Among advertisers and marketers, there's a rush underway to use synthetic data. Research Live calls it “the next new kid on the block for the research sector.” Defined by AWS as “non-human-created data that mimics real-world data,” it comes in various forms. Perhaps the most promising form is the creation of synthetic personas.

Also known as “digital twins,” these personas allow marketers to test messages, ads and even entire campaigns on imaginary people whose attributes and interests are based on a large collection of data about real people. (To avoid confusion, marketers should know that the term “digital twins” is also used in another context, as a reference to simulated real-life objects.)

Marketers and advertisers in a wide range of industries are interested in synthetic data but face big pitfalls. In the race to build digital personas, companies often leave out their most important data: qualitative findings. So, they end up with inadequate, two-dimensional representations of the people they’re trying to reach.

Read: How digital twins can help brands learn about real consumers

Qualitative research is key to 3D

For years, brands have conducted surveys and collected feedback in which people submit open-ended answers. But these brands have generally lacked the personnel, time or tools to piece through and learn lessons from all those answers. One chief insights officer said his company (on the Fortune 50) has huge sums of qualitative data “sitting in a trash can.”

This problem affects B2C and B2B organizations alike. After meeting someone at a dinner party, the CEO of a household name brand felt the person seemed like her company’s ideal buyer. She got this sense not just from the person’s job title, but from how this person described approaching work and even life.

The next day, she asked her marketing team about the attributes of the organization’s typical buyer. Their responses revolved around facts and figures—the kind collected in quantitative surveys involving numerical or multiple-choice answers, as well as transactional data. The marketing team described a typical buyer’s role, the size of the company, the size of the staff and the industries they work in.

She asked for a sense of their thought processes, attitudes and beliefs. She wanted to know what they were like as people. The team said they didn’t have this information. Actually, they did have access to some of it; they just didn’t know it. It was sitting dormant in reams of qualitative and quantitative survey data. 

AI-powered tools can now piece through all of the responses and feedback that people have submitted over the years, across all platforms. When properly configured, platforms drawing on the power of large language models (LLMs) can “understand” all those answers, analyze them and contextualize them with other data sets such as quantitative surveys, social media posts, online reviews and sales histories.

Nearly instantly, these tools can draw insights about how people think and feel, what their pain points are, what their hopes are for their careers and their lives and more. Marketers can ask these tools to dig up even more insights by simply typing in a question.

A virtuous cycle

Companies try to rush ahead with digital personas without first having their AI tools ingest all that precious qualitative data. When that happens, these “twins” often fail to act like their real-world siblings in small but important ways. Their responses to innovations or campaigns can be so generic as to be useless or just plain wrong. Following the advice of a faulty digital twin can result in wasted resources, brand risk and damaged careers.

But when they create digital personas that are enriched with all that qualitative data, marketers garner responses that are much more likely to reflect real-world results. They also find that their AI tools can suggest messaging for campaigns that draw directly on the responses of audience segments that matter the most.

The system is cyclical. As companies tap into their qualitative data as part of a holistic process, they discover more of what they don’t know about their buyers. AI tools can then recommend new survey research to help gather more information and build the kinds of survey questions that people are most likely to respond to with meaningful answers, rather than ignore. As the responses come in, they can use the findings to flesh out digital twins even more.

For companies that chart this course, the potential is unlimited. The better they understand the real people behind the synthetic data, the more customers they reach—and the more they can learn about them. It’s a virtuous cycle, positioning brands for rapid growth in this new era.