What Would Deming Do - Incident Management?

What would Dr. Deming do if he walked into a bank in 2022 and was asked to assess the handling of incidents? He might have recalled what Dr. Shewart did in 1924 when he conceived of Statistical Process Control (SPC). He would assert that quality problems should be solved using mathematics. Defects were inspected manually and categorized into go/nogo groups before Shewhart's SPC. Based on the specifics of the part and the acceptable tolerance levels, there could be over two categories.

Currently, most banks manage incidents on a per-incident basis. They categorized a typical incident into priority categories (e.g., P1-P4). They then placed the incidents in a queue and handled them individually. Even worse, most organizations lack the resources to manage P3 and P4 incidents, so they never get analyzed. They learn nothing from these incidents. Because of their isolation, they also do not learn from the P1 and P2s. Furthermore, most of them are closed and removed from the incident queue without any strategy for systemic learning. Adaptive Capacity Labs' (ACL), John Allspaw, says incidents are investments. ACL uses a unique qualitative approach that I believe Dr. Deming would also find very compelling.

Dr. Deming frequently complained about the tyranny of prevailing management styles, and here is no exception. Dr. Deming would first ask whether the current system effectively identifies sources of resilience or brittleness within the organization. The answer is no. Next, I imagine he would ask why they treat every incident as an individual defect instead of looking at the system. For understanding the system, he would suggest using statistical methods.

He always chose SPC; however, at the very least, he would suggest linear regression or another type of statistical analysis. For example, Deming adapted this approach many times in the Department of Agriculture, the Bureau of the Census, manufacturing quality during WW2, and quality initiatives in Japan and the west after WW2. Instead of analyzing each incident separately, he would use statistical tools to identify trends, processes, and anomalies across all four types of incidents, P1-P4. An analytical approach would enable long-term systems thinking instead of short-term firefighting. Using SPC or other statistical analysis techniques, an organization could understand incidents in process control and ones that are not.

An unstudied solution to a problem may yield immediate results in the right direction, yet in time bring disaster.

Deming, W. Edwards. The New Economics for Industry, Government, Education, third edition . MIT Press.

This could lead to budgetary, training, product, compliance, and regulatory control discoveries across the whole spectrum of an organization's system. What would Dr. Deming do if I asked him to look at modern incident management? He'd probably get furious.