‘We no longer need to speak directly to people to understand them.’ That was the motion tabled at a debate organised by MRS in partnership with the Debating Group, which took place on 27th January at Portcullis House, Westminster.

Chaired by MRS president, Baroness Deborah Mattinson, supporting the motion was Ben Warner, co-founder of Electric Twin, and Kyri Bagdades, regional vice-president for client partnerships (Emea), NewtonX.

Arguing in opposition to the motion was Sir John Curtice, professor of practice – politics, University of Strathclyde and Bridget Williams, head of public communications unit – public affairs, Ipsos UK.

A pre-debate show of hands from the audience revealed a huge majority against the motion.

As a data scientist and advisor to the Conservative government, Warner worked on the 2016 Vote Leave campaign and was “in the room when decisions were made” during the Covid-19 pandemic.

Warner opened with a confession that he was a physicist, not a market researcher. “I spent a decade in quantum mechanics before I fell in love with the hardest, much harder problem – trying to understand, predict, and influence human behaviour,” he said. Electric Twin, the business he co-founded, creates “science based synthetic populations” to deliver instant consumer insights.

Covid-19 would be referenced many times during the debate by both sides of the debate. For Warner, a key feature of his argument in favour of the motion is the slowness – as well as prohibitive cost – of traditional research projects which can take “six weeks and £50,000” to execute, when decisions need to be made in the moment by a wide range of stakeholders. Synthetic solutions are fast and available to everyone, he said.

For Warner, supporting the motion means evolving a future for the research sector; existing ‘traditional’ methodologies are no longer fit for purpose: “Our tools don't actually work. They're not actually providing us with that insight to actually understand people… a lot of these arguments are based on the way that we work today. What we're trying to do is to iterate, to always evolve, to always improve what we're doing at every stage of when we're trying to interact with the world.”

Also in support of the motion, Bagdades said: “We will still need human inputs, no doubt – but not all of us, not all of the time, perhaps not most of the time. We need to focus on what matters to most people. And most people can't speak directly to these people at scale. The future of understanding lies in the synthesis and interpretation of human emotion… it is our duty to embrace this future and capitalise on its benefits for the betterment of our industry.”

Based as it is on historic data, the opposing argument also used the pandemic as evidence – this time of synthetic data’s inability to deliver foresight, or at least keep up with pace of changing behaviours. As James Endersby, MRS chair and chief executive at Opinium, said, models that are never refreshed slowly drift away from reality.

Williams shared this concern, saying: “Yesterday’s context is really out of date ever so quickly… Covid changed the context and it resulted in people very clearly changing their attitudes.”

As a counterpoint, a member of the audience from a marketing agency who supported the motion said that AI helped them to simulate multiple real-world contexts: “With synthetic populations, we can construct a very vast range of environments that people can consume in.”

The audience frequently heard these concepts that are central to the value of research – ‘context’, ‘speed’, ‘accuracy’ – cited by both parties to support their opposing positions.

The debate was entirely civilised and good natured, with Curtice only once protesting when the opposing camp appeared to have sequestered the concept of ‘modelling’. Underlining the sector’s long-standing credentials as being data and technology adept, Curtice reacted: “What do you think those of us who've spent far too many years trying to understand social attitudes have been doing?... modelling the data is exactly what we do.”

Over-reliance on modelling was, however, seen as a weakness in the synthetic approach, with participants often revealing insights in real conversations that a model could easily pass over with its tendency to produce generic and aggregate findings. Curtice said: “We are interested at the collective level, but we are also interested in what individuals think and what individuals do.”

While the ‘say-do’ gap was cited by Warner as a critical weakness in self-reported data (“People do not purchase what they say they will”), there was a counter argument from the opposition that this inconsistency was in itself part of the nuance of understanding real human behaviour and shouldn’t be ironed out. “Understanding people means loving the contradiction… the pauses and people saying, ‘I’m not sure’,” said Williams.

The most impassioned moment of the night came when discussing how the act of listening to vulnerable groups is a core value that the sector should uphold. One audience member asked: how would those groups – for example people suffering from bereavement – feel if they found out they weren’t being talked to, but instead were being simulated? They went on to say: “Children are people, too… they have a fundamental human right to have their voice heard.” Article 12 of the United Nations Convention on the Rights of the Child mandates that children have the right to express views on matters affecting them and to have those views taken seriously.

Despite the adversarial nature of any debate, it was clear that the market researchers in the room were comfortable embracing AI in their hybrid research methodologies and saw it as a valuable tool when working within clear guidelines. “We accept that yes, AI is to be used... but not that we can dispense with all human testimony,” said Curtice.

No surprise then, that the second show of hands at the end of the evening matched that at the start of debate, with the motion defeated by a clear majority.