Dealing with Uncertainty

A lot of decisions in lean manufacturing have uncertainty. How many products will I sell (and what is my customer takt)? Which layout is more efficient? Should I believe expert A or expert B? Uncertainty is a part of life in manufacturing. In fact, the higher up you go in the hierarchy, the more you have to deal with uncertainty. And often these are not just simple “A or B” type of questions, but highly complex and interacting decisions like “What should our new line look like?” Here are some suggestions on how to deal with uncertainty. Please note that they will not answer all of your questions but will help you make better decisions.

The Wrong Way

In industry, I often encounter managers who seem to have an answer for any question. No matter what you want to know, they can tell you what you should do.

Those managers scare me!

Nobody knows all the answers! They just tell you the first thing that pops into their minds. Unfortunately, this is rarely the best and many times not even a good solution. Equally unfortunately, their answers often look good to their bosses and may help them to get promoted.

Please do not get this “instant answer” confused with conviction. Once a decision has been made, the manager should give his people a feeling of certainty and conviction. A commander-in-chief usually should not show his own doubts to his people. But he should have doubts and question his own decisions. Most of all, the manager should not just make up answers on the spot! Here are some suggestions on how to do it better.

Break Problems into Smaller Problems

One approach is to break a big problem into multiple smaller problems. The initial big problem may have lots of uncertainty. If you break it into smaller sub-problems, there will be less uncertainty. Sure, there are still things you don’t know, but others you will know. A mix of sub-problems with varying uncertainty will generally give a better result than looking only at the big problem with lots of uncertainty.

Take for example the problem of designing a new line. It will be much easier to break this into sub-problems. To answer the question of how the line should look, you answer many sub-questions like “What is my customer takt?” “What is the OEE?” “What are my cycle times?” and so on. Even though all of these questions have some uncertainty, overall the result will be much better than creating a line without these details.

You can break down problems into smaller sub-problems. You can also give these problems a hierarchy or sequence. First you decide if you want a flow shop or a job shop, and then you go to the next decisions based on your previous decisions (see also the alternate scenarios below).

Get Good (Enough) Data

Sometimes, uncertainty can be answered by getting and analyzing data. Yet obtaining and analyzing this data takes effort and – more importantly – time. Hence, you have to make a trade-off.

In my experience, Western managers often fall short when obtaining and analyzing data. Take for example sales prediction. In the West, many mathematical models are available to predict future sales data based on past sales. Unfortunately, while mathematically beautiful, they are often not very good. In fact, the most complex models seem to give the worst results.

Toyota, on the other hand, spends a lot of time talking to past, current, and future customers. Much more, in fact, than Western car makers. Hence, the sales predictions at Toyota are often of a very high quality compared to other companies.

Use the Wisdom of Groups

Correctly managed, groups can be more intelligent than individuals. Together, a group has more knowledge and experience than an individual. Hence accessing the knowledge of the group can create better results.

There is a famous NASA exercise, “Ranking Survival Objects for the Moon,” that demonstrates this point. Assume you are stranded on the moon and have a list of 15 possible items that you can take along (from oxygen tanks to a compass to matches). What would be the ranking in priority of the different items. In this exercise, groups consistently perform better than individuals.

Hence, involving multiple people in the decision making often gives better results. You could either interview or talk with different people to get their view and help form your opinion. Alternatively, you could have a group workshop or exercise to access this wisdom of the crowd. Please note that the smoothest talker is not always the most knowledgeable person.

Make Different Scenarios

It can also help to make different alternative scenarios. Sometimes it is difficult to decide between two or more options. In this case it can help make a scenario for the different options.

For example, if you have not yet decided if it will be a flow line or a job shop, make a draft of both and compare them. Do not make a completely designed flow shop or job shop, but merely an estimation of how it could look. This should be enough to compare multiple ideas.

I have seen this multiple scenario problem solving very often in Japan, and I also use it myself. I like this method so much that I wrote a full blog post on it: “Japanese Multidimensional Problem Solving.”

Delegate

In many companies I have seen, it is common to leave the problem for the next level in the hierarchy. Many managers also feel empowered by making decisions and want to make the decisions.

Unfortunately, all too often managers have to make too many decisions while lacking both time and knowledge to make good choices. There is even the common effect of Decision Fatigue for the deteriorating quality of decisions made by an individual, after a long session of decision making. The more decisions a person makes, the worse these decisions become.

Often, it would be much better if the manager hands the decision making back down to the next level in the hierarchy. These people generally know many more details and also can pay more attention to the particular decision. Hence, whenever someone asks you to make a decision, think first if you are the right person to make this decision or if it would be better if you let your people make the decision.

Summary

Making decisions without all the information is tough. Yet in industry this cannot be avoided. In fact, the pay grade tends to reflect the level of uncertainty in decision making. A CEO needs to make decisions with much less certainty about the outcome than an operator on the shop floor. Yet the decisions have to be made. Now, go out, make some decisions – or even better, have them made by your people – and organize your industry!