Fig. 1. Four quadrant diagram of cause and effect.
According to the distribution law of indexes in the four quadrants of cause and effect diagram, it can be concluded that C9, C24, C21, C12, C22, C8, C5 in the first quadrant have a centrality greater than 4.41, which belong to the driving indexes with a relatively large degree of influence on the supply chain data governance ecosystem and play a direct driving role in the sustainable development of the governance system; C16, C11, C2, C18, C17, C17 These six indexes have a higher degree of cause although their centrality is lower than 4.41, and they can be used as support indexes for optimizing the governance system and show a stronger indirect driving effect on the governance system. Although C13, C15 and C15 are in the third quadrant, they should be included in the selection of key indexes for the optimization of the supply chain data governance ecosystem because their centrality is close to 4.41 and the absolute value of the cause degree is larger. Based on the aforementioned influence degree analysis, the six indexes in the fourth quadrant were excluded.
In summary, through the comparison and analysis of impact index ranking and cause-effect four-quadrant diagram, 16 key indexes for supply chain data governance ecosystem optimization are finally screened out, namely C1, C2, C5, C8, C9, C11, C12, C13, C15, C16, C17, C18, C21, C22, C23, C24. to further study the structural system of supply chain data governance ecosystem indexes and provide quantitative thinking for the design of optimization paths.
4.2. ISM methodology
4.2.1. Building the reachable matrix
Based on the 16 key indexes screened above, the overall impact matrix is established. Given that the comprehensive impact matrix obtained by DEMATEL method does not consider the impact of indexes themselves, the comprehensive impact matrix formed by the 16 key indexes is added with the unit matrix to obtain the overall impact matrix , and then determine the reachable matrix of supply chain data governance ( = ), which is used to portray whether there is a pathway between the fixed points of the directed graph, if there is a pathway between indexes fi and fj , then =1; otherwise, then =0, expressed as follows.
Where, is the threshold value. It is set up with the aim of eliminating less influential indexes, thus simplifying the hierarchical structure of the indexes. To eliminate the subjective dependence of data sources, the study relies on the mean of the combined impact matrix\(\mu\) with variance\(\sigma\) for \(\lambda\) In order to eliminate the subjective dependence of the data source, the study relies on the mean and variance of the composite impact matrix. Matlab calculates = 0.091, = 0.019, then = + = 0.11. According to the formula, we obtain the reachable matrix of supply chain data governance optimization indexes (Table A3).
4.2.2. Hierarchy of indexes
According to the reachability matrix, the indexes are divided into a hierarchy, and the index reachable set, the prior set , and the common set are obtained. If=, then is extracted as the first hierarchical index set. The results of the first level decomposition are as follows.