Synthetic data: the AI revolution you didn’t see coming

Picture yourself in an intense strategy meeting. The dialogue gets heated as each person vouches for the credibility of his or her market research data, until someone drops a bombshell: What if all this data is artificial?

Heads turn, eyebrows are raised, and suddenly you have everyone’s attention. Welcome to the new frontier: synthetic data.

What is synthetic data?

For those new to the concept, synthetic data is artificially created to replicate real-world data. Imagine it as a clone of the original — like Dolly the sheep — designed to look, feel and function like the real thing, but created through artificial means. 

Think of fake grass in your backyard. It looks like real grass but doesn’t need watering or mowing. Similarly, synthetic data looks and behaves like real data, but it’s created by computers and doesn’t require the same effort to collect. 

For marketers, this means access to vast amounts of data without privacy issues, data collection restrictions, or the high cost of commissioning your own research.

The synthetic data surge

Here’s a statistic to get your gears turning: Gartner predicts that by the end of 2024 60% of the data used for AI and analytics projects will be synthetic. Real data is about to be overshadowed by its artificial counterpart, leading to it rapidly becoming the digital revolution’s newest star.

The strategy overhaul

Experts suggest synthetic data can be superior to real data. It allows marketers to generate high-quality data sets tailored to their specific needs, opening up endless possibilities for testing and iteration. It’s akin to having a tireless team of AI-driven market researchers. This data can be instrumental in developing a host of strategies for customer engagement, perception mapping, media-mix modelling, and even economic analysis.

Real-time insights and decisions

One of the most exhilarating aspects of synthetic data is its potential to revolutionise real-time marketing. Imagine AI systems that continuously churn through data, optimising targeting, positioning, pricing, and media mixing in real time. It’s like having Doctor Strange’s ability to view 14-million futures and pick the optimal path — only with less spandex and more spreadsheets.

From data to action, seamlessly

Synthetic data doesn’t just end with strategy — it powers flawless execution. 

Picture this: your AI system not only devises your marketing strategy, but also executes it across digital platforms, tweaking and refining it in real time based on performance. The era of “set it and forget it” campaigns is over. The new dawn is “set it and watch it evolve”.

Privacy, bias and cost: the synthetic data trifecta

Synthetic data brings big benefits in privacy protection and bias reduction. By generating data free of personally identifiable information, companies can comply with regulations such as the General Data Protection Regulation and the Health Insurance Portability and Accountability Act, reducing the risk of data breaches. Additionally, synthetic data can create balanced data sets, improving the representation of minority groups and addressing crucial issues in AI model training — not to mention that it’s often more cost-effective to generate synthetic data than to collect and annotate large volumes of real data.

Criticisms to be aware of

Despite its advantages, synthetic data is not without its drawbacks:

• Privacy risks: Poorly anonymised synthetic data can lead to linkage attacks and attribute disclosure, potentially leading to personal details being exposed.
• Fidelity issues: Synthetic data may not capture the full complexity of real-world data, which could result in subpar model performance in real scenarios. Existing biases in real data can also be replicated and amplified.
• Ethical and legal concerns: Misuse of synthetic data, such as creating deep fakes, and ambiguous regulatory compliance present ongoing challenges.
• Technical limitations: Producing high-quality synthetic data is challenging and demands big computational resources.

The democratisation of data

One of synthetic data’s most transformative benefits is its democratising effect. Small and medium-sized enterprises, which often lack the resources for extensive market research, can now access high-quality data at a fraction of the cost. This levels the playing field, allowing David-sized companies to compete with Goliaths in their industry​.

Future forward: what’s next?

Looking ahead, it’s clear synthetic data is more than a fleeting trend. The synthetic data market is projected to grow substantially, potentially reaching more than $1bn (about R18bn) by 2027. It’s a catalyst for a new era of data-driven marketing. As technology advances, we can expect even more sophisticated applications, from personalised customer experiences to predictive analytics that anticipate market shifts before they happen.

Synthetic data is reshaping the landscape of marketing strategy and execution, proving itself an indispensable tool in the marketer’s arsenal. 

• About the author: Alexander Brand is creative editor at Rogerwilco.

This article was sponsored by Rogerwilco.