When to Use a Decision Maker, and When You Should Trust Yourself

The question "should I use a decision maker for this?" has a specific, useful answer. It is not "always" or "never." It is "sometimes, and here is how to tell when." The distinction is between decisions where the tool helps and decisions where the tool gives you false confidence.

The three things a decision maker is for

A decision maker, a coin flip, a random picker — they are all the same tool with different inputs. They answer a specific kind of question: between these N options, which one do I pick? The tool is useful when the question itself is the right one. It is not useful when the question is the wrong question.

1. Analysis-paralysis on equal options

The classic case. You are choosing between two restaurants, three colors, four weekend plans. You have looked at the menus, the reviews, the photos. You have a vague preference but no real reason to pick one over the other. You are spending 15 minutes trying to find a tiebreaker that does not exist. At this point, the analysis is done and any further analysis is just stalling.

The decision maker is the legitimate end of the analysis. The analysis said "these are equivalent" and the tool says "this one, then." Use it. You will enjoy the restaurant you did not pick, because the analysis was actually showing you they were equivalent.

2. Breaking a long deadlock

You and another person have been going back and forth on a decision for 30 minutes. The discussion has stopped producing new information. You are arguing about preferences, not facts. The deadlock is a sign that the data is exhausted and only subjective preference remains.

Flipping a coin (or running a tool) breaks the deadlock. The tool does not give you the right answer — there is no right answer — but it does break the loop. If you feel relief when the coin lands on a particular side, you had a preference you did not know about. If you feel disappointed, you have a preference too — flip again or pick the other side.

3. Choosing between truly equivalent things

Some choices are objectively equivalent. The leftmost empty seat at a conference. Which of two parallel parking spots to take. Which of the six identical-looking USB-C cables to grab. The decision maker is the right tool here. Save the human brain for the choices that actually matter.

When not to use a decision maker

1. When one option is clearly better

If you have done the analysis and one option dominates, do not randomize. Choosing the objectively-better option is a skill, not a bias. If the safer car is 10% more expensive and has better crash ratings, buy the safer car. If the cheaper flight is 2 hours longer and arrives at midnight, take the longer flight if you can sleep on planes. If the local restaurant has 4.7 stars and the chain across the street has 3.4 stars, go to the local one. The decision maker would pick randomly; you should pick the better one.

Randomizing a clear win is not being humble. It is wasting a clear advantage. The "I might be wrong" humility only applies when the data is not clear.

2. When the decision is important and the analysis is fresh

Big decisions — job, partner, house, major purchase, business strategy — deserve real analysis, not random selection. If you have not done the analysis, do the analysis. If you have done the analysis, trust it. The decision maker is for the cases where the analysis is done and the options are equivalent, not for the cases where the analysis has not started.

The failure mode here is using the tool to avoid responsibility. "I flipped a coin and it said take the job" is not a decision; it is the absence of a decision dressed up as one. You still have to own the outcome.

3. When the cost of being wrong is high

Some decisions have asymmetric consequences. A bad restaurant choice wastes an evening. A bad medical choice wastes a life. The decision maker is appropriate when the cost of being wrong is roughly equal across options. It is not appropriate when one option is much riskier than the others.

This is the "don't gamble with things you can't afford to lose" rule. If the worst-case outcome of either option is acceptable, you can flip. If the worst-case outcome of one option is much worse than the other, you cannot.

4. When you actually have a preference but want to feel decisive

Sometimes you have a preference but you do not want to make the call. You want the tool to make the decision so you can blame the tool if it goes wrong. This is fine in low-stakes situations (where to eat lunch) and is a problem in medium-stakes ones (whether to invest in a friend's startup). The tool gives you randomness; the responsibility is still yours. Use the tool to break ties, not to avoid ties.

The two-minute test

When you are about to reach for a decision maker, ask two questions:

1. Have I spent more than 2 minutes on this decision? If no, decide normally. If yes, the tool is probably right.
2. If the tool says option A and the actual outcome is bad, will I regret using the tool? If yes, decide normally and own the outcome. If no, the tool is fine.

These two questions filter out most of the cases where the tool would be misused. If the answer to both is "use the tool," use it. Otherwise decide normally.

How to use the tool well

When you do use a decision maker, the input matters. A list of two things is different from a list of ten things, and a list of options you actually like is different from a list of options you wrote down because you felt you had to consider them.

The best use is to list the options you would actually pick from, not all the options you could theoretically consider. If the list has more than five options, you have not narrowed enough. If the list has only options you do not like, you have not actually engaged with the question.

For two options, a coin flip is enough. For three to ten options, a multi-option picker is the right tool. For more than ten, you have a different problem (probably a missing constraint) and the tool will not help.

What the tool is not

A decision maker is not an oracle. It does not have insight. It does not know your preferences. It does not see the future. It just picks one option. The "decision" part is yours: you decided the question, the options, the stakes, and the moment to use the tool.

The tool is also not a way to avoid thinking. If you have not thought about the decision, the tool gives you a random answer to a question you have not actually asked. That is not a decision. It is a coin that says "yes" or "no" without knowing what the question is.

It is also not a way to make a decision that affects other people. The coin flip you do alone is yours to live with. The coin flip you do for a group punts the responsibility to chance, which is fine for restaurant choice and terrible for "where do we go on vacation for a week."

The honest summary

The tool is for breaking analysis-paralysis on equal options, breaking deadlocks, and choosing between truly equivalent things. It is not a substitute for judgment, not a way to avoid analysis, and not appropriate for high-stakes decisions. The two-minute test catches most of the misuse cases.

Used correctly, the tool is genuinely useful. It saves you the time you would have spent second-guessing, and it gives you a clean way to say "I decided, on purpose, and here is what I picked." The 30 seconds it takes is worth it. The hours of hesitation it replaces is the whole point.

If you find yourself reaching for the tool, it usually means the analysis is done and the next step is just to act. The tool is a way to act when the only thing stopping you is the lack of a tiebreaker. Use it for that. Don't use it for anything else.