The Respondent You Can’t Recruit: How Synthetic Simulations Are Changing B2B Research
By Verve
- article
- Artificial Intelligence
- Generative AI
- AI Personas
- Customer Panels
- Insight Communities
- Online Communities
- Qualitative Research
- Synthetic Data
- Data Analytics
- Trend Analytics
- B2B Research
Summarise with AI
B2B research has always been hard. The audiences are small, the individuals are busy and expensive to reach, and the incentives that work for consumer panels are simply not credible for senior professionals. I want to talk about why I think synthetic simulations are particularly well-suited to addressing these structural problems – not as a replacement for B2B research, but as a way to make it more practical and useful.
Learn more by watching or listening to Andrew on the Founders and Leaders Series podcast here:
Episode 9: Andrew Cooper, Founder & CEO, Verve
The Structural Problem With B2B Samples
Let me start with the basic challenge. In consumer research, the population is large and reasonably accessible. In B2B research, the relevant universe is often small – sometimes very small – and the people in it are not sitting around waiting to take surveys. A cloud computing specialist at a major enterprise organisation is unlikely to turn up for a twenty-dollar Amazon voucher. The idea that access panels can reliably deliver these audiences is, in my view, questionable. You may get someone who claims to be in that role, but how confident are you in that?
The consequence is that B2B research tends to be slow, expensive and methodologically conservative. Qualitative interviews are the mainstay because they are the most credible way to access senior B2B audiences. But they are resource-intensive, and the sample sizes they produce are inherently limited. You might conduct twenty in-depth interviews with independent financial advisers and feel you have good data. You do! But there are limits on how much you can build from it without augmentation.
Where Simulations Enter the Picture
The insight that changed my thinking on this was straightforward: the same data constraints that make B2B research expensive also make it a strong candidate for simulation. Qualitative interviews with senior professionals are, almost by definition, high-quality data. They are detailed, substantive and directly relevant to the client’s actual decision-making context. If those interviews can seed a simulation, then the simulation starts from a much stronger position than one built from consumer survey data.
We have worked with a large Swiss bank on simulations of high-net-worth North American clients, and separately with a financial advisory business on simulations of independent financial advisers. In both cases, the organisations already had qualitative data from their own client programmes. That data became the starting point for building simulations of both audiences. We then augmented with relevant curated sources to produce something more rounded and testable.
The question of whether the simulation is as good as talking to the real person is the wrong question. The right question is whether the simulation is better than the alternative, which is not doing the research at all, or doing it with a panel that probably cannot reliably recruit the audience you actually need.
A Different Use of Human Respondents
One of the practical shifts that simulations enable in B2B is a change in how you use the real respondents you do have access to. Instead of asking them to sit through a long survey or an exhaustive interview that tries to cover every research question, you can use the simulation to do the exploratory and tactical work, and bring the real respondent in at specific points where their direct input is most valuable.
You generate hypotheses using the simulation, test those hypotheses with real human beings, and then feed what you learn back into the simulation. It becomes a cycle rather than a linear process, and it is considerably more respectful of the respondent’s time. A cloud computing expert will engage in a focused thirty-minute conversation about a specific topic. They are much less willing to be the seventeenth person to complete a lengthy questionnaire.
Nothing is Wasted
There is another dimension to this that is underappreciated. When you conduct qualitative research, respondents tell you a great deal more than the brief requires. The client asked about pricing, so the report focused on it. But in the interviews, people talked about service delivery, trust, and competitive switching decisions. That information existed, was recorded, and then was effectively discarded because it was out of scope.
When that interview data goes into a simulation, none of it is wasted. The simulation is informed by the full content of what the respondent said, not just the portion included in the final report. Over time, as more research is conducted and fed back into the simulation, it becomes more rounded and more accurate. It retains institutional knowledge that human researchers, over time, forget or never formally document.
What This Means for Healthcare and Other Specialist Sectors
The same logic applies to healthcare research, which shares many of the structural difficulties of B2B. Specialist clinicians, senior NHS managers, and rare disease patients – all of these are hard to recruit in meaningful numbers, and the research conducted with them is expensive. Simulations based on high-quality qualitative data from these audiences can considerably extend the usefulness of that research.
I should be clear about what I am not claiming. Simulations built on thin or low-quality data will produce unreliable outputs in any sector. B2B does not automatically produce good simulations just because the research is qualitative. The quality of the seed data still determines the quality of the simulation. But where good qualitative data already exists – which it does in most mature B2B research programmes – simulation offers a practical way to do more with it.
The structural problem of B2B research has not gone away, but the tools available to address it have improved considerably. A well-built simulation, anchored in proprietary client data and validated against real respondent hold-out data, can make B2B insights faster, more accessible, and more cost-effective – without compromising the standards that make it credible for serious decision-making.
Learn more by watching or listening to Andrew on the Founders and Leaders Series podcast here:
