5.2.1. Identification of causes and receivers

First of all, the 15 experts pairwise evaluate the causal relationship structure among SFs on a Likert scale from 0 (no effect) to 4 (very strong effect). The individual evaluation matrices X<sup>k</sup> (with k = 1, ith) are described by Eq. (1) and are aggregated to the direct relation matrix A by Eq. (2). After that, the matrix A is normalised to matrix D according to Eqs. (3) and (4). With the help of Eqs. (5) and (6), the final total relation matrix D can be calculated. Table 1 shows the results which describe the direct influence intensity that SF i exerts on a SF j:

For more clarity in the causal structure, only these influence relationships between the SFs are considered in the IRM, of which the influence intensity is greater than the calculated threshold of α = 0.1426 {Eq. (7)}. In Table 1, the sufficiently significant results are marked in bold. Consequently, nearly half of the amount of causal relationships among the factors is specified as above-average causal interdependent and thus can be determined as very performancerelevant relation.


Table 1. Total relation matrix.


Table 2. Classification of the SFs as causes and receivers.

To organise the SFs in the groups 'causes' and 'receivers', the row sum ri and column sum ci as well as their difference have to be calculated in the subsequent step as follows:

According to Table 2, the SFs 'structural circumstances', 'product range', 'product quality' and 'ability to supply' are identified as causes, under the condition that ri > cj and thus the results of their difference are positive. However, the SFs 'financial success', 'competitive environment', 'pricing' as well as 'image' fulfil the condition cj > ri. As a result, the calculated difference between the row and column sum is negative and thus the factors are specified as receivers. Moreover, within the groups 'causes' and 'receivers' the SFs can be clearly ranked by their total influence intensity (ri + ci) in respect of their significance for the performance generation. It can be realised for the group 'causes' that the SF 'product range' is the most influencing factor, which largely determines all other SFs. In contrast the SF 'structural circumstances' has the lowest impact on the whole system. Considering the group 'receivers', the SF 'financial success' is mostly influenced by the other SFs compared to the SF 'pricing', which is less determined by the other ones.

### 5.2.2. Tailor-made impact-relation map

In this section the identified SFs and only their above-average calculated influence intensity from Table 1 as well as the group specification of receivers and causes from Table 2 is finally visualised in an appropriate IRM (Figure 4).

According to the coordinate system in Figure 4, the ordinate represents the difference between received and outgoing effects of a factor. Factors that can be characterised as causes, for example 'ability to supply', are always pictured in the positive range. Whereas, receivers like the factor 'pricing' are depicted in the negative value range. Furthermore, the abscissa displays the overall intensity of the influence relationship of an individual factor. The further away a factor is located from the coordinate origin, the greater its total influence intensity in the whole system is. Following Figure 4, the factor 'financial success' is the most performance-relevant factor in relation to the others.

Performance Management by Causal Mapping: An Application Field of Knowledge Management http://dx.doi.org/10.5772/intechopen.70297 115

Figure 4. Impact-relation map.

To organise the SFs in the groups 'causes' and 'receivers', the row sum ri and column sum ci as

Factors ri ci ri + ci ri ci Characteristic Financial success 0.9445 1.9988 2.9433 1.0544 Receiver Competitive environment 1.3897 1.4927 2.8824 0.1030 Receiver Structural circumstances 1.0446 0.3761 1.4207 0.6685 Cause Product range 1.3080 0.9082 2.2163 0.3998 Cause Product quality 1.3229 0.7980 2.1209 0.5248 Cause Pricing 1.0896 1.3280 2.4176 0.2385 Receiver Image 1.1164 1.5391 2.6555 0.4227 Receiver Ability to supply 0.9121 0.6868 1.5989 0.2253 Cause

According to Table 2, the SFs 'structural circumstances', 'product range', 'product quality' and 'ability to supply' are identified as causes, under the condition that ri > cj and thus the results of their difference are positive. However, the SFs 'financial success', 'competitive environment', 'pricing' as well as 'image' fulfil the condition cj > ri. As a result, the calculated difference between the row and column sum is negative and thus the factors are specified as receivers. Moreover, within the groups 'causes' and 'receivers' the SFs can be clearly ranked by their total influence intensity (ri + ci) in respect of their significance for the performance generation. It can be realised for the group 'causes' that the SF 'product range' is the most influencing factor, which largely determines all other SFs. In contrast the SF 'structural circumstances' has the lowest impact on the whole system. Considering the group 'receivers', the SF 'financial success' is mostly influenced by the other SFs compared to the SF 'pricing', which is less deter-

In this section the identified SFs and only their above-average calculated influence intensity from Table 1 as well as the group specification of receivers and causes from Table 2 is finally

According to the coordinate system in Figure 4, the ordinate represents the difference between received and outgoing effects of a factor. Factors that can be characterised as causes, for example 'ability to supply', are always pictured in the positive range. Whereas, receivers like the factor 'pricing' are depicted in the negative value range. Furthermore, the abscissa displays the overall intensity of the influence relationship of an individual factor. The further away a factor is located from the coordinate origin, the greater its total influence intensity in the whole system is. Following Figure 4, the factor 'financial success' is the most performance-relevant

well as their difference have to be calculated in the subsequent step as follows:

mined by the other ones.

5.2.2. Tailor-made impact-relation map

factor in relation to the others.

visualised in an appropriate IRM (Figure 4).

Table 2. Classification of the SFs as causes and receivers.

114 Knowledge Management Strategies and Applications

Generally, through the construction of an IRM, a better comprehension of the relevant direct and indirect causal relationships can be developed. Besides, the IRM underlines which SFs are most important for the corporate management and the focus should lay on them. Compared to a qualitative causal mapping process DEMATEL strictly distinguishes the SFs between causes and receivers and quantifies their cause-and-effect relationships [53, 54]. However, because the individual evaluations are ordinal a cardinal interpretation of the SFs´ causal relations cannot be provided. Only a systematisation by building a hierarchy among SFs is possible.

### 6. Conclusion

Causal knowledge on SFs underlying financial performance generation is an important prerequisite for an effective PM. For this purpose, important parts of the PM have to be drawn on the subjective experiences and knowledge of the employees. It is the current task for KM to extract the subject-bound tacit knowledge and make it explicitly available for the management of an organisation. Subsequently, by application of a convenient mapping method revealed tacit knowledge has to be aggregated, structured as well as systematised in a more general and for the employees' applicable manner. In addition, the complex financial performance generation process will be represented and analysed as for performance relevance of the SFs and their causal relations. In this way, a general and clarified understanding of the performance generation is achievable among the employees.

The concept of causal mapping and the multi-criteria DEMATEL method illustrate approaches how to construct a causal map from the base of externalised tacit knowledge. Both methods differ in procedures and in results. Causal mapping offers a low quality of the identified causal structures of SFs because of the lack of quantitative assessment and the highly subjective aggregation of the implicit knowledge. However, applying DEMATEL in the mapping context, the subjective bias can be minimised by a systematic and transparent pairwise evaluation of the SFs. Because of its replicability it achieves intersubjectivity. But since the discovered causal relationships among the factors are only interpretable on an ordinal scale, strategic forecasts of future performance developments are only possible to a limited extent.

To achieve an objective validity, the existence of adequate data and the use of suitably selected statistical procedures are necessary. If for all variables of the causal map manifest time series data are available, the validation of causal relationships can be done by using a multivariate time series model. When the variables of the causal map are not directly observable, but can be operationalised as latent variables with appropriate factors, structural equation modelling can be used to validate the cause-and-effect relationships among the SFs [29]. The statistical validation of the causal relationship network objectifies the previous ordinal data in metric forms to achieve relative comparability and clear predictability. So, the significance of the map is optimised compared to the one constructed by DEMATEL. Finally, in the context of PM, the performance realisation and generation can be represented and analysed qualitatively as well as quantitatively by the validated map in a comprehensible way.
