My Ad Age Op-Ed: Why 'I Don't Know' Matters

This op-ed first ran in Ad Age on October 29, 2026.

Here's a business plan no focus group would have signed off on: charge people to nap five minutes from their own bed.

Yet when I took a Kellogg's team exploring the concept of energy to a nap pod salon in New York, the staff let us in on a secret: some of their most loyal customers lived nearby.

Why pay to nap around the corner? Because at home someone always needs you, whether it's the kids or the inbox. The pod had privacy, soft light, music and a scent piped in, and twenty minutes later people walked out energized for the rest of the day.

What they bought was a little time when nobody needed anything from them.

That's the kind of need you miss when you only ask people what they want, and it's why the rush to AI-generated respondents worries me. We're getting very good at generating answers. I'm not sure we're getting better at discovering what we didn't already know.

The trouble with always having an answer

"I don't know" can be worth more than a confident response. It usually means you've hit something people haven't worked out for themselves yet.

People rarely like admitting it, least of all when a marketer is asking. Put us in front of a survey or a focus group moderator and we want to be helpful and sound consistent. So we come up with sensible explanations for behavior we may not have thought much about, and those explanations aren't always a guide to what we'll actually do.

Synthetic respondents are appealing: fast and always ready with an answer. But they're shaped by existing data and the assumptions we feed them. Ask them about a category we already understand and you'll get a polished summary of yesterday's consensus. The real risk is mistaking a plausible answer for a discovery.

What happens when you leave the office

I've spent almost two decades taking corporate teams out of the office to learn from people in other industries.

On a program about trust and safety in Los Angeles, I took Coca-Cola's Scientific and Regulatory Affairs team to a cannabis store and lounge. Not the most obvious field trip for a room of regulatory scientists. Back home in Georgia there wasn't a single licensed dispensary at the time, and people were hesitant at first.

Then the owner, a former executive, and her staff answered questions in real depth, in an airy, beautifully designed space. By the end, the team was far more comfortable.

What struck everyone was how the setting helped establish credibility. The staff knew their subject, and the care put into the space and the way the products were presented reinforced it. That hit home for a team responsible for the science behind what people drink, as consumers scrutinize ingredients more than ever. How do you earn that kind of confidence as a brand?

With Clorox's Global Insights team in San Francisco, we explored food safety, starting from the assumption that people cared about hygiene but didn't always act like it. A food truck crew said customers never raised food safety; they trusted the truck, probably thanks to its artisan feel. A nurse at a hospital infection clinic told us homes don't need to be sterile, since even hospitals can't completely manage that. Common sense and good hand washing go a long way. A food-safety author pointed out that foodborne illness is badly underreported, so many people may never link getting sick to their own kitchen.

At our download session, one team member summed up the learnings with a question about foodborne illness: "Is there really fear?!" The people we met were pragmatic about hygiene. That didn't prove nobody was worried, but for a company that sells protection from germs, it was a jolt. It raised a different question: might visible reassurance that something is clean matter more than another warning about what could go wrong?

Neither the Coca-Cola nor the Clorox program produced a representative survey result, and neither was meant to. What they did was change the question. Coca-Cola's team left asking how to make people feel comfortable, not just informed. Clorox's team left asking what "clean enough" means to people who feel fairly relaxed about germs.

A field toolkit: discover first, then measure

When the brief involves something new, start in the field:

Start with a tension, not a topic. "Food safety" is a topic. "People say they care about hygiene, then cut corners at home" is something a team can go out and investigate.

Go where your category isn't, and talk to the people who watch customers all day. Staff, nurses and food truck crews observe behavior customers can't describe about themselves.

Bring the people who own the budget, not just the insights team. An insight handed over in a deck is easy to nod along to. When the people who make the call see it happen firsthand, it sticks.

Listen for the pauses. Every time someone stumbles into "I don't know," circle it. It may point to the question most worth testing.

Then test it. Once you've found something worth investigating, use surveys, quantitative research and synthetic respondents to test it and size the opportunity.

The answer we didn't know to ask for

Could AI explain the appeal of nap pods today? Sure, now that people are already doing it. But before the salon opened, who would have thought to ask whether people would pay to nap around the corner?

AI can help us make sense of what people have already done. Fieldwork can reveal the first signs of what they might do next. And when someone tells you "I don't know," don't rush to fill the silence. It may be the point at which the real research begins.