Decision trees for founders and operators

Pricing, hiring, a new market: see which option is worth most, how sure you need to be, and what to tell the board.

The problem

  • Every big call is a bet you can't rerun

    A price rise, a senior hire or a second market can't be A/B tested. You get one go, so you need the expected value and the downside before you commit.

  • The board asks why

    "We think it's right" doesn't survive a board meeting. A tree shows the options, the odds you assumed and what each one is worth, on one page.

  • Your gut has the numbers, but not the maths

    You know roughly how customers will react. Put those guesses in and see whether the answer depends on them, or holds either way.

Worked example

Raise prices by 20%?

A software company can raise prices for everyone, raise them for new customers only, or hold. Payoffs are the extra gross margin over the next twelve months compared with holding the price; the cost is the work to change the pricing page and billing.

Best optionRaise for everyoneExpected value 108,000 · 42,000 ahead of Raise for new customers only
Raise prices by 20%?60%40%−5,00070%30%Raise prices by 20%?EV 108,000→ Raise for everyoneRaise for everyoneEV 108,000Customers accept it240,000Churn spikes−90,000Raise for new customers on…EV 71,000Conversion holds110,000Conversion drops−20,000Keep today's price0

Where it flips. If the chance of customers accept it under Raise for everyone falls below 47.3% (it is 60% now), Raise for new customers only wins instead.

0%47.3%100%

Range of outcomes. If you follow the recommendation, the chance of ending up with a loss is 40%; the worst case is −90,000 and the best is 240,000.

40%−90,00060%240,000

Open with the recommendation and how far ahead it is. Then give the switch point: it's the share of customers who must accept the rise for it to beat the safer option, which is the number your board will want to argue about.

This example as text
# Analytical: raise prices by 20%?
Raise prices by 20%?
  Raise for everyone
    Customers accept it (p 0.6): 240000
    Churn spikes (p 0.4): -90000
  Raise for new customers only (cost 5000) # pricing page and billing changes
    Conversion holds (p 0.7): 110000
    Conversion drops (p 0.3): -20000
  Keep today's price: 0

What you get

  • The best option, with its value

    Expected value on every option, and the margin over the next best, updated as you change a number.

  • How sure you need to be

    The switch point for each assumption, so you know which guess is worth checking before you decide.

  • The downside in plain view

    Every outcome with its probability, so the chance of losing money is on the page, not hidden in an average.

  • What research is worth

    The most a customer survey, a pilot or an advisor is worth before you commit.

  • A slide for the board

    Export a 16:9 PowerPoint slide with the tree and the recommendation, or a PNG or PDF.

  • Private until you share it

    Keep strategy work private: five private trees free, unlimited on Pro.

How people use it

  1. 01

    Pricing

    Raise for everyone, for new customers only, or hold. Put a probability on churn and see where the answer flips.

  2. 02

    Hiring

    A senior hire now, a contractor for six months, or wait a quarter. Put a cost on each and a chance on it working out.

  3. 03

    Expansion

    Launch in a second market, pilot one city, or stay focused. The launch template is a ready-made starting point.

Templates for this work

All templates · How to make a decision tree · Expected value and decision trees

Questions

What if I don't know the probabilities?
Use your best guess, then check the switch point. If the answer only flips far from your guess, the guess doesn't matter; if it flips close to it, that's the thing to find out first.
Can I show this to my board or investors?
Yes. Export a PowerPoint slide with the tree and the recommendation, or share a link where they can change a number and watch the answer move.
Is my data private?
Yes. Trees are private unless you publish them, and you can use the editor without an account if you just keep the link.

Start with the example, or your own decision.

The editor is free and needs no account. Describe the decision in a sentence and the AI builder drafts the tree.