Chapter 2: Find out if you're wrong, before you build, lesson 6 of 7

Test your skills

Check what you learned in chapter 2. Four rounds: the key words, the evidence ladder, and then all five lessons at once on a new parking app.

By 1 min read, 20 min of exercises

You have learned five things in this chapter. This lesson checks them, in four rounds.

The first three rounds are the key words and the evidence ladder. You meet each one in a new moment, on a plant-care app. The last round uses all five lessons at once, on a parking app you haven’t seen. You get one score at the end, and you can try again.

Four rounds, one score

Skills test

29 questions in 4 rounds. Answer each one, then check it. Your score comes at the end.

Round 1 of 4

Which word fits?

Each one is a moment from Jo's plant-care app. Pick the word that fits the gap.

Question 1 of 12

Jo's team believes people water their plants on Sundays. Nobody has checked. That belief is  .

Tap a word, or drag it into the gap. Then check it.

Question 2 of 12

Jo only asks happy customers about the new reminders, and they all like them. That's  .

Tap a word, or drag it into the gap. Then check it.

Question 3 of 12

"Our plants will feel happier" can't be proved wrong by anything you could count. It isn't  .

Tap a word, or drag it into the gap. Then check it.

Question 4 of 12

Jo writes: reminders will help new users water on time. We're wrong if under 3 in 10 use them. That's a  .

Tap a word, or drag it into the gap. Then check it.

Question 5 of 12

"Fewer than 3 in 10 new users turn reminders on" is Jo's  .

Tap a word, or drag it into the gap. Then check it.

Question 6 of 12

"We decide on 30 November, whatever the numbers say" is the test's  .

Tap a word, or drag it into the gap. Then check it.

Question 7 of 12

Jo lays out every belief by how much the app needs it and how much proof there is. She made  .

Tap a word, or drag it into the gap. Then check it.

Question 8 of 12

The app can't work without a belief, and nobody has any proof for it. Its   is high.

Tap a word, or drag it into the gap. Then check it.

Question 9 of 12

Jo turns reminders on for 50 users, and can switch them off in a minute. That's a  .

Tap a word, or drag it into the gap. Then check it.

Question 10 of 12

Two loud users asked for a plant shop, and Jo forgot the 5,000 who said nothing. That's  .

Tap a word, or drag it into the gap. Then check it.

Question 11 of 12

A Buy seeds button leads to a "coming soon" note, and counts every tap. It's a  .

Tap a word, or drag it into the gap. Then check it.

Question 12 of 12

Jo texts five users their watering days by hand, before building any reminders. That's a  .

Tap a word, or drag it into the gap. Then check it.

See the answers
  • an assumption: It's a belief the app depends on, and nobody has checked it yet.
  • confirmation bias: She only looked where she expected a yes. Nothing could show her she was wrong.
  • falsifiable: No result could ever go against it. So you can't test it.
  • hypothesis: It's a belief, written with the result that would prove it wrong.
  • kill criterion: It's the result, chosen before she looks, that means stop.
  • timebox: It's a real date to decide on, not an open door.
  • an assumption map: It sorts the beliefs behind an idea, so she can see which to test first.
  • risk: Risk is how much the idea needs a belief, and how little proof you have.
  • reversible step: It's cheap, and easy to undo if the proof turns out wrong.
  • what you see is all there is: She judged from the two in front of her, and forgot everyone she couldn't see.
  • fake door: It counts how many want something before it exists.
  • concierge test: She does the service by hand for a few people first.

Round 2 of 4

Could it prove Jo wrong?

Jo believes new users forget to water their plants. Could each result prove her wrong? Put each card in its group.

Card 1 of 4

Five friends say they forget to water sometimes.

Drag the card to a group, or tap a group. Then check it.

Card 2 of 4

The app's log shows most new users water on time in their first month.

Drag the card to a group, or tap a group. Then check it.

Card 3 of 4

Jo watches her own plants dry out on holiday.

Drag the card to a group, or tap a group. Then check it.

Card 4 of 4

Half the new users get no reminders, and they water just as often.

Drag the card to a group, or tap a group. Then check it.

See the answers
  • Can only agree: Five friends say they forget to water sometimes.
  • Could prove her wrong: The app's log shows most new users water on time in their first month.
  • Can only agree: Jo watches her own plants dry out on holiday.
  • Could prove her wrong: Half the new users get no reminders, and they water just as often.

Round 3 of 4

How strong is the proof?

Jo wants to know if new users forget to water. Put this proof in order, from weakest to strongest.

Weakest proof

  1. Jo thinks new users forget to water.
  2. One user tells Jo her fern died while she was busy.
  3. In a survey, 8 in 10 users say they sometimes forget.
  4. Jo watched 5 users for a week. 3 missed a watering.
  5. The app's log shows 6 in 10 plants miss a watering in the first month.
  6. Half the new users got reminders, and they missed far fewer waterings.

Strongest proof

Drag a card up or down, or use its arrows. Then check your order.

See the right order
  1. Opinion: Jo thinks new users forget to water.
  2. One story: One user tells Jo her fern died while she was busy.
  3. What people say: In a survey, 8 in 10 users say they sometimes forget.
  4. What you saw: Jo watched 5 users for a week. 3 missed a watering.
  5. What you counted: The app's log shows 6 in 10 plants miss a watering in the first month.
  6. A fair test: Half the new users got reminders, and they missed far fewer waterings.

Round 4 of 4

Use it all on a new case

7 questions on a product you haven't seen in this chapter. Say how sure you are each time.

Alex runs a parking app for a big city. His boss wants a new feature this quarter: drivers pay ahead to reserve a parking spot. Alex's team knows three things.

  • Ten drivers emailed to ask for reserved spots.
  • The app's records show 7 in 10 drivers find a spot within 5 minutes.
  • Alex's boss believes the feature will double the app's income.

Question 1 of 7

Which of these is a belief the feature needs, not something already true?
How sure are you?

See the answer

Drivers will pay ahead to hold a spot. The emails and the boss's request already happened. Paying ahead is a bet nobody has checked. Reread: What Peter’s email takes for granted, in lesson 2.1

Question 2 of 7

Which line could prove Alex's team wrong?
How sure are you?

See the answer

Fewer than 1 in 10 drivers reserve a spot by 30 November. It has a number you could count, and a date. The other two can never come out against the team. Reread: The two questions most tests get wrong, in lesson 2.2

Question 3 of 7

Two beliefs: drivers struggle to find a spot, and drivers will pay to reserve one. If drivers don't struggle, what happens to the second belief?
How sure are you?

See the answer

It stops mattering, because nobody pays to fix a problem they don't have. Paying to reserve rests on struggling to park. If the lower belief falls, the one above it falls too. Reread: Question 3, in lesson 2.3

Question 4 of 7

The records show 7 in 10 drivers find a spot within 5 minutes. Where does that sit on the evidence ladder?
How sure are you?

See the answer

Rung 4: what you counted. It's counted from what drivers did. And it goes against the belief that drivers struggle. Reread: How strong is the proof for group ordering?, in lesson 2.4

Question 5 of 7

The team's proof is ten emails. Which step can that proof carry?
How sure are you?

See the answer

Show a sketch to the ten drivers, and look again in two weeks. Ten emails is weak proof. A sketch is cheap and easy to undo, and it comes with a date to look again. Reread: Size the step to the proof, in lesson 2.4

Question 6 of 7

Alex wants to know how many drivers want to reserve enough to tap for it. Which one-day test fits?
How sure are you?

See the answer

A Reserve button that counts taps, then says it's coming soon. A fake door counts what drivers do. Asking only gets kind words, and building takes far longer than a day. Reread: Five tests you can run in a day, in lesson 2.5

Question 7 of 7

Alex plans to raise prices for half the drivers at random, without telling anyone. What should he ask first?
How sure are you?

See the answer

Could the result prove us wrong, and who agreed to it?. Drivers would pay for a test they never agreed to. Ask who agreed, and who pays if it goes wrong. Reread: Who pays if a test goes wrong?, in lesson 2.5

Done

Your score

    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.