Four ideas that stood out at Quirk's New York

Rep Data Team · August 6, 2026 · 4 min read

Eight members of the Rep Data team attended Quirk's New York this year, giving us the opportunity to cover a broad range of sessions. Although the presentations covered different industries and methodologies, our notes kept returning to many of the same topics. Here are four observations from the sessions our team attended.

Research is becoming part of an ongoing decision process

We attended eBay's session, “How eBay uses AI to build the always-on decision engine,” where they described moving from a research-first mindset to a decision-first mindset. The business decision comes first, followed by the research needed to support it. It was interesting to hear about how AI is used to coordinate parts of the research process so work aligns with business decision windows, with some projects moving from four- to six-week timelines toward 24- to 48-hour decision support where appropriate. Throughout the presentation, research was described as part of the decision process rather than a standalone project.

The Mondelēz session we caught at the show described a similar approach through a different lens. Their research on premium indulgence began with foundational learning, then continued through creative development, validation, in-store research and innovation planning. The presenters described it as a "living research foundation" because each phase informed the next. We also attended Bayer's presentation, where concept testing extended into consumer bidding, in-home product experience, a second bidding exercise after product use and business-case validation. Both these sessions treated research as something that continues supporting decisions rather than ending with a final report. That perspective resonated with our team because many of the same questions came up in Rep Data's own focus building confidence throughout the research process.

Research is getting closer to consumer behavior

Another thing from the Bayer presentation we noted was how they followed consumers beyond stated purchase intent. Participants evaluated concepts, bid with their own money, used the product at home and completed a second bidding exercise. They described how ingredients that initially raised concerns among participants became stronger drivers of value after product use, influencing positioning, packaging and pricing.

Symrise’s session examined another part of the purchase journey where consumers buy fragrances online without trying them first. The researchers described studying how opinions form through digital communities, organic social conversations and the language people use to describe products. The session also explored how community consensus influences purchasing decisions.

These presentation topics reinforced the value of combining stated responses with behavioral and contextual evidence. That broader view of research also informed Rep Data's discussions throughout the conference about building confidence in business decisions.

Building confidence requires multiple sources of evidence

Rep Data's fireside chat, “From Access to Trust,” traced two decades of change in online sample, from the rise of proprietary panels and programmatic marketplaces to today's AI-assisted participation. The discussion focused on how the industry's priorities have shifted alongside those changes. As access to respondents, research platforms and AI-generated analysis has expanded, greater attention has turned to the quality of the underlying data and the confidence organizations place in the findings they use to make decisions.

That theme carried into the panel session we moderated, “Building Confidence in Modern Research.” Panelists discussed selecting methods based on the business question, the decision being made and the level of confidence required, rather than relying on a single methodology. And in their session, Symrise described combining social content, community analysis and language analysis to understand how opinions form before purchase. 

The theme of all these described organizations selecting different combinations of evidence based on the decision they were trying to make. The emphasis was on matching the research approach to the business question rather than relying on a single source of information.

AI is changing the role of insights teams

AI appeared in nearly every presentation our team attended, but many of the discussions focused on how insights teams organize their work. During the eBay session, speakers distinguished between using AI to improve existing workflows and using it to expand the role of the insights function. Their examples centered on creative development, customer understanding and business planning, with insights teams contributing earlier in the decision process. One line from the session summed up that approach: "Don't wait on the agenda. Create it."

Rep Data's panel explored similar questions through the experiences of research leaders adapting workflows, team structures and quality practices as AI becomes part of day-to-day research. Our panelists stayed focused on how researchers apply judgment, evaluate evidence and maintain quality as AI becomes part of the workflow. This reflected many of the same questions our team heard throughout Quirk's New York and reinforced why data quality, research design and confidence continue to be central to the decisions organizations make.