# Fluff: vague answers and the say-do gap > How to spot vague answers (“I always”, “I would”, “I might”), why the say-do gap makes them useless, how to anchor them to one real past event, and how to tell complainers from customers. - Course: [Customer Development: Talking to Customers](https://productovi.com/courses/customer-interviews.md) - Module: Bad data - Lesson 5/15, ~40 min - Source: https://productovi.com/courses/customer-interviews/fluff ## Olivia, take two **Alex Tutor:** Back at CatChow, you're still interviewing customers about the food subscription. Olivia got a second call. **Sofia:** I called Olivia back and didn't mention the subscription once. No compliments this time! But my notes still feel… empty. **Alex Tutor:** Read me what she said. **Olivia:** Oh, I always buy food in advance, I'm very organised about it. If there were a delivery service, I'd definitely use it. I might even order treats as well. **Sofia:** Nothing wrong with it, right? But I can't use any of it. **Alex Tutor:** Because it's **fluff**: answers that sound like data but describe nothing that happened. Fluff lives in the **say-do gap**, the distance between what people say they do and what they actually do. Three sentences, three kinds: “I always…”, “I'd definitely…”, “I might…”. ![Three fluffy clouds float above a talking cat owner; the interviewer reaches for one and his hand passes straight through it.](https://productovi.com/courses/customer-interviews/inline/fluff-illustration-2.webp) *No date, no place, no money: if you can't picture the scene, there's nothing to hold.* **Alex Tutor:** Before I explain them, try sorting them yourself. > **Exercise:** Match each thing Olivia said to the kind of answer it is. ## Three kinds of fluff **Sofia:** The first one I know from module 1: a **generic claim**. The ideal version of herself. **Alex Tutor:** Right. “Always”, “usually”, “never” describe who she'd like to be. Real life has exceptions, and the problem usually lives in the exceptions. **Alex Tutor:** The second is a **future promise**: “I would”, “I will”. It's the most dangerous, because it sounds concrete. But people are hopelessly optimistic about their future selves. Everyone starts the gym next Monday. **Sofia:** And the third, “I might”? **Alex Tutor:** **Hypothetical** answers: “I might”, “I could”, “I'd probably”. Even softer than a promise: a polite way of saying “in some other world, sure”. **Alex Tutor:** All three share one thing: no date, no place, no money. If you can't picture the scene, it's fluff. > **Exercise:** Answers from Sofia's other calls. Which are **fluff**? Select all that apply. **Sofia:** Sorting them on paper is one thing. How do I catch fluff live, in the middle of a call? **Alex Tutor:** Use the scene test. Can you picture one moment: when, where, what it cost? A real event usually comes with a day, a place, often a number. **Alex Tutor:** Watch out for habits dressed up in detail. A vividly described routine is still a routine, not one event. **Alex Tutor:** Try the test on four answers. > **Exercise:** Use the scene test: which answer describes something you could actually picture happening? ## Generic and hypothetical questions that invite fluff **Sofia:** Honestly, I think I asked for it. My first question was “Do you usually buy food in advance?” **Alex Tutor:** Questions shape answers. “Do you usually…?”, “Do you ever…?”, “Would you…?”, “Could you see yourself…?” all invite fluff. **Sofia:** So I should never ask them? **Alex Tutor:** You can, sometimes. “Do you ever run out?” is a fine way to open a topic. The mistake is treating the answer as data. Use it as a bridge: if she says yes, ask about the last time. **Alex Tutor:** The rule: a fluffy question is fine only if your next question is about one specific time. > **Exercise:** “Do you ever run out of cat food?” is a useless question that should never be asked in an interview. ## Anchoring: ask about past behaviour **Sofia:** So after “I always buy in advance” I should have asked…? **Alex Tutor:** “When did you last buy food? Walk me through it.” That's **anchoring fluff**: tying a vague answer to one real past instance. Try it. Call her back. ![A person ties a floating cloud by a rope to a stake hammered into one square of a big calendar lying on the ground.](https://productovi.com/courses/customer-interviews/inline/fluff-illustration-3.webp) **Sofia:** Done! I said I'd forgotten to ask something. Here's what she told me: **Olivia:** Last time? Two weeks ago, at the shop by the metro. Well, the time before was a mess. My daughter was ill, I forgot completely, and on Saturday night the bag was empty. **Olivia:** I borrowed a bowl of food from my neighbour and ordered online with express delivery. Paid almost double. **Alex Tutor:** “I always buy in advance”… except when life happens, and then it costs her double. That's the pattern: the generic answer shows the ideal, the specific one shows the problem. **Alex Tutor:** Now you have two versions of Olivia in your notes. Which one would you hand to the product team? > **Exercise:** Two notes from the same customer. Which one gives CatChow's team something to decide on? ![All three kinds of fluff get the same follow-up, “When did you last buy food?”, and Olivia's “I always buy in advance” becomes a story of paying almost double.](https://productovi.com/courses/customer-interviews/diagrams/m2-l2-anchor-fluff.png) > **Exercise:** Put this conversation in order: how a fluffy start turns into a real story. ## Curious, not suspicious **Sofia:** It felt a bit awkward, as if I were checking whether she lies. **Alex Tutor:** Tone matters. You're not catching her out, you're curious about the details. “Oh, and how did it go last time?” sounds like interest, not an audit. **Alex Tutor:** A bit of self-irony helps too: “I always plan to buy ahead and never manage it. When did you last buy food?” It makes the messy version easy to admit. **Sofia:** And future promises? “I'd order every month if it were cheaper.” There's no “last time” for a product that doesn't exist. **Alex Tutor:** Anchor the part that already exists. She promises to switch for a lower price. Has she ever switched shops to save money? What does she pay now? Past behaviour predicts future behaviour better than any promise. > **Exercise:** A customer, Vicky, says: “I'd definitely order every month if it were 10% cheaper than the shop.” Write 1–2 questions that **anchor** this future promise to her real past behaviour. **Sofia:** Back to tone for a moment. How do I ask about the last time without sounding like an auditor? **Alex Tutor:** Show interest first, then ask: “Oh nice, and how did it go last time?” Never question whether the claim is true. **Alex Tutor:** And don't close the topic for them. “So you'd never need delivery?” invites a quick yes and ends the story before it starts. **Alex Tutor:** Pick the follow-up that sounds curious. > **Exercise:** Olivia says: “I always buy food in advance, I'm very organised.” Which follow-up anchors this and sounds curious, not suspicious? ## Complainers vs customers **Megan:** Good news from my team! One man spent ten minutes complaining about pet-shop prices: “Someone should finally do cheap food delivery!” That's a hot lead, right? **Alex Tutor:** Maybe. Did anyone ask whether he's looked for cheaper delivery? **Megan:** Hmm. Yes, actually. He said no, never got round to it. And that there are probably loads of them already. **Alex Tutor:** Then he's a **complainer**, not a customer. He talks about the problem but hasn't lifted a finger to solve it. If he never looked for a fix, he won't look for yours either. ![Left: a man lounging on a sofa complains toward an empty cat bowl. Right: a woman compares three cat-food boxes on her desk, a delivery app open on her phone.](https://productovi.com/courses/customer-interviews/inline/fluff-illustration-4.webp) *A real customer has already tried something; a complainer only talks.* **Sofia:** And what would a customer sound like? **Alex Tutor:** Like someone who's already tried things: compared three shops, tested a delivery app, dropped one because of the courier. Searching, spending and switching are the signs. > **Exercise:** Four people all say buying cat food is annoying. Who is most likely a real customer for a delivery service? ## Testing a hot lead **Megan:** OK, but my team hears complaints like this every day. How do they tell who's who during the call? **Alex Tutor:** Don't argue with the complaint; agree it sounds annoying. Then ask what he's done about it: what he's tried, what he's looked for, what it costs him now. **Sofia:** And if it's all “no”, we mark him as a complainer and move on? **Alex Tutor:** Yes, politely. It's still a useful call: now you know his complaint never turns into action. Now write the questions for Megan's team. > **Exercise:** A caller tells Megan's team: “Pet shops are a rip-off! Someone should finally do cheap food delivery.” Write 1–2 questions that show whether he's a **complainer** or a potential customer, without arguing or pitching. **Sofia:** Some complaints sound exaggerated. Should Megan's team push back, like “Is it really that bad?” **Alex Tutor:** No. Arguing makes people defend their story and dig in. Agree that it sounds annoying, then ask what they've done about it. **Alex Tutor:** How bad it really is shows in their actions, not in a debate. Searching, spending and switching tell you more than any counter-argument. **Alex Tutor:** Let's lock that in. > **Exercise:** When a customer complains loudly about a problem, the interviewer should first question whether it's really that bad, to filter out exaggeration. ## A whole call of fluff **Sofia:** One more call to review: Andrew. It's from last week, before our lessons, and reading it now, it's almost pure fluff. **Alex Tutor:** Then it's perfect practice. Mark every place where you could have anchored, and what you'd ask. **Sofia:** And separately, what we actually learned, even if it's tiny. **Alex Tutor:** Especially if it's tiny. One real fact beats a page of “usually”. > **Exercise:** Analyse Sofia's call with Andrew: find the fluff, the missed moment and what the team really learned. ## What to remember **Sofia:** So fluff isn't lying. It's just blurry. **Alex Tutor:** Exactly. People describe their best selves and best intentions. Your job is to bring the picture into focus: one date, one place, one story. > **Key ideas** > > - **Fluff** sounds like data but describes no real event. It lives in the **say-do gap**: what people say vs what they do. > - Three kinds: generic claims (“always”, “usually”, “never”), future promises (“I would”, “I will”) and hypothetical maybes (“I might”, “I could”). > - Fluffy questions (“Do you ever…?”) are fine as a bridge, never as the answer. > - **Anchor** it: “When was the last time? Walk me through it.” For promises, anchor what already exists: current spending, past switching. > - A **complainer** talks about a problem; a customer has already tried to solve it. **Alex Tutor:** Next lesson: the third kind of **bad data**, and the most tempting one. Customers start designing your product for you. - Previous lesson: [Compliments and how to deflect them](https://productovi.com/courses/customer-interviews/compliments.md) - Next lesson: [Feature requests: digging for the real problem](https://productovi.com/courses/customer-interviews/feature-requests-and-digging.md) --- Productovi Learning. Free, no sign-up needed to read.