Today we’re speaking with Milena Traikovich, a Demand Gen expert with extensive experience in analytics and performance optimization, who helps businesses craft campaigns that nurture high-quality leads. We’ll be diving into a fascinating shift in market research, where the long-standing wall between “qualitative” and “quantitative” methods is being torn down by new technologies. This conversation will explore how the traditional trade-off between the depth of human nuance and the power of large-scale data is becoming a thing of the past, and what that means for how we understand consumers. We’ll discuss how AI is enabling a new kind of reasoning at scale, what this means for the structure of insights teams, and what skills will be crucial for the market researcher of the future.
Marketing has often operated on the assumption of two paths: “nuance” for qualitative research and “numbers” for quantitative. Could you explain the practical limitations of this dichotomy and share a real-world example of how it may have hindered business insight in the past?
Absolutely. This split between “nuance” and “numbers” has been a foundational, and I would argue, limiting, aspect of marketing for decades. We created these proud identities—the “qual researcher” who understands human feeling and the “quant researcher” who measures markets—but in doing so, we often forgot that we’re studying the exact same things: people, products, and brands. The practical limitation is that you get an incomplete picture. For instance, a company might conduct a massive survey and find that 70% of customers are satisfied, a solid quantitative metric. At the same time, a separate qualitative team runs focus groups and uncovers a deep-seated frustration with the product’s user interface. Because these teams operate in silos, the “quant” data leads to a decision to maintain the status quo, while the “qual” insight, seen as anecdotal, is ignored. The business misses a critical opportunity to innovate and prevent future churn, all because the two halves of the story were never connected.
Imagine a researcher is tasked with analyzing one million tweets to find business insights. How does this scenario, which involves both messy human thoughts and massive scale, challenge the traditional separation of quant and qual methods? What new capabilities are needed to tackle this?
That’s a brilliant thought experiment because it’s a perfect modern koan for our field. When I ask, “Are you doing quant or qual research on those million tweets?” both answers feel equally right and wrong. The scenario fundamentally breaks our old models because it contains both the immense scale we associate with quantitative analysis—a million data points—and the profound depth of qualitative research, with all the messy, random, and nuanced human thoughts embedded in each tweet. Traditionally, a quantitative researcher would be lost without structured data to analyze, while a qualitative researcher would be completely overwhelmed trying to synthesize meaning from such a vast volume. This challenge reveals that our old division wasn’t a strategic choice; it was a technological limitation. The new capability we desperately need is the ability to perform synthesis at scale, to find the nuanced human story within the mountain of data.
You’ve pointed to abductive reasoning as a key shift, where we seek the most likely explanation for a surprising observation. Can you walk me through a step-by-step example of how a marketing team might use this “abductive loop” to better understand customer behavior?
The abductive loop is an incredibly powerful, intuitive process that good researchers have always used, but now we can apply it more systematically. Imagine a marketing team notices a surprising anomaly in their sales data skincare product that has been dormant for years suddenly sees a 30% sales spike in a specific region. That’s the first step: noticing the surprise. Instead of just noting the increase, they start the abductive loop. Their next step is to form the most likely explanation, a hypothesis. They might guess, “Perhaps a local celebrity mentioned it.” To test this, they gather more data, focusing on social media and local press in that region. If that yields nothing, they loop again with a new explanation: “Maybe a competitor’s product was discontinued.” They investigate, and so on. This iterative process of observing a surprise, forming the best explanation, and then seeking new data to confirm or deny it is the essence of abduction. It’s a dynamic search for truth, moving beyond just pattern-spotting to actively explaining the ‘why’ behind the data.
Generative AI is described as enabling “computational abduction,” or reasoning at scale. Beyond just processing data, how does this technology fundamentally change the synthesis part of research? Could you provide a specific example of how an LLM can derive nuanced insights that were previously unfeasible?
This is where the real revolution is happening. “Computational abduction” isn’t just about faster processing; it’s about delegating the act of reasoning itself. The synthesis part of research has always been the human bottleneck. A researcher could read maybe a hundred customer reviews to get the gist of a problem, but they couldn’t possibly read ten thousand. Generative AI, specifically a Large Language Model, can. For example, an LLM could analyze those million tweets we talked about and go far beyond simple sentiment analysis. It could identify that a specific slang term is being used ironically by a younger demographic when discussing your brand, indicating a potential perception problem that a keyword count would completely miss. It could connect a spike in mentions in one city with a local weather event and a specific product feature, revealing an unforeseen use case. This is true synthesis—deriving a complex, multi-layered explanation from vast, unstructured data. It’s an insight that was simply unfeasible to uncover when a human had to do all the reading and connecting of dots.
As the lines between depth and scale blur, how should insights and market research teams restructure their workflows and skill sets? What practical steps can a leader take to move their team away from a “qual” or “quant” identity and toward a more integrated approach?
The first and most crucial step for any leader is to actively dismantle the “qual” versus “quant” identities. Stop hiring for one or the other; start hiring “insights researchers” who are curious and method-agnostic. Practically, this means restructuring workflows around problems, not methods. Instead of having a survey team and a focus group team, create a “customer retention” team that is empowered to use any tool necessary—large-scale data mining, in-depth interviews, statistical analysis—to solve that problem. Leaders should invest in cross-training. Have your data scientists sit in on ethnographic interviews to feel the human context behind their numbers, and get your qualitative researchers comfortable with AI-powered synthesis tools to see the patterns in the noise. The goal is to build a team whose primary skill is not a specific technique, but a relentless curiosity and the flexibility to choose the right tool for the job, whatever it may be.
What is your forecast for the future of market research roles over the next five years?
My forecast is that the most valuable and sought-after market researchers will be those who are “bilingual”—fluent in the languages of both data and human experience. The pure specialists, the ones who cling to a “quant” or “qual” identity, will find their roles increasingly automated. The future belongs to the synthesizer, the storyteller, the strategist who can stand at the intersection of a massive dataset and a deep human truth and weave them together into a compelling business direction. The role will become less about executing a specific research methodology and more about asking brilliant questions, critically evaluating the outputs of AI models, and translating complex, hybrid insights into actionable strategy. The core of the job will shift from data collection to insight creation, making it a more strategic and indispensable function than ever before.
