Simplifying Probabilities for Better Decision Making

Ever find yourself stuck trying to gauge the likelihood of an event?

What if you could transform your approach to probability assignments with just one simple trick?

On this episode of Quality During Design, we talk about simplifying probabilities for better decision making. We uncover a method to assign probabilities and occurrence ratings during preliminary assessments.

We share a technique that involves breaking down an event into smaller, more manageable parts, helping you understand and analyze it better. This method makes it easier to assign likelihoods because it provides better understanding of the event, clearer context, and consideration of what may drive things to happen.

Then, we relate these parts to conditional probabilities. We offer example explanations and practical applications to help you grasp conditional probabilities.

Plus, for those looking for additional resources, we have a cheat sheet that can further simplify these concepts for you. Not a subscriber yet? Visit qualityduringdesign.com to sign up and start receiving valuable insights directly in your inbox.

Join us to learn how these techniques can simplify probabilities, improve your assessments, and boost your team's decision-making.

Visit the podcast blog for more.

Transcript

Breaking Down Probabilities

Speaker 1 0:00

When we're doing preliminary assessments of things this could be on our own or with a team there's usually some sort of occurrence assignment that's happening. We're asked to provide an occurrence rating or a probability assignment or the likelihood what's the likelihood of this bad thing or this good thing happening? I've seen it time and again with myself, with other people observing other people making decisions, and then also in a team setting, where a team is trying to make a decision and assign some occurrence rating to something and this could be with hazard analysis, fmea and even project management. We can get stuck in trying to figure out what likelihood to assign to whatever it is we're assessing. We get stuck and we don't know how to move forward with that. I have a trick that I can share with you that I think will help. Listen in after this brief introduction. Hello and welcome to Quality During Design, the place to use quality thinking to create products. Others love for less. I'm your host, diana Deeney. I'm a senior level quality professional and engineer with over 20 years of experience in manufacturing and design. I consult with businesses and coach individuals on how to apply quality during design to their processes. Listen in and then join us. Visit qualityduringdesigncom.

Speaker 1 1:37

Probabilities and occurrence ratings, and some people like to call them likelihoods. They can be a bear to try to nail down. We're often asked to do assessments of what we think could happen. Well, that's what a probability is right. We're assigning a likelihood to an event, to something that could be happening. We may not have a whole lot of information, or sometimes a study or a test about this event is just impossible or perhaps infeasible infeasible in its setup, in the study length and even the cost of it. We're being asked to assign a probability to an event and we're stuck. I have a go-to solution, and it's generally my go-to solution for a lot of things it's to break it down into smaller parts so that we can better assess things. What's the old saying? There's only one way to eat an elephant one bite at a time.

Speaker 1 2:43

Well, when we're trying to assign probabilities or occurrence ratings to things, when we break things into its constituent parts, or if we take one bite of the elephant at a time, this helps us in several ways. First, we're breaking down an idea into its parts, which can help us better understand what it is that we're evaluating or analyzing. Second, it provides context against which it's easier for us to assign likelihoods or probabilities. It's much easier to assign a probability to a potential event if we have more context about the event itself. And third, breaking apart events to assign probabilities helps us explore the why of an event. Being able to answer or at least ponder why this event is occurring can help us to not only better understand the context so a double benefit of why but also helps us to start thinking along the lines of chains of events or things that might occur that lead to that event happening. That can give us a better idea of what kind of likelihood is possible for the event itself if we better understand the things that lead up to it. Those are three reasons why breaking apart an event that you're trying to assign an occurrence or a probability to is helpful.

Speaker 1 4:17

Breaking apart a big problem into smaller chunks or to smaller problems that you think you can solve isn't a new concept. This is something that you're probably doing daily or within your projects to help you move forward with the project and make some progress. When we're faced with too big of a problem, we can get stuck, so we know to break it down into smaller problems that we can solve. It's the same kind of approach here when we are trying to assign a probability to an event, we can solve. It's the same kind of approach here. When we are trying to assign a probability to an event, we can break the event up into pieces so that it's more manageable.

Speaker 1 4:53

And for those three reasons that I just listed, to break down our events into things that we can better analyze for probabilities is to use a statistical method which is conditional probabilities. Or, if you work in the medical device industry, you may be familiar with probability one and probability two. Those are also conditional probabilities. Let me explain how this would work for us and for our team. To assign probabilities to events, picture in your head an equation. On the one side of the equation is our total event, what it is. We're being asked to provide a probability for that event. We're going to break up into two different parts. We're going to break it up into its situation, whatever situation is causing the event, plus the impact that that situation has caused because it occurred. So we have our big event that we can break down into a particular situation, and then the impact that it has on whatever it is we're analyzing.

Speaker 1 6:04

Let's first take a look at this in the context of, maybe a project management risks, where they are also considering occurrences and likelihoods and probabilities. We're asked to prioritize something about our project based on the likelihood of bad things happening. So for our project, we have a risk event bad things happening. So for our project, we have a risk event, some sort of situation that could happen, and it has its own probability of occurrence. We also have the impact of that risk event. The impact is really what's the impact, given that the risk event has occurred? A risk event has to happen first to lead to some sort of an impact. The impact also has its own probability of occurrence. So now we have our total event. We have the situation that has a probability and we have the impact of that situation if it occurs, and that has a probability, and we have the impact of that situation, if it occurs, and that has a probability. So now, instead of us trying to figure out what's the likelihood of this event, now we're trying to figure out two probabilities, which it sounds worse but it's really not. We're breaking it apart, remember, so that we can more easily understand the context and the why and just the details of what it is we're analyzing. Now we have two probabilities we need to assign. One is what's the likelihood or probability that the risk event occurs, and the second one is what's the probability of the impact, given that the risk event has already occurred. Once we map these things out, once we break them into their parts, they're easier for us to assign a probability to.

Speaker 1 7:50

Since I started on the project management example, let's continue with that. Let's go through an example. Maybe we have a project and one of the risk events that could happen or that we're worried about is that the tool and die that we rely on to be able to make some new designs and new parts they're delayed and we end up losing our test window opportunity. In this situation, we need the parts to be able to assemble and we have already scheduled our test house, and if we lose it, it could be months worth of delay. So that's the risk. So how bad of a risk is this? It's sort of hard to just assign a probability to that, so we break it off into its different parts.

Speaker 1 8:37

All right, the first part, the situation or the risk event in this case, is that the tool and die are delayed and we don't get our parts. Now what is the likelihood of just that happening? What are the kind of things that you're starting to think about now. Chances are you're thinking about things, fact-based things, or drivers that could lead to the tool and die being delayed and us not being able to make parts, and you can start to understand and put some context around what kind of problem this is, how big of a problem it is, and you're probably more easily able to assign a probability of just that happening, of just the tool and die being delayed. But that's only one half of our problem. The other half is that we lose our test window opportunity. What's the likelihood or probability that we would lose our test window opportunity? What's the likelihood or probability that we would lose our test window opportunity? Now, when you consider just that, you're probably thinking of a whole set of different drivers that could lead to us losing that test window opportunity. There might be more variables that you come up with that are different than the first one, with the tool and die being delayed, just by breaking the problem apart, once we're able to more clearly see the nuances and the drivers and we can more easily assess the likelihood of occurrence or the probability of each of these individual things.

Speaker 1 10:23

Now you may be saying but that's not what I was asked. I was asked to give the likelihood that the tool and die are going to be delayed and we lose our test window. Well, with the magic of conditional probabilities, you would multiply the probability of the risk event and the probability of its impact and that gives you the overall probability of the total event. So now you can prioritize it against other things in your project and since you've broken it out, you can also better consider what kind of contingency plans you can set up or the types of things that you need to be monitoring to ensure things stay on track. So that's one example with project management, but the same can be true when we're looking at symptoms of our customers experiencing bad things.

Speaker 1 11:16

Perhaps you're evaluating field data and you're getting customer complaints. You need to prioritize what you're going to fix first, based on the information that you're getting in the complaints. One of the things you would want to look at is how often are things happening and how bad it is. But we're focusing on occurrence for this episode. So again, we would want to break it out into two separate parts. We want to break it out into the situation that's occurring with our product and then the impact that it's having on our customers. We assign a probability of each of those and then we multiply the probabilities together to be able to better prioritize and understand our events. There's a chance you may discover that you really can't assign a likelihood, but now you have a better information about where you need to investigate, investigations or tests or surveys do you need to conduct to be able to better answer what the likelihood of the impact is, for example, no-transcript. Now, really, you have three probabilities that you can prioritize on. You have the probability that a situation happens. You have a probability that an impact happens, given that the situation occurred, the overall likelihood of the outcome and the impact occurring. You could choose to prioritize on any one of these probabilities that you're assigning.

Speaker 1 13:05

Breaking apart events that we're trying to assign a probability to, whether it's risk-based, project-based or if we're trying to figure out how to best meet a customer benefit. Breaking things out helps us to think about things a little bit differently and helps us to assign probabilities based on more information and context than if we just left it as a whole big thing that we're trying to figure out. Another hang-up I see with assigning probabilities is that people are afraid that they might be wrong. What if I assign a probability and it turns out to be wrong? Chances are at some point it's going to happen.

Speaker 1 13:52

You need to do the best that you can with the information that you have today and you need to allow yourself grace and allow other people grace to change their minds, given new information, just because in January you assigned a probability to an event as 80%. But now that you've learned something new about it, learn new information, perhaps you performed a study. It comes to May and you think, boy, I was really wrong, go ahead and change the probability. It's not admitting fault when you update a likelihood of occurrence, given new information. You are not a bad engineer or a bad project manager, or just not good at your job because your assessment or your assignment of a probability was off.

Speaker 1 14:47

Having said that, there are things that you can do to improve your occurrence and probability estimations. One is to just consider the information that you have, I mean and actually consider it. Involve other people, because with other people comes new information. They may have an understanding about something that you don't or they may be looking at it from a different viewpoint. So understanding the information that you do have and then involving other people can also increase your pool of information with which to make a decision.

Speaker 1 15:27

Another thing you can do is just practice. You don't even necessarily have to practice on your field of study, although I can see that would be helpful too. Although I can see that would be helpful too. If we're being asked to make probability assignments on a particular product line or in a particular field, then let's make ourselves as knowledgeable about that field as we can. You can also practice with other things, other things that are printed in the newspaper or in the world records, kind of thing. Just practice making estimates of occurrence. You can do this within sports, if you like to follow baseball. Baseball has a lot of data and information and statistics in it. You could probably verify your estimations with the real data that it has.

Speaker 1 16:21

Douglas Hubbard is the author of the Failure of Risk Management and he calls this kind of practice calibration training. He has a series of questions listed in the back of his book, but the whole point of calibration training is just to train ourselves to be able to make decisions with the information we have. So what's today's insight to action? When we're stuck, especially when assigning probabilities and occurrence rankings to things, let's break it apart and evaluate the parts of whatever we're looking at. If it's occurrence, we're going to break it up using the statistical rules of conditional probability. It'll help us to better evaluate and think about events so that we can prioritize and make decisions.

Speaker 1 17:09

If you're not sure about this whole conditional probability idea and how it's relating to breaking things apart, just respond to one of the weekly newsletters that I send out asking about conditional probabilities and I'll send you a cheat sheet about it. If you're not part of the newsletter yet, go ahead and sign up at qualityduringdesigncom. There are forms in blue that you can sign up for the weekly newsletter. You'll immediately get an email from me to confirm that you want to sign up for the weekly newsletter. You'll immediately get an email from me to confirm that you want to sign up for the newsletter and we can start communicating after that. This has been a production of Dini Enterprises. Thanks for listening.