connecteddale

Strategy Coach = Clarity + Alignment

Decision Trees

A Decision Tree maps a decision as a branching sequence of choices and chance outcomes, each with a probability and a value, so you can calculate which branch is worth most before you commit.

Five stages turn a branching decision into one number: draw the choices, add the odds, and calculate what each path is really worth.

1 Draw the choices 2 Add chance events 3 Assign a value 4 Calculate expected value 5 Choose, then check risk
The sequence for turning a branching decision into an expected value.

Reach for this when…

How to run it

  1. Draw the first decision point and every real choice branching from it.
  2. Add the chance events that follow each choice, with their probabilities.
  3. Assign a value, cost, revenue, time, to each end outcome.
  4. Calculate the expected value of each branch by multiplying value by probability and summing.
  5. Choose the branch with the best expected value, then check it against your actual risk appetite.

A worked example

Situation. Amina Njoroge ran Uzima Credit, a small business lender in Nairobi, Kenya, deciding whether to extend a risky loan or refer the applicant elsewhere.

Applied. She built a decision tree branching approve versus decline, then default versus repay under each, with probabilities drawn from the applicant's sector and repayment history.

Result. The expected value slightly favoured approving with a higher interest rate and a shorter term, a structure the loan committee had not considered when they were only comparing approve or decline outright.

1 Draw the choices 2 Add chance events 3 Assign a value 4 Calculate expected value 5 Choose, then check risk
Fjord Kredit's expected value favoured a structured approval the committee hadn't considered.

The catch

The whole calculation rests on the probabilities you put in, and those are usually estimates dressed up as numbers - get them wrong and the expected value is precise and wrong. It also collapses a genuinely risky, high-variance branch into the same single average as a safe one, which can make a gamble look identical to a sure thing on paper.

An expected value slightly ahead of the alternative is not a reason to ignore how much you'd lose if the bad branch actually happens.