, depends on the extent to which they represent primary variables, i.e. on the variance of primary variables contained in them. All principal components carry 100% of the variance of primary variables. If the sum of the variances for some first components constitutes a significant part of the total variance of primary variables, then these principal components can substitute the primary variables to a satisfactory degree. It is assumed that this variance should be reflected in the principal components in more than 80%.

Kaiser Criterion.

The Kaiser criterion says that the principal components, which we want to leave for interpretation, should have at least the same variance as any standardized primary variable. Due to the fact that the variance of each standardized primary variable is 1, according to the Kaiser criterion only the principal components with an eigenvalue that exceeds 1 or is close to it are valid.

Scree Plot.

The plot illustrates the rate of decrease of eigenvalues, i.e. the percentage of the variance explained. The point on the plot, at which this process stabilizes and the descending line becomes horizontal, is called the end of the scree (the end of downward trend of the information about the primary variables that is carried by principal components). The components located to the right of the end of the scree represent a negligible variance and mostly present a random noise.

In the light of the above criteria, the decision was made to leave the following variables in the model: X1 (‘expenditures on human capital’) and X2 (‘expenditures on R&D’) as fully representative and crucial for the explanation of the phenomenon. Since a satisfactory result was obtained, the plans to build a synthetic variable were abandoned.

The variables presented above can therefore be firmly considered as crucial for the retention and accumulation of knowledge in a company operating on the American market.

4 Conclusions and Future Work

Knowledge is subjected to processes of identification, acquisition, exploitation, exploration, development and retention – irrespective of the sources, from which a company draws knowledge, the character and characteristics of knowledge, as well as the conditions, needs, manners and skills that determine the possibilities of managing the knowledge. Each phase of this endless process is a part of a knowledge system that requires the use of appropriate tools to assist its management. This means that modern companies, irrespective of their phase of development, have to go a long way and overcome many hindrances to switch from a traditional model of managing to a knowledge-based management modelFootnote 11.

These phases accurately represent the problem that companies have to solve. The results of the research [8, p. 29] form a view that a vast majority of modern companies is still unaware that knowledge can be managed or has no experience or opportunities to do so.

In fact, the following quotation from a study by Zack refers to this situation: Although there is much talk about linking the knowledge management with the business strategy, in practice this is widely ignored [8, p. 30]. Finally, it is worth paying attention to two further limitations associated with the philosophy of knowledge management in a company. These limitations are associated with the organizational culture and thus with the system of values and the business model applicable in a given company. It seems that the rigid bureaucratic model with a fixed hierarchy and professional specialization is still prevailing, which hinders its diffusion. An additional, important factor limiting the absorption of the idea of knowledge management is the time pressure associated with quick reactions to the changes occurring in the company’s environment. On the one hand, employees, especially managers, do not have time to learn and, on the other hand, both groups do not have time to share their knowledge. If we add to these factors also a lack of willingness to share the knowledge and the connivance for leaks of knowledge outside the company, there should be a call for strong leadership, which should also include the responsibility for the protection of knowledge.

Will it be possible to adopt such a concept in the near future? The companies struggling with the global reality must respond to the problem formulated in such a way. Their response must take into account a number of perspectives, including the responsibility for knowledge, definition of requirements for the knowledge management system, and priorities to be adopted when implementing such a system. In other words, regardless of the type, character and degree of advancement (level) of knowledge, the company must continue efforts to multiply (quantity) and enrich (quality) it, as well as work on methods of its protection (retention). Otherwise, the issue of knowledge management in a company will remain a theoretical problem.