Key results
10x
more research, with the same headcount and budget
92%
accuracy match, against The Times’ 642,000 subscribers
3 teams
running synthetic research independently
Seconds
to customer feedback, instead of weeks
What they did
Conducted initial product testing to decide which features to prioritise on the roadmap
Ran creative testing to decide on a name for a new podcast launch
Evaluated which new partnership deals would help them reach new subscribers
The problem
The Times had an Insights team doing great work. The problem wasn’t quality: it was capacity.
Product teams needed answers in days, not weeks. Marketing wanted to test more ideas before committing budget. Commercial partnerships needed audience modelling that existing systems couldn’t deliver at speed. With 642,000 subscribers, the data existed but there were limits to how much the Insights team could ask of The Times’ reader panel.
The question was practical: how do you get better answers, faster, without compromising quality or exhausting the goodwill of your audience?
“We went from rationing research to running it on demand. The same team, ten times the output, and we could prove it matched our real audience”
Chris Courtney-Smith
Director of Data Operations, The Times
Three use cases
The Times integrated Electric Twin’s synthetic audience platform across three business functions: Product, Marketing and Commercial.
01
Product use case
Initial concept testing: better product roadmap prioritisation
Product teams used synthetic audiences to test product and design concepts before committing development time. Instead of waiting weeks for panel results or relying on internal opinion, teams could now identify potential audience reactions in seconds.
This tightened the feedback loop between insight and delivery.
One example revealed interesting differences in their audience: when asked about new ways to experience The Times, loyal readers expressed scepticism about AI, while prospective readers wanted summaries, AI explainers and video. That kind of segment-level contrast, invisible in aggregate research, directly informed product prioritisation and what to build next on their roadmap.
02
Marketing use case
Creative testing: naming a new business podcast
Marketing applied the platform to CRM optimisation, copy testing and campaign evaluation. The team had always struggled with limits on the number of questions they could ask. With synthetic audiences, the team could explore a far greater volume of ideas, without additional fieldwork cost or panel fatigue.
A standout example: naming The Times’ business podcast. The team tested four potential names with their synthetic audience. The winning name, “The Business”, scored well with the general reader panel but performed exceptionally with the specific demographic that the podcast was targeting.
Precision like that would have been difficult and expensive to achieve through traditional research.
03
Commercial use case
Audience discovery: checking whether a partner’s audience was the right fit
Before committing to a partnership, the Commercial team used Electric Twin to check: does this partner reach the same people we already reach? That evidence, which didn’t exist before, helped them back the right deals, so they were more efficient in allocating budget to reach new audiences and grow their subscriber base.
The impact
10x
increase in research volume, with the same Insights team headcount
92%
accuracy against real panel responses, validated using a 10% hold-out from 642,000 subscribers
3 teams
running synthetic research independently
Seconds
to real customer feedback
1 new name
for a podcast product, tested across different audience segments
Bigger reach
from audience discovery testing to find out which partnerships would result in new subscribers
Beyond the panel
product teams testing initial concepts that couldn’t justify traditional fieldwork
How we validate
We took data from The Times’ 642,000 subscribers and built a synthetic audience modelled from 90% of their real answers.
We then handed over the audience to The Times’ Insights team for accuracy testing. They ran their own hold-out tests and used the remaining 10% of their subscribers’ real answers to see how closely our synthetic personas’ answers matched what those subscribers had said.
That check returned 92% accuracy on NDAM, compared with the 94% benchmark for real-world surveys. This validation was critical in building trust within The Times, enabling further adoption of Electric Twin.
Looking ahead
The Times shared two key takeaways from using Electric Twin:
First, expanded research capability for every team means managing the divide between the sceptics and the supporters. It requires ongoing validation against real responses, visible human oversight and a consistent message that synthetic testing expands what researchers do rather than replaces it.
Second, how you frame value matters. Rather than measuring whether individual decisions improved, The Times measured research volume uplift and speed to delivery. That framing was honest, concrete and compelling to senior leadership.
The Times is working on expanding Electric Twin to more users across its organisation, so that all teams can test any idea with their subscribers for data-driven results.
At a glance
Industry
Media & publishing
Audience
642,000 subscribers
Location
UK
Teams
Commercial, Marketing, Product
Used for
Brand messaging and audience testing