AI-built focus groups can fill the gaps in your audience research, as long as you know which questions to ask them.
I expected to be delivering unwelcome news.
We had run synthetic research for one of our clients, and the report flagged a gap in their pricing structure: one of their products had no middle-tier price. Instead of pushing back, the client told us they had already spotted the same issue in their own customer data and were creating a product to fill it.
Think about what that means. Working only from a model of the client’s audience, synthetic research reached the same conclusion that their own data had. For an organization that hadn’t found the gap yet, the same report would have pointed straight to a product worth building.
That’s why I think synthetic research deserves a place in your planning. It’s also why I think it’s easy to misuse.
Every destination marketer knows a strategy is only as strong as the audience insight behind it. Research tells you who’s most likely to visit and what would get them to book. Yet most destination marketing plans rest on less research than they need.
Primary research, such as quantitative surveys and qualitative focus groups, takes budget and time that many destination marketing organizations (DMOs) simply don’t have. So teams pull whatever secondary research is available and start planning. Any question the research leaves unanswered becomes an assumption in the plan.
Synthetic research offers another option. These AI platforms let you build focus groups of virtual people modeled on your target audiences and put questions to them directly. The approach is still in its infancy. But because it finds patterns and gaps much faster and more cheaply than primary research, I think it’s on its way to becoming a dominant research tool.
That speed and easy access are exactly what make it easy to misuse. Before synthetic research becomes a standard part of your planning, you need to know where its answers hold up and how to use it well today.
Use synthetic research to find what’s missing
Synthetic research does what AI does well. It recognizes patterns, and it finds correlations between pieces of information from unrelated sources. That makes it useful for spotting gaps in what you think you know about an audience and for testing whether a message or call to action is clear.
The pricing gap is a good example. It was a factual, structural question: does the product lineup leave a hole? The model didn’t need taste or judgment to answer it. It needed to see the pattern, and it did.
Keep creative judgment out of synthetic research
Where I’ve seen synthetic research fall apart is when the question is subjective. AI understands the information it was trained on. It doesn’t understand subjectivity, so strip opinion and creative judgment out of every question before you run it.
For example, don’t ask a panel of virtual people to decide which of two campaign ideas they like better. Plenty of great campaigns looked like a terrible idea on paper. Describe an ad where an older woman gets tackled into the mud during a pickup football game, and it sounds like a nonstarter. Cast Betty White in that role, and you have the Snickers ad that ran during the 2010 Super Bowl, an idea that worked brilliantly.
Look, human research has limits here, too. A statistically projectable panel of a hundred people is still only a hundred people, and most of them aren’t trained to judge creative. They can’t tell you why something should be blue instead of red, or why one photo of a trailhead at dawn will make someone book and another, nearly identical shot won’t.
The experts inside your organization and your agency can, and their judgment deserves weight no matter how much research you run. I’m not saying you should ignore research when you make creative calls. I’m saying no study, synthetic or human, should get the final word.
Instead, use synthetic panels for factual or structural questions, like:
- Are we missing a price point?
- Which call to action reads most clearly?
- What gaps exist in what we know about this audience?
- What patterns show up across the data we already have?
Layer synthetic research with the data you already have
Many DMOs skip a step in their research. They might run a few focus groups or gather secondary research, but they never layer that data to reveal what no single source can show on its own.
Synthetic research can help you close that gap. Bring together your CRM and booking data, the reviews travelers leave, and whatever primary and secondary research you have. Then use synthetic research to look for the gaps between them. The combined picture gives you research that can shape your strategic planning, not just support it.
At Off Madison Ave, that layered research is the first step of a client’s Behavior Blueprint. A Blueprint is our research-backed strategic direction, grounded in Behavioral Marketing, so everything we create starts with what moves your audience to act.
Layered research matters most at the next stage, when we build Behavioral Archetypes for each audience. Demographics and personas tell you who your audience is. An Archetype also has to explain what motivates those travelers, what stands in their way, and when they’re most likely to take action. No single data source answers all of that.
I’ve seen teams move straight into strategic planning with thin research or off-the-shelf personas. They usually end up with a stack of tactical channel plans and no long-term thinking holding them together.
Use synthetic research to test your plan’s assumptions
Synthetic research will keep improving. I’ll be honest with you, though: for now, it handles a narrower set of questions than the hype suggests. Its best use is answering the factual questions a plan would otherwise leave to assumption.
So, start with your current plan. Look for the assumptions that rest on a factual question you’ve never tested. Does your website answer the questions a first-time visitor is likely to have? Do the itineraries you promote leave out experiences your target audience usually looks for?
Those are good candidates for a synthetic panel. Layer what it reports with your own data, and leave the creative calls to the people trained to make them.
Remember that pricing gap. Our client found it in their customer data, and synthetic research found it from a model of their audience. I expected that report to be unwelcome news. It turned out to be proof of what synthetic research does well.
Use it to turn assumptions back into answers. Ask it to stand in for human judgment, and you’ll get confident answers to questions it was never equipped to judge.
Off Madison Ave builds Behavior Blueprints for destinations and hospitality brands. To see how Behavioral Marketing clears the path from wanting a trip to booking one, read The Behavioral Marketing Playbook: How to Remove Barriers and Prompt Action.