lockstep
4 min readNaman Jain

Decision tree template with a worked example

A free decision tree template for business decisions: decision, chance, and end nodes, how to calculate expected value, a build vs buy example, and pitfalls.

A team is deciding whether to build a search feature, buy a vendor's, or wait a quarter. The debate goes in circles because everyone argues from a different scenario: the engineer from the case where the build slips, the PM from the case where the vendor goes bust. Each scenario is plausible. Nobody has weighed them.

A decision tree weighs them. This guide gives a template and a worked example, using the conventions in the decision tree article on decision analysis, checked on 16 November 2026.

The parts

NodeShapeMeans
DecisionSquareA choice you make
ChanceCircleAn uncertain event, with a probability per outcome
EndTriangleThe final result and its payoff

Trees are drawn left to right: a decision, then the chance events that follow each option, then the end results.

How to build one

  1. Write the decision in a square: "How do we add search?"
  2. List the options as branches: build, buy, wait.
  3. Add chance nodes after each option for the uncertainty that matters: "build succeeds or slips."
  4. Assign probabilities to each outcome. They must add up to 1 at each chance node.
  5. Put a payoff on each end node, in the same unit for all of them.
  6. Note each option's cost.
  7. Roll back: for each chance node, sum probability times payoff. Subtract the option's cost. That's its expected value.

Worked example: build, buy, or wait

Payoffs are estimated two-year value in thousands.

OptionCostOutcomeProbabilityPayoffProbability × payoff
Build200Ships on time, works well0.7600420
BuildSlips two quarters0.310030
Buy120Vendor works well0.9450405
BuyVendor problems, migrate off0.1505
Wait0No change1.000

Expected values:

  • Build: 420 + 30 − 200 = 250
  • Buy: 405 + 5 − 120 = 290
  • Wait: 0

Buying has the highest expected value. It also has the lower downside: the worst build outcome nets −100 (100 − 200), the worst buy outcome −70 (50 − 120).

The spreadsheet template

OptionCostOutcomeProbabilityPayoffP × payoff
[Option A][Cost][Outcome 1][0.0 to 1.0][Value]=D2*E2
[Outcome 2]=D3*E3
[Option B][Cost][Outcome 1]
[Outcome 2]

Expected value for an option: =SUM(F2:F3)-B2. Check that probabilities for each option sum to 1.

Reading the result

  • Expected value is an average, not a prediction. If the worst case would be fatal, risk matters more than the average. The Wikipedia article notes analysis can include a decision maker's risk preference.
  • Test the probabilities. If a small change flips the answer, the decision depends on that estimate. Spend time improving it.
  • Close calls are ties. Decide on other grounds.

Decision tree vs decision matrix

Use a decision tree whenUse a decision matrix when
Outcomes are uncertainOutcomes are fairly clear
You can estimate probabilitiesSeveral criteria matter
Sequence matters (decide, then events happen)You're comparing options side by side

See the decision matrix template for the other one.

After the tree

The tree supports a decision; it doesn't record one. Once the owner decides, write down the choice, the probabilities you assumed, and when you'll revisit. If the vendor's reliability turns out worse than 0.9, that note is what tells you to reopen the decision.

Where Lockstep fits

Lockstep keeps that record. It captures decisions from documents, Slack, and pull requests, routes each to its owner for approval, and briefs the engineers and Claude Code sessions building against it. When the team revisits the call, the new decision supersedes the old one and both stay visible.

See how it works.

Sources checked 16 November 2026

For choosing a framework, see product decision-making frameworks. For recording the outcome, see the decision log template.

Frequently asked questions

What is a decision tree?
A tree-shaped model of a decision, its possible consequences, and their chance outcomes and costs. In decision analysis it's used to compare alternatives by calculating their expected values. It's a different thing from decision trees in machine learning, which classify data.
What do the shapes in a decision tree mean?
By convention, squares are decision nodes where you choose, circles are chance nodes where outcomes happen with some probability, and triangles are end nodes showing the final result or payoff.
How do you calculate expected value in a decision tree?
For each chance node, multiply each outcome's payoff by its probability and add the results. For each option, subtract its cost from that total. The option with the highest expected value is the best on average, before considering risk tolerance.
When should I use a decision tree instead of a decision matrix?
Use a decision tree when the main difficulty is uncertainty: outcomes that depend on events you can estimate but not control. Use a decision matrix when the difficulty is weighing several criteria for options whose outcomes are fairly clear.
Can I make a decision tree in Excel or Google Sheets?
Yes. List each path as a row with its option, outcome, probability, payoff, and cost, then calculate probability times payoff and sum by option. Drawing the tree itself works in any diagram tool, but the math fits in a spreadsheet.

Keep reading