# Better Decision Making Under Uncertainty

There is an entire area of Industrial Engineering known as Operations Research. It’s dedicated to making optimal decisions under circumstances of uncertainty. Complicated matrix manipulation can be used to figure out the highest expected return for any decision, even when there is no clear choice. For the scope of this essay, though, we’re going to focus on the big takeaways that don’t require complex math. The reason being, algorithms to find optimal solutions rely on the expected value of the decision. It requires assigning a utility value, a dollar value, or some metric to a possible outcome then multiplying it by the probability of getting that outcome. In many cases, though, the expected value is itself uncertain, and the probability is assumed, making the math nothing more than a fancy guess. If there is real data and a real expected value, it’s no longer decision making under uncertainty. So, let’s look at some non-technical ways to make decisions when there is uncertainty.

Eliminate the wrong choices first. In general, when comparing multiple alternative options, there is one choice that is definitely worse than the other options. Whether deciding on a restaurant for dinner, adding a new product line, or looking at what new CNC machines to purchase, if there is an option that is clearly worse, eliminate it first.

Protect the downside. If one of the alternatives being considered poses a potentially better upside, but is not obviously a better decision due to the downside risk, go with the safer option with a limited upside. If the upside is worth some risk, go for it. If the risk could end your business, do not go for it. (catastrophic risks are rarely worth it)

Change the frame of the decision. Or, in other words, read between the lines. Daniel Kahneman and Amos Tversky famously ran an experiment where they stated that the US was preparing to have a disease outbreak that would cause 600 deaths (if nothing was done). They posited the following options.

Plan IA: Save 200 people

Plan IIA: A ⅓ chance at saving 600 people

In this group, most people chose Plan I. In the next round of participants, the experimenters posed the next plans;

Plan IB: 400 people die

Plan IIB: A ⅔ chance that 600 people die

This time, most people chose Plan IIB. In the first experiment, Kahnamen and Tversky believed people chose Plan IA because Plan IIA had a large chance of everyone dying. In the second experiment, most people chose Plan IIB because people recognized a chance that no one would die. Both samples are identical though. The only difference being in the phrasing of the solutions. Khanamen and Tversky’s conclusion was that the frame of reference you use when looking at a decision will impact how that decision is made. In the case of your business, if there is uncertainty looming, look for ways to change the reference point.

Any decision that can be easily reversed should be made quickly and reneged quickly if it proves to be a poor choice. Any decision that cannot be reversed needs to be made slowly, unemotionally, and with as much diligence as possible. If your business can make two decisions, one wrong and one right, faster than a competitor can make a single decision, you’ll generally have the upper hand.

Another option for making decisions under uncertainty is to use the 70% rule. Popularized by Amazon founder Jeff Bezos, the rule states that when you are 70% sure of the decision, make it. 100% certainty is unattainable in most circumstances, but 70% certainty is attainable. The remaining unknowns will often work themselves out in the process.

Lastly, don’t forget that doing nothing is a viable alternative in many cases. When it’s true, it might just be the preferable choice. If a problem will work itself out without your intervention, it might be a better use of your time and energy to let that happen.