What Don Draper got right about research and what he missed.

There's a scene in Mad Men I come back to a lot. A researcher presents what her focus group told her, and Don Draper isn't having it. His argument is that "a new idea is something they don't know yet", so it was never going to come up as an option in the room. Put the campaign on TV for a year, he says, hold the group again, and maybe it will. You can't predict how people will behave from how they've behaved before.

He's not entirely wrong. Creatives have a genuine instinct for what will land, built on experience, taste and craft that most consumers couldn't articulate if you asked them directly. If every ad needed consumer approval before it could exist, we'd have lost a lot of the work that actually changed how people felt about a brand. Sometimes the right move is to trust the person who's spent twenty years doing exactly this.

But instinct isn't always right, and knowing the difference is the hard part. Some briefs sit on category assumptions nobody's checked in years. Some ideas test brilliantly in the room and mean nothing to the audience they're meant to move. Some campaigns have a barrier buried in them that nobody on the creative side can see, because they're not the ones it's a barrier for.

That's what research is for. Not auditioning every idea for consumer approval. Finding the places where there's a real opportunity nobody's spotted yet, or a real barrier getting in the way, and leaving the rest alone.

Where Don is right, and where he isn't

He's right that asking people to predict their own behaviour doesn't work. Ask someone whether they'd respond to an idea they've never seen and you'll get a polite, unreliable answer. People are poor witnesses of their own behaviour, and they can't react to something that doesn't exist yet.

Where he's wrong is in thinking that's all research can do. Good research doesn't ask people to predict what they'll do next. It looks at what they actually do, brings them into the process while the idea is still being built, and puts the finished thing in front of them to see how they react. That's a very different job to a group of people being asked what they'd like.

I think about advertising research in five stages, because different projects need different things from it. Not every campaign needs all five. Plenty don't need any. The value is in using the right one at the right moment, to unlock an opportunity or clear a barrier that's actually there, not running every stage on principle because that's what the process says to do.

The five stages, with Don in the room

Stage one: audience understanding. Before anyone writes a brief, understand the audience properly: the category context, what motivates them, what gets in the way, how they actually make the decision in question. Don's right that this can't tell you what the new idea should be. It isn't meant to. It tells you where the idea needs to land, so the brief is built on something real and the creative team starts from a map instead of a blank page.

Stage two: ideation and co-creation. This is the direct answer to "a new idea is something they don't know yet". If people can't react to an idea that doesn't exist, bring them in while it's being made. Not to approve it, but to help build it. It earns its place when the creative team needs specifics that no amount of internal thinking will produce: the exact moment a behaviour happens, the exact words someone would actually use, the small real life detail that makes an idea land rather than just sound plausible. Clients often want to be in that room too, working alongside their own audience to get the thinking off to the right start.

I led the research on a project with Nestlé Purina and the University of Sheffield, developing a dog obesity health pack grounded in the COM-B model of behaviour change. The behavioural science said what should work in principle. It couldn't say what a dog owner's actual week looks like. Co-creation workshops with dog owners and vets filled that gap: a way to track weight week to week, a card for handling the moment a child slips the dog a biscuit under the table, a tag on the collar as a reminder every time the lead goes on. None of that specificity comes from a brief, however good the thinking behind it. It comes from sitting with the people who live the problem daily, and building the solution with them rather than for them. The pilot pack was tested with 78 dog owners, and the feedback shaped what went into the final version distributed through vets.

Stage three: creative refinement. Once there's something to react to, find out what's actually working, what's confusing, what's memorable, and what would make it stronger. By now the idea exists, so people aren't being asked to imagine it. They're reacting to the real thing. This is diagnosis, not a verdict. The point isn't to mark the work, it's to tell you precisely where to spend the next round of effort.

Stage four: pre-launch testing. Don's answer to the researcher was to run the campaign for a year and ask again. It's a fair test, and an expensive one. Pre-launch testing is the same idea, done before you've spent the budget: test the creative against what it actually needs to achieve, so you know which route, message or execution gives you the best chance of success without having to find out in market.

Stage five: live ad learning and tracking. And when you do run it for a year, this is how you find out properly. What's cutting through, what people are taking from it, what action it's actually driving, and what should change next time. Don's year on TV only teaches you something if someone's measuring what happened. A campaign's first burst in market is information, not a final verdict.

The real source of the tension

Watch the scene again and the argument isn't only about method. It's about territory. The researcher arrives with the findings, presents them, and the creative takes it as an intrusion into something that isn't hers to comment on.

That's worth taking seriously, because it's usually a process problem rather than a people problem. Research that turns up at the end and delivers a verdict will always feel like that. Research that's involved early, shaped with the creative team and designed to answer the questions they actually have, feels like part of the work.

The brief stays the creative team's. What research adds is detail nobody in a meeting room would invent alone. Used on the right project, it sharpens an idea. Used on the wrong one, it slows things down for no reason, which is exactly why it's not something to run by default.

A way to think about it

If you're stuck in a maze, and someone offers you a map out, you take it. Research is the map. It doesn't tell you which way is the most creative way to walk. It just stops you wasting effort on the wrong turns.

What you get from me

I work on my own, not through an agency chain. You brief me on the campaign, the brand or the question, and I build the right stages of the CLEAR framework that is my own design, not all five by default, just the ones that actually move the work forward.

That means audience understanding sharp enough to brief from, creative refinement that tells you exactly what to fix, pre-launch testing that gives you real confidence before you spend, and tracking that tells you what's actually happening once a campaign is live. Clear direction, not a pile of data to interpret yourself.

Minimal effort on your end. The point of research at this stage is to make the creative process faster and more confident, not slower and more defensive.

The question worth asking

Not "does this need research?"

But: is there a real opportunity here we haven't spotted, or a real barrier we can't see from the inside?

If the answer is no, do what Don would do. Trust the instinct and get on with it. If you're not sure, that uncertainty is exactly what research is for, as long as it's the right kind, at the right moment, asking people about what they do rather than what they think they'd do.

Developing a campaign and want research that sharpens the creative rather than slows it down? I'd be glad to talk it through. Get in touch.