Chapter 2: Find out if you're wrong, before you build, lesson 3 of 7

Test the riskiest belief first

An idea rests on several beliefs, and there is rarely time to test them all. Three plain questions show which belief to test first, so the week isn't wasted.

By 5 min read, 15 min of exercises

Warm-up

From last time. You turned Peter's belief into a test. Is each of these ready to test, or not a test yet?

  1. Fewer than half the managers who get the link use it twice in two weeks. We decide on 16 October.

    Right, it's ready to test. It has a result that could prove Peter wrong, and a real date to decide.

    Not quite. It's ready to test. It has a result that could prove Peter wrong, and a real date to decide.

  2. Office managers will love the link.

    Right, it's not a test yet. Love is a feeling. Nothing you could count would prove it wrong.

    Not quite. It's not a test yet. Love is a feeling. Nothing you could count would prove it wrong.

  3. We'll watch the link until the numbers look good.

    Right, it's not a test yet. There's no date. It only ends if you're right.

    Not quite. It's not a test yet. There's no date. It only ends if you're right.

Last time, you turned one of Peter’s beliefs into a test. Peter, Tiffin’s head of sales, wants group ordering: one link, and everyone adds their own meal. Your test had a number that could prove him wrong, and a date.

On Friday, Mia, Tiffin’s research lead, read it.

Think first

Mia, research lead, after reading your test: Good. But Peter’s idea rests on four beliefs, and we only have time to test one this week.

These are the four beliefs, from lesson 2.1:

  • Offices order lunch as a team, most days.
  • Placing the order is the hard part.
  • People will open a link and add their own meal.
  • It will bring more orders than it costs to build.

How would you decide which belief to test first?

You

Which of these is closest to your answer?

Shirish

Easy is tempting. You’d have an answer by Monday. But what if it’s a belief that couldn’t sink the idea anyway? Then you’d have a quick answer that changes nothing.

Which belief would hurt most if it turned out wrong?

Shirish

There’s something to that. The belief everyone is surest of is often the one nobody checked. But being sure isn’t the same as having proof.

For each belief, what proof does the team actually have?

Shirish

Yes, start there. But look at the four again. The idea might not live without any of them.

If all four matter, what would tell them apart?

Shirish

It feels safe. But one week split four ways makes four tests too small to prove anything.

If you had to bet the whole week on one belief, how would you pick it?

You

What this means

Mia doesn’t pick by feel. She asks three plain questions about each belief:

  1. If it’s false, does the idea die?
  2. How much proof do we have?
  3. If it’s false, which other beliefs stop mattering?

You’ll ask all three now, one at a time.

Questions 1 and 2: could it sink the idea, and is there proof?

First, Mia asks Nora Kim, Tiffin’s growth manager, to check the order records. Tiffin already has a group cart. One person adds each meal, one by one, and pays for all of them.

Ryan Brooks, the product manager, adds a belief of his own to the list.

Practice

For each belief, ask Mia's first two questions. Could it sink group ordering if it's false? And is there any proof yet?

  1. Offices order lunch as a team. Nora's records: 34 of Tiffin's 60 office accounts use the group cart every week.

    Right, it's a belief that could sink the idea, but has some proof. With no team orders, there's nothing to group. But the records count what offices really do. That's real proof.

    Not quite. It's a belief that could sink the idea, but has some proof. With no team orders, there's nothing to group. But the records count what offices really do. That's real proof.

  2. Placing the order is the hard part. Nobody has asked offices which part of team lunch goes wrong.

    Right, it's a belief that could sink the idea, with no proof yet. If the hard part is somewhere else, a link fixes the wrong thing. And nobody has checked.

    Not quite. It's a belief that could sink the idea, with no proof yet. If the hard part is somewhere else, a link fixes the wrong thing. And nobody has checked.

  3. People will open a link and add their own meal. No office has seen a link yet.

    Right, it's a belief that could sink the idea, with no proof yet. If nobody uses the link, group ordering is dead. And there's nothing to look at yet.

    Not quite. It's a belief that could sink the idea, with no proof yet. If nobody uses the link, group ordering is dead. And there's nothing to look at yet.

  4. It will bring more orders than it costs. Peter thinks it could double office orders.

    Right, it's a belief that could sink the idea, with no proof yet. If it costs more than it earns, it's a loss. And Peter's guess is an opinion, not proof.

    Not quite. It's a belief that could sink the idea, with no proof yet. If it costs more than it earns, it's a loss. And Peter's guess is an opinion, not proof.

  5. Ryan's belief: managers will want to use the link on a laptop, not only on a phone.

    Right, it's a belief the idea would survive without. If it's false, you change the design. Group ordering still works.

    Not quite. It's a belief the idea would survive without. If it's false, you change the design. Group ordering still works.

Look at what’s left. Three beliefs could sink the idea, and none of them has proof. The first two questions leave a three-way tie.

The third question breaks it.

Question 3: what else falls if it’s false?

Think first

Say Mia finds that placing the order is not the hard part. Offices find it quick.

What happens to the other two beliefs with no proof: the link, and the extra orders?

You

Pick the answer closest to yours.

Shirish

They are written separately. But picture it. Ordering is already quick. Would a manager bother sending a link to fix it?

And if nobody sends the link, where do the extra orders come from?

Shirish

Yes. When that belief falls, the two above it fall with it. A test of the link would be a wasted week.

So which of the three should be tested before the others?

Shirish

Some might, out of curiosity. But would they keep using it, for a part that was never hard?

What would bring them back the second week?

You

What this means

Some beliefs rest on others. If a lower one is false, every belief above it stops mattering. So testing a belief above it first can waste the week.

Answer question 3 first. Then you'll see Peter's idea as a stack.

Here is Peter’s idea, drawn as a stack.

  1. It will bring more orders than it costs. No proof yet
  2. People will open a link and add their own meal. No proof yet
  3. Placing the order is the hard part. No proof yet
  4. Offices order lunch as a team. Proof: Nora's records
Each belief rests on the one under it. If one is false, the ones above it stop mattering.

So test the lowest belief that has no proof. For Peter’s idea, that’s placing the order is the hard part. That’s what Mia tests first. It’s also the answer to your bet from lesson 2.2.

I learned to ask this third question on an idea of my own.

Shirishfrom my own work

Why the teacher came before the students

I worked on an idea for teachers. When a student hands in code, show how it was made: built up over days, or pasted in whole. Then a teacher could judge what the student really knows, even when AI helped.

Two beliefs worried me. Could a student fake a history of trial and error? And would the teacher open the history at all?

Faking looked like the big danger. But grading by what you already think of a student is the easiest way to grade. If that’s how a teacher grades, he may never open the history. And nobody fakes a history that nobody looks at.

So the teacher’s belief came first. Until he opens the history, faking it doesn’t matter.

What you just did

You just made an assumption map. It sorts an idea’s beliefs by how much the idea needs each one, and how much proof you have. Teams often draw it as a grid of sticky notes.

How much the idea needs a belief, and how little proof you have for it, is the belief’s risk. Test the riskiest belief first.

Most assumption maps stop at those two questions. Mia’s third question, what else falls if it’s false, is the one that broke the tie. Keep it.

Words from this lesson

Assumption map
The beliefs behind an idea, sorted by how much it needs each one and how much proof you have. In this lessonPeter's four beliefs, sorted by whether they could sink the idea and how much proof there is.
Risk
How much your idea needs a belief, and how little proof you have for it. In this lessonPlacing the order is the hard part: the idea needs it, and nobody has checked.

Back to your first answer

Your first answer

Your answer

Your answer will show here after you write one at the top of this page.

You've read the lesson. Would you change your answer?
Good. Which part of this lesson backs your answer?

Your bet: is three offices asking enough to start building?

Your bet

Mia plans her test for next week. But Peter, the head of sales, doesn’t want to wait.

Peter, head of sales, at the team meeting: Three offices asked for it. Can’t we just start building?

Is three offices asking enough proof to start? Make your bet, and say how sure you are.

Your guess
How sure are you?

You'll find the answer in Lesson 2.4: Match the decision to the evidence.

Try this during the week: take an idea from your own work. List three beliefs it rests on, and ask Mia’s three questions about each. Which belief sits lowest, with no proof?

Sources

  • David J. Bland and Alexander Osterwalder, Testing Business Ideas, 2019. The assumption map: a grid of how much the idea needs a belief, against how much proof you have.
  • Teresa Torres, Continuous Discovery Habits, 2021. Test the riskiest assumptions before you build.
  • The code history idea is from my own work. Tiffin, Peter, Mia, Nora and Ryan are made up for this course.
Portrait of Shirish Shikhrakar

Written by

Shirish Shikhrakar

Shirish is a Google-certified UX designer and UX engineer based in Kathmandu, Nepal. He started as a software developer, moved into product design for Silicon Valley startups, and has taught UX foundations to hundreds of designers since 2019.

AboutLinkedIn