How do you make decisions when you don’t have all the data?

Don’t wait for complete data, because in a small business you will almost never have it. Treat the decision as a bet: list the possible outcomes, give each one a chance in percent, even as a wide range, and write down why you’re choosing this path. Afterwards, judge first how the decision was made, and only then how it turned out.

A good decision can turn out badly

Annie Duke played professional poker for twenty years and then wrote Thinking in Bets, a book about deciding under uncertainty. She starts with a mistake poker players call “resulting”: judging a decision by its outcome. When a few hands go badly, it’s tempting to change your strategy. More experienced players warned her against this from the start.

Duke compares two games. In chess, everything is on the board and chance plays almost no role, so whoever lost made a mistake somewhere. In poker, some cards are face down, and part of the outcome is plain good or bad luck. A business is more like poker. You don’t know what your competitors are planning, what a client isn’t saying out loud, or whether in six months someone will dig up the street in front of your premises.

I agree with her main point: the quality of a decision shows in the process, not the outcome. A good process starts with an honest admission of how much you know and how much you’re guessing. The mistake cuts both ways. After one unlucky outcome it’s easy to throw out a sensible way of working, and a lucky break can lock in a bad habit.

Every decision is a bet, including “I’ll wait”

Duke writes that whenever you decide, you choose among options, you risk money or time, and you assume what is likely to happen. Choosing not to bet is a bet too.

Take a hypothetical physiotherapy practice with six people. The schedule is booked three weeks ahead, and the owner is considering a second location in another part of town. She has three paths: open it in January, wait six months, or rent a single treatment room from someone else a few days a week as a trial. “I’ll wait” is a bet that the patients on the waiting list won’t go to a competitor in the meantime.

Now some fictional numbers for this example. The owner doesn’t know whether the second location will fill up, but she does know something about it. She writes down: the chance that after a year the schedule will be at least half booked is 40–70%. The chance that the location will have to close is about 20%.

Duke suggests rating your confidence on a scale from zero to ten, in effect a percentage, or giving a range. The less you know and the more depends on chance, the wider the range. Even 20–80% beats having no estimate, because it moves you away from the hidden assumption that something is 0% or 100% certain.

Another help is the question Duke used as the title of her second chapter: “Wanna bet?” If you had to put your own money on your belief, ask yourself the questions Duke lists:

how do I know this, and who did I hear it from;
how fresh is this information;
what else did I feel sure about that turned out to be false;
what other possibilities are there.

With a range of 40–70%, a trial with one rented room looks more sensible than a full location from January. After three months the owner will know more, and the range will narrow.

Premortem: imagining the failure in advance

Gary Klein has spent years studying how people decide under pressure, starting with fire commanders. In Sources of Power he describes an exercise he named the premortem. You imagine that a few months have passed, the plan was carried out, and it failed. That’s all you know. Your job is to explain why. Klein writes that imagining the failure and its causes takes less than ten minutes, while the discussion afterwards can last an hour. In his trials the exercise seemed to lower confidence in the original plan, which is what he expected.

In our practice it looks like this: it’s December next year, the second location has closed, and the team lists the reasons. The physiotherapists didn’t want to commute. Patients in the new part of town had someone else closer by. The owner spent half the week in the car. Each of these reasons is a question to check before signing the lease.

This is where AI helps. It has no stake in you being right, so ask it for ten reasons your plan failed before you show the plan to your business partner.

Checklist before a decision without complete data

name the decision in one sentence and write down the date by which it has to be made;
list at least three options, including “I do nothing” and a trial version;
for each option, write down the possible outcomes and give them percentages; a range counts too;
ask yourself whether you would bet your own money on it, and check how you know what you know;
run a premortem: it’s a year from now, the decision failed, write down why;
write down the choice, the reasons, your level of confidence and the review date;
at the review, judge the process first, then the outcome.

When this doesn’t work

For small decisions that are easy to reverse, this whole procedure costs more than a mistake. Decide quickly and correct course if you need to.

Percentages won’t save a bet whose worst outcome means the end of your business. If the bad version wipes out your cash, in my view you don’t go in, whatever the odds. You look for a smaller version or a trial.

Estimates turn into decoration if you never compare them with what actually happened. Then “70%” means no more than “I guess so.”

For tomorrow

Take one decision you’ve been putting off for weeks and write down three possible outcomes for it, with percentages.

Is there a decision you keep putting off? Let’s talk about it

How this text was written: the research, the choice of sources and the final version are mine. AI helped me prepare a draft; I checked every fact, and I am responsible for this text. — Maciej Chmura

Sources

Annie Duke, Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts, introduction “Why This Isn’t a Poker Book”; ch. 1 “Life Is Poker, Not Chess”; ch. 2 “Wanna Bet?”; ch. 6 “Adventures in Mental Time Travel.”
Gary Klein, Sources of Power: How People Make Decisions, MIT Press 1998, ch. 5 “The Power of Mental Simulation,” p. 71.