Part 1: https://bankingfrontiers.com/solving-product-design-puzzle-by-startups/
Part2: https://bankingfrontiers.com/solving-product-design-puzzle-by-startups-part-2/
Sandeep Sethi, Leadership Team at ICICI Bank
In the previous posts I emphasized that Startups should adopt at least 3 key Quality Tools before they feel they are ready to go full hog into the market. Even at the expense of repeating, adopting these 3 Quality Tools will prepare us for answering the following key questions:
- Is my product aligned to customer requirements?
- Mistakes will happen, how can I pre-empt what can go wrong and accordingly have a mitigating or quick response plan if things do not go as planned?
- Is my product only technically sound or is it easy to use also?
We discussed that most appropriate tools for these would be Quality Function Deployment (QFD), Failure Mode and Effects Analysis (FMEA) and Customer Experience Journey Mapping (CxJM). In the previous post we covered QFD. I won’t say we covered in detail, but we did cover enough using which one can make an attempt before finally watching Youtube videos and taking it to conclusion. In this post let’s improve our understanding on FMEA.
- Failure Mode and Effects Analysis (FMEA) – “Ek machhar saala aadmi ko hijda bana deta hai !”
Perhaps this Nana Patekar dialog from the 1996 movie Yeshwant provides best analogy to what a small mistake can do to bring to halt an otherwise great thing!! Doesn’t O-ring defect, which led to crash of Shuttle Challenger in 1996 qualify in the same category? The limitations, whether technical or process, or even creative genius of those using it (where normally those involved in creation do not even dream that an error like that can be made) provide enough opportunities to fail (Failure Modes). These failures can lead to Effects in the form of either minor irritants or even to extreme cases of serious injury or fatality. The success lies in anticipating (Analyzing) where all our product or process can go wrong and have a mitigating plan in place. Thus the name FMEA.
FMEA is used in many variations. Of these Process FMEA, Design FMEA and Product FMEA are most commonly used. The fundamental purpose of any FMEA, however, is to identify, evaluate and take actions to reduce risk of failure, hence the approach remains similar.
There are few more key terms: The intensity of effects is termed as Severity. Some failures occur more frequently than others. Thus Occurrence is a key term. And third key terms is Detection i.e what is the ability of current controls to detect if anything goes wrong. The final key term is RPN or Risk Priority Number. This number helps us prioritize our work. It is simply a multiplication of the three S-O-D.
[RPN = Severity x Occurrence x Detection]. All three (Severity-Occurrence-Detection) are rated on a scale of 1(min)-10(max). It is self-understood that Severity and Occurrence get high score if Severity is high or Occurrence is frequent. Can you guess how do we rate Detection? If detectability is low i.e. we have low ability to detect a failure, should we assign high rating or low rating? A low detectability in actual is more dangerous thing for the process. Moreover, if we assign low score, it will result into lower RPN and that cause will not become highlighted. This is against the overall objective of the approach. Hence Detection follows reverse rating logic. A low detectability is assigned high score and high detectability is assigned low score. Again a team activity, this yet another easy to understand tool, cannot be overlooked if we are planning to become big daddy in our field. It helps us ask questions and then answers following key issues:
2.1 Has our team applied brains to list where all our process can break down or the product can fail? This is to be done for each process step and component. These are called failure modes. Implied need here is to capture the process steps in detail rather than just capturing high level steps. So start with first listing the process step and against that list all the possible failure modes.
2.2 Do we understand the Severity associated with each failure? Mentioning the severity in qualitative terms helps.
2.3 Can we quantify the Severity? If we understand it to be more severe, naturally we would be
more cautious planning that. Example, severity rating of 10 would mean the failure could lead to someone’s death in case of an equipment failure. On the other hand, a simple discomfort will carry low rating.
2.4 How frequently is that failure likely to Occur? Capturing the information in terms of % or instances per month, etc. helps.
2.5 Can we assign Occurrence number to the failure? A higher number indicates a frequently failing part/ step. A frequently failing step/ part either needs to be changed altogether or failure in that area should get quickly addressed.
2.5 What are the current controls in place that will help detect/ check that failure? This obviously helps in alerting about areas where controls are missing?
2.6 How well are above controls likely to catch the failure from causing damage? As discussed above, if they cannot, the Detection rating will be high.
2.7 In listing down step wise or part wise Severity, Occurrence and Detection, we understand the
complete standing of our product/ process. Can we multiply S x O x D to arrive at Risk Priority Number?
Thus, theoretically, the maximum RPN can be 10x10x10 i.e 1000. You might like to separately list top 10 (or as appropriate) in descending RPN order. For improvement, start with the action plan on how you will reduce RPN for steps that have highest RPN number and keep moving down in descending order. A time bound action plan with clear ownership for deployment brings in results.
A simple template can be as follows:
One of the simplest examples that I picked up from internet (credits unknown) and use in my sessions to explain the concept to participants of diverse groups and functions is from the field of cricket, the game whole of India is crazy about, and hence able to participate equally. Alternatively, we can relate it to any game. While the actual exercise will have many rows (steps) in first column and accordingly the chart can run into a few pages, for simplicity in explanation the example picked here has only one row:
The action plan should actually mention responsibility and timelines also. Once that is closed, we should re-calculate new RPN. So consider above and below format to be placed side by side. Since extending the format above would have meant further reduced font size while reading when we adjust the complete format in single shot, that’s why I am taking liberty to continue the above format.
Just in case you haven’t noticed, against a failure mode, the effect of that mode and the severity associated with that remains same. However, there can be multiple causes leading to that failure and hence while creating FMEA on excel, merge cells till Severity rating as appropriate and have multiple rows thereafter depending upon the number of causes.
Few words of caution – Even if two teams are working on the same process FMEA in the same company, there can be difference in the steps listed as part of process, difference in the identification of possible ways in which a step can fail and difference in the understanding of controls. This would result in different output of FMEA. But even if the two groups are supplied with the same steps, same possible ways of failure or same controls, they can still arrive at different points to focus based on RPN values. This is because they might give different ratings of S-O-D. This is the limitation of FMEA approach. We can to an extent work on eliminating difference due to different S-O-D ratings. It is recommended that firstly a standardized rating sheet is prepared. This standardized rating sheet should be used for similar nature of processes.
The good news however is that even if the teams arrive at different RPN for same step, the difference is only notional because ultimate objective is to work on higher RPN and keep on reducing RPN as we move down in descending order of RPN.
A sample standard rating sheet that ‘can be used’ in financial services is as below:
Similarly, a startup in the area of physical product will have a different rating list. Just to bring home the point of differentiation, assume a startup that works on hardware or has made a physical product. A medium failure could lead to physical injury and ultimate failure could result in a fatality. So in such cases, the rating sheet shared above need to be re-created.
For ease of remembering, the whole FMEA process is summarized below:
Again, similar to QFD, this is a team activity and real benefit lies in maintaining it as a live document. Like QFD, I encourage you to pull out the template available on internet and go ahead. Even if you are not fully satisfied with the outcome, I am sure you will anyway be better than where you were earlier. Practice will help us become more confident in using this.
To be continued…
Currently, Sandeep is Group Head – People and Process Optimization, as General Manager. He has been in various leadership roles at ICICI Bank in the areas of Quality, Innovation, Infrastructure Facilities Management, Business Process Re-engineering, Branch Customer Experience & Service Improvement. A 6 Sigma MBB, he is also a Lean Manufacturing expert and brings rich experience of working in Manufacturing, Business Process Outsourcing and Banking while working with organizations like LG Electronics, Videocon, HCL Hewlett Packard, Wipro and now ICICI Bank.
Contact the author at [email protected]
