Book Review: The Mom Test

Book Author: Milo Frank

When you are trying test a new product idea it is difficult to cull the signal from the noise. This book is called the Mom Test because it helps you run interviews where even your Mom won’t lie to you.

Customers, and people in general will tell you what you want to hear rather than tell you the truth. Even more dangerous people don’t often know what they really need and will lead you down a the wrong path by talking about what is easy to talk about. You have to avoid leading questions, dig down to facts, read between the lines, and triangulate on real problems in order to build a product that fulfills needs instead of painting over symptoms.

First, don’t ask leading questions. When you are asking for feedback, don’t tell them about your idea! This puts the focus on your idea instead of what problems exists. This will bias them and makes it hard to contradict you. They will tell you what you ask them to tell you.

There is a great conversation in the book in the opening chapter and translates what you say to subtext. Won’t repeat it here but highly recommend reading that conversation.

In any case, the goal isn’t to get opinions – it’s to gather facts and information about behavior and problems. Instead of asking if they’d like a cookbook app, you can ask when the last time they bought a cookbook was or what the hardest part about cooking is. What frustrated them most in the last week? What took the longest. Stick to concrete and specific facts.

Disonfirming data is good! It prevents you from going down a deadend. Don’t be attached to your ideas (embrace stoicism).

Three simple rules of the mom test:

• ask about their life, not your ideas
• ask about specific past actions and not future opinions
• listen

Rule of thumbs:

• customer conversatons are bad by default you must mindfully fix them
• if you’ve mentioned your idea people will protect your ego
• anyone will say your idea is great if you are annoying enough about it
• the more you talk the worse you are doing
• you should be terrified of at least one of the questions you are asking in every conversation

Three types of bad data:

• compliments (and other lies to protect your ego)
• fluff (generics, future, opinions)
• ideas

Three types of fluff:

• future tense promises – I would, I will…
• hypothetical maybes – I might, I could…
• generic claism – I like usually, I never, I sometimes…

Fluff is vague and nebulous.