In 1974, a psychologist named Elizabeth Loftus showed a group of people a film of a car accident. Afterward, she asked half the group one question: how fast were the cars going when they hit each other? The other half got the same question with a single word swapped: smashed instead of hit.

The smashed group estimated higher speeds. Same film. Same crash. One word.

A week later, everyone came back and answered a new question: Did you see broken glass? There was no broken glass in the film. The smashed group remembered it anyway.

One word inside one question changed what people reported seeing. Then it reached back and rewrote what they remembered.

Somebody did this to you this week. A question steered your answer, and you never noticed.

Last week, I showed you that your brain cannot refuse a question. You hear one, you start answering, whether you agreed to or not. This week is the other side of that power. If every question forces an answer, then how the question is built decides which answer you get back. Questions have an anatomy. Almost nobody looks at it.

Before we're done, a motorcycle taxi in Bangkok is going to show you what a question built right can find. Let's get into it. =====

Every question you ask carries three things, whether you put them there on purpose or not.

It carries assumptions, the things it treats as already settled. Ask "Why did the launch fail?" and you've ruled the launch a failure before anyone speaks.

It carries scope, the range of answers it permits. "Coffee or tea?" permits exactly two answers. "What should we drink?" opens the room.

And it carries a load, the specific words that steer the answer. That's what "smashed" did. Nobody in that study felt steered. The steering is invisible to the person answering, and most of the time to the person asking too.

Once you see the parts, you can take any question apart. Take "why did the launch fail?" one more time: it assumes failure before anyone answers, its scope is only explanations, and "fail" is the loaded word doing the steering. Hold on to that one; we'll rebuild it later.

Let's start with the ones built to do damage.

The worst question in business

"That presentation was fantastic, wasn't it?"

That's a tag question: a statement dressed up as a question, built so every answer is closed off except agreement. The person asking isn't really asking, not in any way that risks hearing something different. They're collecting a signature. When lawyers use them in court, it's called leading the witness. When managers use them in conference rooms, it's called alignment.

Tag questions did damage for years at HP, in the design reviews, every product went through before customer briefings. One review was for a prototype, one of our first laptops in what's now called the "thin and light" category. The product team wanted to lead that category without straying too far from what already sold.

That's the tradeoff tag questions live in. Nobody wants to sound negative about it. So the pushback arrived dressed as agreement: "We'll still have four USB ports, right?"

You don't get thin and light with a case full of legacy ports, and the designer knew it. I watched the exasperation cross his face. Before it became a battle, I stepped in and rebuilt the question: "What's the right mix of ports for this segment, and why?" Then: "Have we tested that mix with the target customers?"

The answer came back fast. "No need. We know what the customer wants." I nearly smacked my forehead. One rebuilt question had uncovered the real problem, and it wasn't ports. It was an untested assumption sitting in the middle of a flagship product plan.

That's what tag questions cost you. The entire point of asking a question is to get information, input, or ideas. A tag question collects compliance instead. Enough of them, and people stop bringing you anything you don't already believe.

The two kinds of good questions

Real questions are split into two categories: factual and investigative.

A factual question retrieves information. How many units did we sell last week? You may not know the answer, but you know exactly how to get it: one phone call. Factual questions keep the world running. They just can't uncover anything because they only retrieve what somebody already knows.

An investigative question can't be answered with a yes, a no, or a lookup. It's divergent: more than one correct answer exists, so the person answering has to go investigate. You felt this difference last week. "What is half of thirteen?" is a factual question. "How many ways could you answer: what is half of thirteen?" is an investigative question. One is arithmetic. The other is the one that a classroom answered thirty-two different ways.

Socrates built his entire way of teaching on investigative questions, pushing every student past "I've heard it said that..." until they could say what they themselves thought, and why. The first step toward knowledge, he insisted, is admitting yours is incomplete.

So here's the definition I've worked from for years. A great question makes the other person genuinely think before they answer, and it surfaces an insight that had been eluding them. Both halves matter. The first half costs you real effort. The second half is the payoff: you see something you've been looking at for years and never actually seen, and it surfaces possibilities you'd never have found on your own.

How to build one

How I build them is with four moves, in the order I use them. You already watched three of them work in that design review.

Strip the answer out. Read your question and find the smuggled conclusion. "Right?" and "wasn't it?" come off first. Then the buried assumptions. "Why did the launch fail?" becomes "what happened when we launched?"

Open the exits. If it can be answered with yes, no, or a single lookup, it's factual. That's fine when information is all you need. When you want discovery, rebuild it until more than one correct answer exists. The rebuild is usually small. "Should we do this?" becomes "what are the ways this could work, and what are the ways it dies?"

Weigh every word. This is Loftus's lesson. "How do we make this cheaper?" and "what happens if we triple the price?" point at the same product, and they will never produce the same ideas. There is no neutral wording. Every word carries load, so choose it deliberately instead of inheriting it.

Aim it where nobody is looking. If you can already predict the answer, the question is just decoration. Point it at something unexamined: the assumption everyone stopped checking, the customer nobody ever talks to. A great question should surprise you.

What a well-built question uncovers

Late 1994, Bangkok. Half an hour stuck in traffic, twenty-eight minutes until a critical meeting. At eighteen minutes, I gave up, jumped out of the limo, and hailed a motorcycle taxi. 125ccs of terror, flat out through the heat, my briefcase clutched to my chest.

Somewhere between the tuk-tuks, I started thinking about the machine underneath me. Cheap, efficient transportation. The streets were flooded with them. And on the other side of the world, Harley-Davidson was selling motorcycles as big-ticket luxury. A product that began as cheap transportation for the American working man had become a premium item that people waited months for.

In that evolution, somebody had to ask a question like this: What if we stop trying to make this cheaper and go the other way? Everything needed to see that opportunity had been lying in plain sight for decades, waiting for someone to ask for it.

That's the payoff of this craft. The next discovery in your business is probably sitting in an assumption nobody's checked in years, waiting on the right question to find it.

Practice Exercise: Before your next meeting, write down the one question you most need to ask, exactly as you'd naturally say it. Then take it apart on paper. What does it assume? What answers does it permit? Which words are steering, and in what direction? Rebuild it with the four moves and ask for the rebuilt version instead. Pay attention to what comes back that the original never would have surfaced.

Conclusion

This is part 2 of a three-part series on questions as the skill behind better thinking, better ideas, and better innovation. Next in on the specific questions I spent twenty years collecting and testing, the ones that reliably uncover what everyone else misses.

Full show notes for this episode are up at philmckinney.com. And if you want more than the show episodes, check out my articles over on Substack, including the first four chapters of my next book.

Subscribe on YouTube or wherever you get your podcasts so you don't miss the next episode, and don't forget to hit the notif...