It is indeed an honor to have both brothers Bob Latino and Mark Latino join me for a lightning round discussion on failure analysis. Failure is part of our personal and professional lives. Failure can come in the form of bad equipment, bad processes, or human factors. All root cause analysis of failures requires introspection.
The Latino family has been involved with root cause analysis and reliability engineering for more than 50 years. Father Charles Latino build the first reliability department in the US in 1967 at Allied Chemical. He founded the Reliability Center in the 1980s. Siblings Mark, Diane, Julie, Bob, and Ken followed in the family tradition and became experts in their own rights in Root Cause Analysis (RCA) and Reliability Engineering (RE). Their trademark RCA approach is called PROACT. Later this year, this company will release a universal Root Cause Analysis software that is independent of the PROACT methodology.
Bob specializes in higher-level facilitation, involving aspects related to human factors and human performance. Mark is much more a metallurgical specialist who also does facilitation and training. Ironically, both have more experience in facilitation, physics of failure, human factors, and training than most experts who only specialize in one area.
I have a focus on high-level facilitation of problems, and failures, associated with organization structure, business processes, human factors, and physical components. I chose to moderate this session while both adding some insights and egging-on the Latino brothers.
Is there usually a single root cause of failure? (and how have the different ways of describing root causes confused the non-expert)
BL: Our biggest competitor is our customers’ definition of root cause analysis. There are a lot of people who believe one root cause is the way to go — the 5 whys crowd — and there are other people who know better
ML: I think there are some things that confuse people. The 5-Whys tends to confuse people because the why, why, whys get to only one cause. Others have casual factors. We call it something different.
BL: A lot of people just view RCA as a commodity. These are the people that believe there is one root cause.
ML: Most people have a job already and they have RCA added to them. Many times, they are just looking for something simple. Often too simple.
BL: I don’t blame the people who do a poor analysis as I do those who accept it.
Can anyone perform a root cause analysis?
BL: We are brothers and our third brother is Ken. I hope he sees this. If Ken can do it, anybody can do it! (ha!)
ML: Anybody can do a root cause analysis — based on your definition. The differences are the depth and breadth of what and how you do it.
BL: Humanity makes you a root cause analyst. Yes, we do it in our everyday life. If you are using deductive analysis you are doing some form of root cause analysis.
ML: One criterion is to pick someone who wants to do it. If someone wants to do RCA well, they can do it well.
BL: We never have the time and budget to do it right, but we always have the time and budget to do it again.
What are the minimum types of training that you recommend before someone leads a root cause analysis?
ML: All of it. Physics of failure, decision making, human factors, communication.
BL: Our greatest competition is people’s definition. If people think it’s just the physics of failure — Root Cause Failure Analysis (RCFA)– then that is all they will train in. If you believe it is more than the physics of failure, then you will train in other aspects.
ML: Everything in RCA is about getting people to make the right decisions.
BL: The big difference in RCFA and RCA is getting the “F” out!! RCA is much broader, including the human and organizational systems aspects, and people must be trained in the other areas.
There are several commercial systems for performing a root cause analysis, including yours. Is buying a software product necessary for completing a root cause analysis?
BL: No, that would presume the software does all of the thinking for you.
ML: The software is not required for success. The software makes it easier and faster to do the analysis
BL: A software tool is only helping you to manage an investigation, to document, to communicate, to share. It makes the process more efficient.
ML: The software also helps you trend the results. You can trend the global results together and share the trends.
BL: An analysis is only as good as the analyst.
Should we do a root cause analysis for all failures? Are there gradations?
ML: When we started doing RCAs a long time ago, we did them on everything because we did not have a proven criterion. Then we got away from the fifty-dollar failures. With PROACT we started using $300,000 as a cut-off.
BL: The ones under $300,000 are with the “hand you are dealt”. Not very good because you are relying on opinion instead of making time to collect the data, but maybe sufficient for the resources you have.
ML: In a base failure analysis we still want to learn, but the cost factor may not be the driver.
BL: No company can do them on everything. Usually, organizations develop compliance type triggers around costs, lost lives, or some other values. That is not always where your money is.
ML: Understanding what is happening is important in many applications.
BL: People need to be able to make the business case for RCA on chronic failures. In our latest book, I use an example of the thousands of times a hospital needs to redraw blood. There was no trigger on a $300 redraw but it happened 10,000 times each year. That was $3M+ hidden in plain sight! That’s a lot of money laying around based on frequency.
How does an RCA team know when to call in an expert to do advanced testing or evaluation?
BL: When I am not 100% knowledgeable in a certain area, I call in an outside expert.
ML: I come at this as a metals-type expert, but as a facilitator I still call in specialists when we are uncertain. We recently did this on a chemical-related RCA in Virginia even though we are experts in some of the areas.
BL: You are not always going to have all the resources you need on a team, no matter how cross-functional it is.
ML: The facilitator is important in making these decisions. They must be well-rounded and humble enough to know when to call in for assistance.
What is the role of human factors?
BL: It is very important. There is a lot of people that think it is either the physical side OR the social side. It is hard to get people to understand that they are totally dependent on each other.
ML: I spend a lot of time on the human side. Usually, the physical side prompts me to look at what people are doing.
BL: Social scientists suck at the physical sciences and those that excel at physics typically suck at the social sciences.
ML: Most of the time, mechanics and millwrights do what they are told. They shape their behavior around the procedures and what their managers want.
BL: You need both social sciences and physical sciences to produce an effective RCA. They are dependent on each other.
Both of you have worked in many different fields. What has surprised you most when it comes to root cause analysis?
BL: I am surprised at what people think RCA is — that so many people see it as a checklist.
ML: Most managers think RCA is something that fixes everything. They think some training and doing the process just fixes the problem.
BL: Managers do not know how to measure RCA effectiveness. Which means they don’t expect much from it and do it only for a regulatory purpose. We see it differently — RCA should produce value. Managers should clearly set that expectation of value.
ML: A second one is how many people do not follow the methodology. Then they make it fit into the RCA methodology…they do their own thing and call it ‘RCA’.
What is the best single tip that you would give someone who is preparing to lead an RCA or failure-related investigation?
ML: Go to wherever the problem occurred, do a first pass observation without any team, and do an evaluation. Then the facilitator can decide who they need on the team and make a second pass.
BL: Do it right or do it again. If you stop short for whatever reason — time, resources, politics — the problem will happen again, and you will be doing the RCA again. At some point, a progressive management would question why someone was getting so much RCA practice!
ML: My second one is that you have an answer for everything in your logic tree. Anything less than a “5” in terms of confidence factors (strength of evidence), you have to be able to explain before you go in front of management and they start shooting holes in it.
BL: Hearsay does not fly as fact in RCA…contrary to popular belief!
1. Whenever you are in the presence of greatness, sit back and listen.
2. RCA is painful. If you do it right, it is introspection.
3. We often try to find what management wants to hear. That is not helpful, but that is often how people survive and keep their jobs.
4. We run through too many cycles of fully integrating, and then not integrating, human factors into RCA. RCA cannot be done effectively without integrating physical factors and human factors.
5. We often outthink ourselves and get too philosophical — in one camp or another. Most things, including Root Cause Analysis, comes down to doing the basics, well.