# Why “do you like my idea?” is useless > Why asking customers if they like your idea gets polite yeses, how false positives make a team confidently wrong, and how to tell opinions from facts. - Course: [Customer Development: Talking to Customers](https://productovi.com/courses/customer-interviews.md) - Module: Why customers lie - Lesson 1/15, ~43 min - Source: https://productovi.com/courses/customer-interviews/do-you-like-my-idea ## Where you work **Alex Tutor:** Welcome to the course. You've just joined the product team at **CatChow**, a small online shop of about thirty people that sells cat food, treats and litter. Next on the team's list is a food subscription: a monthly delivery, a bit cheaper than buying in the shop. **Alex Tutor:** Before anyone builds it, you'll check whether people actually want it. Over this course you'll learn to interview customers so they tell you what really happens in their lives, not what you hope to hear. Today we start with the question almost everyone asks first. ## The validation plan **Alex Tutor:** So, the subscription. How will you find out whether people want it? What's the plan? **Sofia:** Simple! I'll call twenty of our customers, describe the subscription and ask if they like it. If most say yes, we build it. **Alex Tutor:** And what do you expect to hear? **Sofia:** Yes, obviously. Megan already asked people at the pet expo last weekend. Look what she sent to the whole team: **Megan:** Great news from the expo! 38 of the 40 people at our booth said they'd subscribe. The idea is validated. Let's aim for a March launch! **Alex Tutor:** Picture the booth: free samples in hand, a smiling rep asking if they'd like her new thing. What did it cost them to say yes? ![At a pet-expo booth stacked with blue cat-food bags, a smiling rep hands out free samples; visitors nod and give thumbs up while a cat peeks out of a carrier.](https://productovi.com/courses/customer-interviews/inline/do-you-like-my-idea-illustration-2.webp) *Saying yes at a friendly booth costs nothing, so it proves very little.* **Sofia:** Um… nothing? **Alex Tutor:** Nothing at all. Hold that thought. First, let's see how you'd read Megan's numbers. > **Exercise:** Megan reports: “38 of the 40 people at our booth said they'd subscribe.” What does this result tell the team most reliably? ## Why nice people give you wrong answers **Sofia:** But why would people lie? Take Olivia: she's been buying from us for two years. She has no reason to mislead me. **Alex Tutor:** She has a very good reason: she likes you. You arrive excited about your idea, and saying “meh” feels rude. So she gives you the pleasant answer instead of the true one. **Alex Tutor:** It isn't malicious; usually she doesn't even notice. You're asking her to predict her behaviour in a future that doesn't exist yet, and people guess generously when someone nice is listening. **Sofia:** So I should ask strangers instead, people who don't care about my feelings? **Alex Tutor:** Strangers are polite too: remember the expo. The problem isn't who you ask, it's what you ask. Nobody can honestly answer “Is my idea good?” Only the market can, when people pay. **Alex Tutor:** What you need are **three rules for questions that get honest answers**. Follow them, and even someone who loves you and wants to be kind can't mislead you with their answer. We'll go through the rules in the next lesson. > **Exercise:** The main fix for polite, misleading answers is to interview strangers instead of friends, family and loyal customers. **Sofia:** So they're not lying to me on purpose? **Alex Tutor:** Almost never. Most people believe their own “I'd buy it” while they say it. They just can't know the future, and they round up for you. **Sofia:** So catching them out won't help. Changing the question will. **Alex Tutor:** Exactly. Let's make sure that one sticks. > **Exercise:** Most customers who give you a misleading “yes” are lying on purpose, to end the call quickly. ## False positives: worse than no data **Sofia:** Fine, the answers are a bit inflated. But surely some data is better than no data? **Alex Tutor:** That's the trap. When you know nothing, you stay careful: small tests, small bets. When you have 38 enthusiastic yeses, you feel certain. You order packaging, book a courier contract and plan the March launch. **Alex Tutor:** That's a **false positive**: evidence that says “go” when the real answer is “no”. It's worse than ignorance, because it raises your confidence without raising your chances. ![A cheerful person with a cat-food bag under one arm strides toward a bright doorway that is only painted on a brick wall, while a cat watches doubtfully.](https://productovi.com/courses/customer-interviews/inline/do-you-like-my-idea-illustration-3.webp) *A false positive looks like an open door and leads straight into a wall.* **Alex Tutor:** And you'd only find out at launch, when hardly anyone subscribes and three months of work are gone. A badly asked question is worse than no question at all. ![A free “yes” at the booth became 38 of 40 yeses, the team placed big bets on them, and at launch hardly anyone subscribed: confidence rose, chances didn't.](https://productovi.com/courses/customer-interviews/diagrams/m1-l1-false-positive.png) **Alex Tutor:** Let's trace how a nice chat turns into an expensive mistake. > **Exercise:** Put the steps in order: how one polite answer grows into a failed launch. **Sofia:** Is every wrong signal a false positive? Say customers hate an idea and it succeeds anyway? **Alex Tutor:** That's the mirror image, a **false negative**: the signal said “no” and the answer was “yes”. It's rarer in interviews, because people are polite rather than harsh. **Alex Tutor:** And having no data isn't a false positive either. No data keeps a team careful. A false positive makes it confident and wrong. **Alex Tutor:** Now try to spot a false positive among situations that look alike. > **Exercise:** Which of these situations is a **false positive**? ## Opinions vs facts **Sofia:** OK, but how do I tell a useful answer from a polite one while I'm still on the call? **Alex Tutor:** Sort every answer into two piles. **Opinions**: what someone thinks, likes or says they would do. **Facts**: what actually happened. What they did, bought, paid, forgot, complained about. ![Two piles on a desk: fluffy speech bubbles with hearts on the left; a crumpled receipt, an empty cat-food bag and a calendar page on the right.](https://productovi.com/courses/customer-interviews/inline/do-you-like-my-idea-illustration-4.webp) *Sort every answer: what people think goes left, what actually happened goes right.* **Alex Tutor:** “I love the idea” is an opinion. “I'd order every month” is a guess about the future: same pile. “Last month we ran out on a Sunday and I drove to a 24-hour pet shop” is a fact. **Sofia:** The last one doesn't even mention our subscription. **Alex Tutor:** That's why you can trust it. She can't flatter an idea she hasn't heard. And look how much it tells you: the problem is real, it happened recently, and she already spends time and fuel solving it. **Alex Tutor:** Let's practise the sorting. > **Exercise:** Six things cat owners told Sofia. Which are **facts** you can use as evidence? Select all that apply. **Alex Tutor:** One more, closer to a real call. Here are two things Olivia might say. Which one goes into your evidence? > **Exercise:** Two things Olivia might tell you on a call. Which one can CatChow use as **evidence** about the subscription? ## The first call, replayed **Sofia:** Confession: I called Olivia this morning, before we talked. Here's exactly what I said: “We're launching a subscription: food delivered every month, 10% cheaper. Do you like the idea?” **Olivia:** Oh, that sounds lovely! Whiskers would be thrilled. I'd definitely sign up. Will there be treats in the box too? **Sofia:** I wrote down “strong yes, wants treats”. Please tell me that's useful. **Alex Tutor:** Let's count the facts. “Sounds lovely” is a **compliment**. “Whiskers would be thrilled” is a guess on the cat's behalf. “I'd definitely sign up” is a **future promise**. And the treats? **Sofia:** A feature idea? **Alex Tutor:** A **feature request**, and mostly her being a good conversation partner. Zero facts, three kinds of noise. We'll handle each in module 2; for now, just learn to spot them. > **Exercise:** Match each line from Sofia's calls to what it really is. ## What to ask instead **Sofia:** So what should I have asked her? **Alex Tutor:** Something about her life, not our idea: “When did you last buy food for Whiskers? How did that go?” There's no polite lie for that. Either there was a hassle or there wasn't. **Sofia:** Why exactly is there no polite lie for that? **Alex Tutor:** Because it's a memory, not a judgement. There's nothing of yours in the question to praise, so she just tells you what happened. **Alex Tutor:** Before you write your own, pick the question that gets an answer nobody can fake. > **Exercise:** Sofia wants to learn whether running out of cat food is a real problem for Olivia. Which question gets an answer that can't be a polite guess? **Sofia:** And if there was no hassle at all? **Alex Tutor:** Then you've learned in five minutes what a launch would teach you in five months: maybe food runs just aren't a problem for her. That's not a failed call. That's a saved budget. **Sofia:** Five minutes on a call versus five months on a launch. I'll take the five minutes. **Alex Tutor:** Try it yourself: take the question you were planning to ask and turn it into one about her past. > **Exercise:** Sofia's planned question is: “Would you use a subscription that delivers cat food to your door every month?” Rewrite it as one or two questions about the customer's **real past behaviour** that don't mention the subscription. ## Breaking the news to Megan **Sofia:** Megan's going to be upset. She's already told the CEO the idea is validated. **Alex Tutor:** Then help her rather than prove her wrong. She's right that people were friendly. She's wrong about what friendliness proves. Keep it short and don't lecture. **Sofia:** Something like “it's not you, it's the question”? **Alex Tutor:** Good start. Anyone would collect 38 yeses with that question. Now write the actual message, and suggest what to do next. **Alex Tutor:** Before you do, compare two drafts. Which one would make Megan check the idea rather than defend it? > **Exercise:** Megan has already told the CEO the subscription is validated. Which message is more likely to get the team to check properly? **Alex Tutor:** Now your own version: short, friendly, with a next step. > **Exercise:** Write a short, friendly message (2–4 sentences) to Megan explaining why “38 of 40 said they'd subscribe” doesn't validate the subscription, and what the team should do instead. ## What's really in the notes **Megan:** Fair enough, I got carried away. But then how do we know anything? Do we cancel March? **Alex Tutor:** We don't cancel, we check properly. You made three more calls this week. What's in your notes? **Sofia:** Mostly “loves it”, I'm afraid. Although one man complained a lot about delivery. **Alex Tutor:** Complaints about real events are gold. Read the notes as if a stranger wrote them. > **Exercise:** Analyse Sofia's notes and write what the team actually knows so far. ## What to remember **Sofia:** So the problem was never Olivia or the expo visitors. It was me asking them to judge my idea. **Alex Tutor:** Exactly. People are kind, and the future is easy to promise. Ask about their real life and there's nothing left to be polite about. > **Key ideas** > > - “Do you like my idea?” gets polite answers, not honest ones. > - A **false positive** is worse than no data: the team becomes confident and wrong. > - Opinions and promises aren't evidence; facts about the past are. > - Ask different questions, not different people. **Alex Tutor:** Next time: the three rules for questions that get honest answers, and how to follow them on a real call. - Next lesson: [Three rules for questions that get honest answers](https://productovi.com/courses/customer-interviews/three-rules-for-honest-answers.md) --- Productovi Learning. Free, no sign-up needed to read.