FreeNo sign-upYour plan stays on your computer

Riskiest Assumption Test Tool

Plan small, cheap tests of the assumptions that would kill your idea, write each pass mark down before the test starts, and judge the results against them.

What to test

Test the riskiest first. Tests can run at the same time; each keeps its own pass mark.

The test

What kind of assumption is it?

The pass mark

Write it down before the test starts, and don't move it afterwards.

of them

Already settled by data you hold?

Skip the test only if real data already answers it, such as sales records or supplier prices you have paid. Say what the data is.

Your plans never leave your browser

The assumptions, plans and your notes stay on your computer and in your progress file. To write the diagnosis, only the numbers, such as each pass mark and result, are sent to our server. How it works

About the tool

Why it exists

No framework can tell you whether people will pay

Only a test can. The riskiest assumption is the one that would kill the idea if it were wrong and that you are least sure of. Some are about demand: will people pay, and how? Others are about delivery: can you do it, at a cost you can live with? Each gets the smallest, cheapest test that could prove it wrong, often doing the service by hand for a few customers before building anything.

Write each pass mark down before the test starts, and don't move it afterwards. A result judged after the fact can always be made to look like a success. This tool works on its own, and as stage 4 of our playbook for validating a business idea, where the assumptions come from the Lean Canvas in stage 3.

The method

How the tool works

What to test. In the playbook, the list starts with the assumption stage 3 chose, then every other one that would kill the idea and that you are not sure of, in rank order. You can add others. Tests can run one after another or at the same time; one trial can check two assumptions, each with its own pass mark.

Two kinds of test. "Will people do it?" counts how many commit something real: money, time, or only a click, which is the weakest. "Can we do it, at this cost?" measures a number against a limit, such as the cost of serving each customer. The tool suggests a plan for the empty fields from what it knows, including a starting pass mark of about 60%; the pass mark is yours to set.

The pass mark, written down. Locking it records the date. You can still change it, but the change is recorded, shown in the report, and the test is judged against the mark you wrote down first. A test can be skipped only when data you already hold settles it, and the report says so.

The verdict. Passed when the count reaches the mark, or the number is within the limit; just short when within one person or 80% of a count, or 10% of a limit; otherwise failed. If the riskiest fails, change the idea and test again, or stop.

What leaves your computer. The assumptions, plans and notes stay in your browser and your progress file. To write the diagnosis, the page sends our server only the numbers, and the server keeps nothing.