The Whole Picture

Use phenomenon-mechanism analysis to understand a problem in its totality

by Ramaswamy Ganesan

The word "phenomenon" means the outward manifestation of an abnormal event.1 An abnormal event could be a defect, breakdown or malfunctioning of a machine. When a phenomenon occurs due to chronic causes,2 the conventional problem-solving approach may be the cause and effect (CE) diagram. This method has shortcomings, though, because it ignores causes that are insignificant individually, but significant in combination.‘

A suitable method in this situation is phenomenon-mechanism (P-M) analysis.3 It works similarly to process failure mode and effects analysis (PFMEA) and allows users to solve the problem in totality.

Online Figure 1 shows the analysis of the problem of poor bicycle braking through a conventional CE diagram. There are two main downfalls of the CE diagram approach: Only reasons that are presumed to be relevant and predominant to the problem will be listed. The causal factors that can contribute to the problem silently on a secondary level may be ignored.

Online Figure 1

There could be two reasons that, individually, don’t have an effect, but together, may cause the problem. For example, brake shoe material or insufficient force may not be a problem independently, but together, they cause poor braking. This type of combination might not be considered in a traditional CE diagram.

Now, consider how P-M analysis tackles this problem. The first step in a P-M analysis is defining the phenomenon. In this case, it is "poor braking." Through a physical analysis, you determine the mechanism that causes the phenomenon in this case as "the frictional resistance between the rim and the brake shoe is inadequate." Conditions that cause such a mechanism are called constituent conditions.

The constituent conditions are further expressed in terms of the primary and secondary 4Ms—machine, method, man and material. The standard value and respective measured value for each constituent condition and 4M is compared and corrected in case of abnormality. For example, the rim surface finish is to be 0.5-0.6 micrometer (µm) roughness average (Ra). If it’s found to be 0.4µm Ra, it’s an abnormality.

Each of the 4Ms will have quantitative or qualitative standard value. Constituent conditions may or may not have standard values. If the standard value is available for the constituent condition itself, it may reduce the work of enumerating its 4M condition. So, using this method, the phenomenon is physically analyzed through the mechanism of the equipment in terms of 4M, and the problem is resolved.‘

How a P-M analysis compares to the conventional CE problem-solving method can be seen in Online Figure 2. The main difference is that in the CE method, you proceed to the causes directly from the phenomenon. In P-M analysis, you first understand the mechanism of the phenomenon, followed by the constituent conditions and the causes.

Online Figure 2

P-M analysis is a technique for problem solving, and PFMEA is a technique for identifying problems that have not occurred yet. The phenomenon and physical analysis in P-M analysis are similar to the effect and failure mode in a PFMEA. Online Table 1 shows a P-M analysis solving the problem of poor braking. Note that physical mechanisms of defects can be inferred through failure modes in a well-made PFMEA (Online Table 2). It may be a good idea to introduce a column for constituent conditions in PFMEA.

Online Table 1

Online Table 2

P-M analysis applies only for quality problems or breakdowns in a particular machine. A full understanding of the machine and mechanisms is required to carry out P-M analysis. Conventional approaches of problem solving may be used to the extent possible, and then P-M analysis may be used.


  1. Kunio Shirose, Yoshifumi Kimura and Mitsugu Kaneda, P-M Analysis, CRC Press, 2004.
  2. Kenneth S. Stephens, ed., Juran, Quality and a Century of Improvement, ASQ Quality Press, 2004.
  3. Shirose, P-M Analysis, see reference 1.

Ramaswamy Ganesan is a quality and management consultant and former advisor at Axles India Ltd. in Sriperumbudur, Chennai. He has a master’s degree in mechanical engineering and production engineering from Madras Institute of Technology in Chrompet, Chennai. Ganesan is a member of ASQ.

Mr. Ganesan makes an excellent point that "there could be two reasons that, individually, don’t have an
effect, but together, may cause the problem."

In my experience we don't often systematically use the available process knowledge (i.e. mechanisms) to guide us in developing hypotheses for the occurrence of a phenomenon. The traditional C-E method doesn't explicitly nudge us to do it, and we forget.

I was so impressed by the piece that I bought "P-M Analysis," to develop a deeper understanding of the method.
--Shrikant Kalegaonkar, 06-12-2017

Average Rating


Out of 1 Ratings
Rate this article

Add Comments

View comments
Comments FAQ

Featured advertisers