Synthetic data: taking business over the rainbow

Leila Callaghan at Sullivan & Stanley describes how the wonderful wizard of synthetic data can take businesses over the rainbow

 

With Wicked hitting the silver screen recently, it’s hard to ignore the parallels between the film’s Wizard of Oz character and business leaders. Just as the Wizard used hidden mechanisms to appear all-powerful, businesses have long used data to create magical experiences for customers. 

 

Third-party data was the glamorous assistant of choice from the late ‘90s, offering the chance to track users’ movements and gain insight into their preferences. By the 2010s, privacy regulations such as GDPR and the California Consumer Privacy Act (CCPA), as well as technological shifts like Apple’s iOS 14 update and Google’s proposed cookie phase-out, saw first-party data take centre stage. 

 

Fast forward to today and synthetic data is the transformative force stealing the spotlight. This AI-generated data simulates human behaviour without relying on real-world responses - powering both data-driven decision-making and watertight privacy compliance. And while this may sound like science fiction, it’s already demonstrating uncanny accuracy. Marketing expert Mark Ritson notes that AI customers can deliver a 95% match to real survey results. 

 

But for all the hype and anticipation, a word of caution. Synthetic data, at this point in time, isn’t a standalone magical solution. Remember, the Wizard needed a human to pull the strings behind the curtain - the same goes for synthetic data. It’s powerless on its own. In order to tap into its true potential and drive meaningful outcomes, it needs a human touch.

 

Peeking under the hood

It’s not difficult to see why synthetic data is already being heralded as the next big innovation in customer experience (CX). After all, what business wouldn’t want an endless stream of rich customer insights?

 

At its most basic, synthetic data provides a comprehensive understanding of your audience. This is the foundation for reshaping operating models and injecting new life into service designs - for B2B and B2C organisations alike. It empowers businesses to experiment with new experiences and stress-test them in a risk-free environment. This doesn’t just help ensure GDPR compliance; it also reduces costs while improving time to market. 

 

The “fail fast, fail safe” approach enabled by synthetic data allows businesses to uncover areas of friction before they escalate into real-world backlash. Just as valuable is its ability to shine a light on promising developments. History is awash with cautious businesses that may have thrived if armed with synthetic data’s insights. 

 

Imagine if Kodak had been able to test the market’s appetite for digital cameras instead of abandoning the project early. Would Blockbuster still be relevant if it had simulated how personalised streaming services resonated with audiences? Could modelling ecommerce trends to anticipate customer needs have saved Toys R Us from administration?

 

Synthetic data doesn’t just thrive from a strategic standpoint. The tool excels as a diagnostic asset. Using it to conduct root-cause analysis - be it evaluating call centre response times or testing chatbot accuracy - offers businesses the chance to identify bottlenecks while improving consumer trust. This technique can also help foster inclusivity by simulating the accessibility of a website or service amongst underrepresented groups, such as disabled or elderly users. 

 

All that glisters is not gold

It’s easy to become swept up in the abundance of promise that synthetic data has already shown. But beneath its shiny surface, there are a few problems that require tuning-up before it can be labelled the full package. 

 

One of the most significant hurdles is synthetic data’s inability to accurately capture human emotions. Whether it’s trying to replicate consumers’ anger at a sub-par experience or the delight of a positive interaction - even the most advanced models can fail to mimic these nuances. This problem extends to loyalty-driven behaviours, which are often based on emotional connections.

 

Another pressing concern is the possibility of bias in the underlying models. Synthetic data is only as reliable as the information it’s based on. Granular auditing and consistently monitoring the output is paramount for businesses that want to avoid skewed results. 

 

The reality is that at this moment in time, synthetic data isn’t ready to shoulder the entire decision-making burden for businesses. And nor should it. Strategies should always be grounded in genuine customer behaviour, which is why human oversight needs to be layered in. 

 

Going back to my original point: it doesn’t take a trip to the Wizard to drive CX innovation. It simply requires acknowledging that real-world insights should be augmented, not replaced. And this starts by pairing qualitative research that understands complex human emotions with the speed and privacy compliance of synthetic data. 

 

---

 

Leila Callaghan is Consulting Client Director at Sullivan & Stanley

 

Main image courtesy of iStockPhoto.com and Lindsey Martin Webb