As a marketer in the B2B space, getting your hands on reliable customer research data can feel like trying to spot a kiwi in the wild – challenging, time-consuming, and often elusive. Recently I’ve been hearing more and more about synthetic data, and started wondering: could this be the answer to our problems? I decided it was time to dive in and suss out what it’s all about.
Now, I’ll admit, I was initially a bit sceptical. How can synthetic data possibly compare to real-life human data? And is this just another AI trend that we marketers love to hail as the cure-all for our woes? But then, I recalled Mark Ritson mentioning it during a talk at a Marketing Association event last year, and I figured there might be something to this…
If you’re not yet familiar with synthetic data, don’t worry – you’re not alone. In simple terms, synthetic data refers to artificially generated information that mimics the characteristics, patterns, and statistical properties of real customer data. It’s like the stunt double in a movie, stepping in to perform all the tricky manoeuvres while the real star takes a break. But unlike most stunt doubles, synthetic data can actually be better than the real thing – safer, more versatile, and surprisingly accurate.
This AI-powered data is created using advanced algorithms and machine learning models trained on actual customer datasets. The result? Data that replicates realistic customer behaviours and attributes, ready for your analysis.
Synthetic customer research data has shown some pretty impressive results. Various sources report a 90-95% correlation between synthetic data and actual survey results – which is remarkably high. While synthetic data tends to perform well for practical, factual questions, it may be less reliable when capturing complex emotional or subjective customer insights. But for many B2B scenarios, where objective data is king, synthetic data is more than up to the task.
Now, you might be wondering, “This sounds interesting, but what’s in it for me?” Synthetic data could just be your ticket to more effective, efficient, and secure customer research, offering several key benefits:
1. Cost effective and time efficient – We all know that in B2B customer research data can take a long time to gather. Getting access to your customers, convincing them to respond to a survey or participate in in-depth qualitative research, and then analysing the data can often take months. Synthetic data, on the other hand, can be produced in a fraction of the time (typically in days rather than months) and for a fraction of the cost.
2. Scalability and flexibility - Synthetic data allows you to generate large, diverse datasets that may be difficult or impossible to obtain through conventional methods. Have you ever wished you could go back and ask a different question or probe a bit deeper? Synthetic data is more flexible. It’s not a one-and-done situation - you can keep asking questions and delving deeper to get the insights that you need.
3. Enhanced privacy and compliance - With increasing privacy regulations, synthetic data offers a way to work with realistic data without compromising individual privacy or violating data protection laws. It’s a win-win: you get the insights you need without the headaches of data compliance.
While synthetic data is powerful, many experts suggest using it as a complement to, rather than a replacement for, traditional research methods. It can be particularly useful for boosting samples sizes or reaching niche or hard-to-access audiences. A balanced approach using both synthetic and real-world data is likely to provide the best results.
The use of synthetic data for customer research is relatively new, but there are already some notable players making waves:
Evidenza
Founded by Peter Wienburg and Jon Lombardo (formerly of the LinkedIn B2B Institute) and Brian Watroba (from Facebook), Evidenza is pioneering the creation of synthetic customers – or “impersonas” as they call them - that closely match real-world profiles. Their platform has shown impressive correlations with traditional survey results.
Their websites states “With synthetic research, even the hardest to reach customers are now available on demand. Surveying AI-generated audiences make it 1000X easier to generate a winning go-to-market plan.
Here’s a great podcast where Jon Evans, in his Uncensored CMO podcast, speaks with Jon and Peter about B2B brand building in the era of AI, and where they talk about their new venture, Evidenza.
And this is what Mark Ritson had to say in his Marketing Week column “It’s not just that synthetic data is cheaper and faster and potentially more accurate. It’s that it opens up the ability to do mind-bending things that simply are not possible with traditional organic subjects.” Note: Mark Ritson is an advisor and minor shareholder in Evidenza.
Yabble
Yabble offers a platform called Virtual Audiences, which uses AI to generate synthetic data from multiple sources. Their approach focuses on creating “augmented data” that combines knowledge from diverse sources, potentially offering marketers a more comprehensive view of their target audience.
Their websites states “Built with a combination of custom algorithms, 50,000+ hours of training and world-class Large Language Models – Yabble is fully secure, and is the leading solution for every stage of research.”
PersonaPanels
PersonaPanels uses AI-powered synthetic respondents to provide consumer insights, creating machine-learning representations of key market segments by exploring the internet and uncovering trends.
Their (not particularly impressive) website states “Unique in their design, these AI-driven Synthetic Respondents are more than digital constructs; they are designed based on extensive human data, mirroring real-world market segments with impressive accuracy.”
Synthetic Users
The Synthetic Users platform allows testing ideas or products using AI participants, offering a conversational interface for creating surveys and building AI models of customers to simulate personas. Their website states that they offer “User research. Without the users, recruitment, scheduling, synthesising, or cost.”
So, all up, I’d say I’m now convinced that synthetic data for customer research is hugely appealing, especially for us B2B marketers. Using synthetic data, we can gain valuable insights that have typically been too time consuming and costly to access. From my perspective, I’m super excited that there’s a potentially viable new option out there.
I guess my big question now is whether there is anything like this based on New Zealand audience data? Most providers seem to be US-based; however, I note that Yabble does have an office in Auckland and some local New Zealand customers, so maybe there is hope…