101 on synthetic data and why it could be big in advertising

As privacy concerns grow and the cookie-less future approaches, marketers are turning to smarter, safer ways to power their campaigns. One such solution gaining traction is synthetic data, a technology Forrester has listed among its Top 10 emerging technologies for 2025.

Here's a breakdown of what synthetic data is: Synthetic data is increasingly seen as a strategic asset in advertising, with agencies and martech players in Asia Pacific already exploring its potential. According to Forrester, 42% of surveyed firms say they’re using it as a privacy-preserving technology, but the full extent of its capabilities is only beginning to surface.

Don't miss: Agentic AI for dummies: 101 on how marketers can leverage on the trend

Defined as artificially generated data that mimics real-world datasets without any direct link to actual individuals, synthetic data can come in many forms, including structured, transactional, image, or audio data. It is especially valuable when real data is limited, restricted, or unsuitable for use.

By enabling organisations to bypass data privacy barriers, synthetic data opens new pathways for testing, optimisation, and innovation. Forrester notes its growing application in marketing and customer experience (CX) analytics, as well as for developing and testing code in safer, controlled environments. Ultimately, it’s paving the way for more responsible and scalable AI development.

Unlike first or third-party data, which is collected from actual users, synthetic data is entirely machine-generated. Despite not being “real,” its applications in media and advertising are very real, and growing fast.

More than a privacy patch

In conversation with MARKETING-INTERACTIVE, three digital and media agency professionals shared what the use of synthetic data looks like in the field of marketing and advertising. 

“Synthetic data is data that’s generated artificially to look and behave like real-world marketing data,” said Inna Weiner, VP of product at AppsFlyer. “Think user sessions, conversions, campaign performance, but without tying back to actual individuals." She added:

In advertising, it’s especially useful for safely training models, testing new ideas, and simulating scenarios when you can’t use real data due to privacy restrictions.

AppsFlyer has helped brands such as Coca-Cola, Nike and Pinterest build better products and preserve customer privacy with its advanced measurement, data analytics, data clean room and privacy-preserving technologies.

For marketers wary of regulatory risks, this makes synthetic data particularly compelling. “Privacy laws are getting stricter, and user-level data is harder to access. Synthetic data helps us simulate that missing behavioral layer while staying compliant,” Weiner said. "Anonymised data is fragile, but synthetic data doesn’t trace back to anyone. It’s made from scratch.”

Laurent Thevenet, head of creative technology, Asia Pacific at Publicis Groupe, shared a similar view:

Synthetic data gives us the ability to leverage ‘the known’, and to simulate ‘the unknown’.

"It guides our decisions in more informed and impactful ways. Our teams in Japan have been piloting this, and now we’re doing it at scale regionally," she said. 

Publicis has already begun using synthetic data for campaign development and creative testing. “We’re testing creative with a full virtual focus group comprised of data-backed personas to assess audience fit at scale,” added Thevenet. “Synthetic data augments our current data sets to imagine new futures and possibilities.”

For Kuhan Kumar, CEO of Malaysian digital agency Digital Symphony, synthetic data is more than a workaround, it’s a paradigm shift. “Synthetic data is an engineered signal, not scraped, tracked, or bought,” he said.

It’s how we break the tradeoff between innovation and privacy. You get full modeling power, and zero privacy baggage.

Kumar's team is already using synthetic data in real-world applications, from media planning to AI model training. “We use it to simulate audience behavior, conversion likelihood, and media efficiency curves. Especially in markets or segments where real data is sparse or noisy,” he said. “It’s become a core part of our AI prototyping loop.”

Endlessly adaptable

Across the board, the experts agree that synthetic data’s real promise lies in its flexibility. As Weiner pointed out, “Synthetic data isn’t static. You need to regenerate it regularly to keep up with seasonality and shifts in user behavior. Think of it as a living tool in your AI toolkit, not a one-off dump."

There’s also growing recognition of its role in inclusivity. “Synthetic data lets us balance datasets, not just reflect historical bias,” said Kumar. "If we’re serious about inclusive advertising, we have to move from observation to simulation."

Still, the experts advise caution and clarity. “Without good input, you won’t have good output,” said Thevenet. “The issue isn’t synthetic data itself. It’s the approach to creating it.”

Weiner echoed this sentiment, adding, “You don’t want synthetic data to be too perfect. The goal is also to explore what-ifs, edge cases, even outliers.”

Their advice to marketers? Start small, but start now. “Pick a use case with clear constraints. Message testing, media simulation, or audience expansion. Treat it as a competitive edge, not just a privacy patch," Kumar said, adding that:

The agencies that master synthetic data early will lead the next decade of advertising.

Related articles: 
Agentic AI and the CX reset: The tech changing how brands serve and retain customers
Study: 81% of APAC shoppers want AI-powered shopping tools
InfoSum CEO Lauren Wetzel on privacy, performance and the myth of data clean rooms