# How do you differentiate between what customers say they want and what they actually need? Give me an example.
> How to answer the question about what customers say they want versus what they need, with a story structure, pitfalls, and a spoken meal-kit example.

Source: https://mokk.ai/pmm-questions/market-research/what-customers-say-vs-need (updated 2026-09-22)
Markdown: https://mokk.ai/pmm-questions/market-research/what-customers-say-vs-need.md

Say in one sentence that you trust behavior over stated preference, then prove it with one story. Name the request in the customer’s words, the behavior that contradicted it, and the underlying need you found. Say what changed because of it and how you measured the result. The strongest answers also admit the moment the team nearly built what customers asked for, which shows the method working under pressure.

## What the wants-versus-needs question is checking

Customers describe solutions. Product marketers are paid to find the problem behind them, and this question asks for evidence that you can. It scores mainly on market expertise, where the rubric’s Strong level mentions understanding buyer motivations beyond stated needs, and on analytical rigor, because your method has to hold up when the interviewer pushes on it.

The words “give me an example” matter. A tidy philosophy of customer research followed by no story is the most common way this answer loses points. Spend about twenty seconds on method and the rest on one case.

## Choosing a story where the request and the need came apart

- **A request you can quote.** Something customers asked for in their own words, repeatedly, such as more filters, more variety, or a lower price.
- **Behavior that contradicted it.** Usage data, a workaround, or something you watched people do.
- **A different response.** The need led you to build, say, or sell something other than the literal request.
- **A result.** A number if you have one, or a clear change in how customers or sales behaved.
- **A moment of temptation.** Ideally the team was ready to ship the request, which makes the insight cost something.

Seniority matters less than evidence here.

## The order a say-versus-need answer should follow

1. One sentence of method, for example that you ask about the last time someone tried to do the job and what happened.
2. The request, in the customers’ words, and how often it came up.
3. What you saw that didn’t fit, with its source.
4. The need you named, and how you checked it against more than one source.
5. What changed, and the result.
6. What you would do sooner next time.

Make the method sentence match the story. If you claim to watch behavior, the evidence in the story should be behavior, and another round of survey answers won’t do. The [quant and qual question](https://mokk.ai/pmm-questions/market-research/quant-and-qual-together) covers the data side of this method in more depth.

## A meal-kit example, told out loud

**Strong answer (invented example)** "I trust what people did last week more than what they say they’d like. At a meal-kit subscription, the exit survey had one clear top answer, which was more recipe variety. The menu team was ready to double the weekly options. Before they did, I pulled order history for about two thousand cancelled accounts and saw something odd. People rarely cancelled after a boring week. They cancelled after skipping two weeks in a row, and the skipped weeks were heavy on recipes over forty minutes. So I called fifteen people who had cancelled. Most said variety when I asked why they left. When I asked about the last box they didn’t cook, the story changed. It was a Tuesday, the recipe took an hour, and they ordered takeout. The need was a weeknight dinner they could trust to take half an hour. We added a thirty-minute badge, let people fill a box with only those recipes, and led our win-back emails with it. Skips fell by about a fifth the next quarter. Next time I’d pull the behavior data before the survey results start a debate."

The pivot comes from asking about the last box. A specific past event turned “variety” into a problem the team could solve.

## How well-prepared answers to this question still lose points

### The method lecture

**Where it goes wrong (invented example)** "I rely on jobs-to-be-done interviews, because customers describe solutions and not problems. I ask about the last time they hired a product, I listen for workarounds, and I triangulate with usage data before I act on anything."

Everything in it is right, and there is no example. The interviewer has to ask for one, which costs you control of the pacing. Cut the method to a sentence and get to the story.

### The need that is the request in disguise

Customers asked for faster exports, and the need turned out to be faster exports. A story like that shows diligence without insight. Choose a case where the need pointed to a different response.

### The customer who didn’t know better

Lines like “customers never know what they want” land badly with interviewers who run research. Treat the survey answer as a real signal that needed interpreting.

## Follow-ups after a wants-versus-needs story

Follow-ups on a story about stated wants and real needs
| Follow-up | What sits behind it | How to respond |
| --- | --- | --- |
| How many customers did you need before you believed it? | Comfort with small samples | Give the count and name the second source that confirmed it |
| What if customers kept insisting on the original request? | Whether you dismiss customers | Say you shipped a small version or explained the trade-off, then watched what they used |
| How do you avoid leading questions? | Interview craft | Ask about specific past events, never about hypothetical features |
| How did you get the team to drop its original plan? | Influence | Put the behavior data next to the survey result and let the gap make the case |

## More questions in this category

- [Market sizing](https://mokk.ai/pmm-questions/market-research/market-sizing)
- [Building buyer personas](https://mokk.ai/pmm-questions/market-research/buyer-personas)
- [Analyzing a funded competitor](https://mokk.ai/pmm-questions/market-research/analyze-well-funded-new-entrant)
- [Useful sales battlecards](https://mokk.ai/pmm-questions/market-research/battlecards-sales-will-use)
- [All market research questions](https://mokk.ai/pmm-questions/market-research)

## Frequently asked questions

**Is it a mistake to ask customers what they want?** No. Their answers are useful signals about what hurts. The mistake is treating the suggested fix as the specification, so check stated wants against what people do.

**Can the example come from a product I didn’t market?** Yes. A story from support, sales, or product work shows the same skill if you ran the discovery. Make clear which part was yours.

**How long should the wants-versus-needs story take?** About two minutes. Spend most of that on the evidence and the change, because those are what interviewers follow up on.

## Practice

**Answer it out loud.** A Mokk session asks this question, follows up on your answer, and scores it on five competencies.