About DecisionBranch

Why it exists

Every drawing tool can draw a decision tree, and none of them will tell you which branch to take. The tools that do the arithmetic are academic and look it. DecisionBranch sits in the gap: write the tree as a few lines of text, see it drawn, and get the expected value of every option, the recommended path, the switch points, the value of information and the risk profile, all recomputed as you type. The math is the product; the picture is how you read it.

Principles

• The text is the truth. A tree is plain text you can read, paste into an email and keep with your files. Every view, the canvas, the public page, the embed, the AI builder, reads and writes that same text.
• Never change the math to make a screen look right. Every number DecisionBranch shows is checked against answers worked out by hand. If a number looks wrong, it is proved by hand before anything moves.
• Free core value. Building, analyzing and sharing a tree are free and need no account. Paid plans add depth for professionals and teams, not the basics.
• No tracking for its own sake. No advertising, no cookies, and analytics that count events without identifying people. See the privacy policy.

Who makes it

DecisionBranch is made by Tov Studios LLC, a small software studio. The math, the editor and this site are all built in-house, and the person answering hello@decisionbranch.com is the person who builds it.

The math, briefly

Decision trees with expected value are standard decision analysis, the method in Raiffa's Decision Analysis (1968) and every operations research course since. DecisionBranch implements rollback (backward induction), sensitivity analysis with switch points found by bisection, the expected value of perfect information per chance node, and the risk profile of the recommended policy. The expected value guide explains each with a worked example.

Get in touch

Questions, bugs, course licenses, custom-domain embeds, press: hello@decisionbranch.com. See the contact page for what to include.