Signal and Noise

Brian spotted an Insights Association poll asking one simple question: what should the insights industry’s biggest priority be in 2027? The four options were responsible use of AI, data quality and trust, proving the value of insights, and building next-gen talent. He sent it to Andrew, and they decided to just talk through all of it on the pod.

What followed was one of their more honest episodes in a while. They work through each priority on its own terms: what it actually means and whether the industry is actually equipped to tackle it. The responsible AI conversation turns into a sharp critique of conference culture, where nobody is getting on stage to talk about where AI breaks. The data quality section goes deep on why transparency is still the industry’s biggest structural gap. And the next-gen talent discussion raises a question nobody has a clean answer to: how do you build the next generation of researchers when the job itself is changing faster than any curriculum can track?

Key Takeaways:

• Why responsible AI use is meaningless without someone willing to get on stage and talk about where it fails

• Why data quality and trust are two different problems that keep getting bundled into one unanswerable priority

• What happens to the industry if a billion-dollar AI company acquires a major traditional research firm and applies VC growth expectations to it

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn (ROI Rocket)

• LinkedIn (Signal & Noise)

• YouTube

• ROI Rocket

After a small break, Brian and Andrew are back. Both came with independent lists, sat down, and went back and forth for an hour on something they have been doing together for over a decade: running brand tracking studies the right way.

Between them, they have overseen the in-field transition of more than ten large continuous tracking programs, launched dozens more, and spent years studying exactly how panel bias shows up in longitudinal data in ways that most researchers never catch. This episode is the playbook they wish existed when they started.

The conversation covers how to think about blending panels for a tracker, why your anchor panel is the most important decision you will make, how inherent panel bias gets built in through recruiting practices that panels will never volunteer to you, why over-engineering the blend is its own form of data corruption, and why NPS in a tracker is a trap waiting to spring.

If you run a tracker, manage a tracker, or are about to set one up, this one is worth a full listen.

Key Takeaways:

• Why your anchor panel is the single most important structural decision in a blended tracker

• How panels develop inherent brand bias through their own recruiting, and why they will never tell you about it

• Why NPS is dangerous in a tracker and how to use it without blowing up your trend line

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

No guest, no agenda, no notes. Brian and Andrew just riffed for an hour on one of the biggest questions in the industry right now: what does sampling look like in 5 years?

Brian is presenting on the future of sampling at the Future of Insights Summit in Athens, Georgia, and this episode is the unfiltered, longer version of that talk. The current model is broken and nobody serious is arguing otherwise. The real question is what replaces it, and who wins when it does.

They work through where today's panels fall short, what synthetic data can and cannot actually do right now, and why the first truly useful synthetic products will likely be built inside brands, not sold to them. P&G, Google, and Meta are sitting on behavioral data sets that dwarf anything a panel company owns. That has consequences.

The conversation also gets into why the brands that build proprietary synthetic tools will have a moat that nobody can buy, why psychological data will never fully be replaced, and why the next 5 years might be the most interesting and terrifying stretch in the history of this industry.

Key Takeaways:

• Why the current online panel model is broken and what the industry is not doing about it

• Why the first useful synthetic products will be built by brands using their own data, not by research suppliers

• Why human panels are not going away, and what role they play in a world increasingly run by synthetic models

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Andrew flies solo and sits down with Joey Maddox and Sebastian Babis from Verisoul, fresh off a $9M Series A and the most creative go-to-market run market research has seen in years.

Times Square billboards. A Reds game with hundreds of attendees. Branded cigarettes at a European conference. None of it is random. Verisoul mapped their top 200 prospect companies, identified the 10 to 15 buyers at each, plotted where they all live, and built activations around them. Strategic chaos.

The conversation covers how a team from banking and identity fraud disrupted a market that had been hearing the same messaging for years, why in-person trust-building is non-negotiable when your product is fraud prevention, and what it looks like to run go-to-market across wildly different verticals at once. Also, all three of them are newly married or about to be, and the wedding vendor pricing rant is worth the listen alone.

Key Takeaways:

• How Verisoul built a geographic database of prospects and designed its entire events strategy around it

• Why an outside-in approach to fraud detection is beating well-funded incumbents head-to-head on accuracy

• The marketing ideas that did not make the cut, including blowing up a building with a competitor's logo on it

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Connect with Joey:

• LinkedIn

Connect with Sebastian:

• LinkedIn
  

Brian and Andrew sit down with Karine Pepin of The Research Heads and Andrew Gordon of Prolific to break down one of the most important data quality studies the industry has produced. They analyzed 5,200 respondents across 10 platforms and the findings will surprise you.

AI agents are not in your data. Outside of MTurk, detection rates were below 0.5% across every platform type. And the things getting flagged on MTurk? Old-school bots, not LLMs. When they ran actual AI agents through the same survey, those agents outperformed real human respondents on nearly every quality measure.

The actual crisis is your human respondents and the platform you are buying them from.

Key Takeaways:

• Agent infiltration is not the crisis the industry thinks it is

• What is actually getting flagged as bots on MTurk, and what it really means

• Why the human quality gap between direct panels & marketplace platforms is the story nobody is telling

• Why the cheapest sample per click is often the most expensive data you will ever buy

Link to the study "AI Agent Prevalence and Data Quality Across Multiple Online Sample Providers"

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Connect with Karine Pepin:

• LinkedIn

Connect with Andrew Gordon:

• LinkedIn
  

Brian and Andrew sit down with Angela Chavez, Insights Manager at Dig Insights and UGA MMR graduate, ahead of her first-ever conference presentation at the Future of Insights Summit in Athens this August.

Angela's talk is called "Learning Research in a Rapidly Changing World," and this episode is a preview of exactly that. She walks through what it has actually been like to enter the industry as AI was exploding, going from hesitant early adopter to someone who opens ChatGPT as a reflex, all while growing from analyst to manager.

The conversation also gets into Dig Insights' tech-forward approach and what Angela's early career experience reveals for senior leaders trying to figure out how to onboard, train, and develop the next generation of researchers right now.

Key Takeaways:

• What it is actually like to learn research and AI at the same time, in real time

• How Angela balances keeping up with new tools while still doing the work

• Where the gap between academic foundation and real-world practice still shows up

• Why early career researchers have a perspective that senior leaders should be paying attention to

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Connect with Angela Chavez:

• LinkedIn

• Dig Insights
  

Andrew is back from his honeymoon, and Brian is very glad to have him. This one is part reunion, part rant, and entirely worth listening to.

First, the honeymoon stories. American Airlines lost three out of four bags for six days. The Vatican Special Police detained Andrew and his wife in the Sistine Chapel after mistaking them for a couple who had made terrorist threats. The Secret Service showed up at the wedding because of a drone and a Vice President flying nearby. All of this is true.

Then it gets into industry territory. Brian recaps the Verisoul-sponsored Cincinnati Reds event that brought together 40 companies and became an impromptu who's who of the Cincinnati research community, with a surprise appearance from Insights Association CEO Anita Watkins. The conversation builds into a genuine debate about who actually has the microphone in market research right now, whether the loudest voices at conferences are the most credible ones, and what it means that the biggest booths increasingly belong to companies with outside capital rather than the most rigorous methodologies.

Brian and Andrew close with a quick AI check-in, a plug for the upcoming July 9th webinar on the trust deficit in sample, and a tease of future episodes where they plan to screen share and show what they are actually building.

Key Takeaways:

• Why the Vatican Special Police detained Andrew and Brooke on their honeymoon, and how it ended

• Why Verisol's Cincinnati Reds event was one of the best community-building moments the local research scene has seen

• Whether the people with the loudest voices at industry conferences are the right ones to have the microphone

• What traditional research firms can do to compete for visibility in a world where outside capital buys the biggest booths

• Why AI is now part of daily life, not just a work tool, and what Brian has been quietly building with it

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Brian sits down with Samuel Cohen, CEO of Fairgen, for a no-fluff conversation about what is actually happening in synthetic data and who is positioned to win as the space matures.

Samuel cuts through the vendor noise fast: most companies claiming to train models on millions of survey responses are not doing that. What is under the hood is usually far simpler, and the lack of transparency is quietly eroding trust in the whole category. Fairgen's answer is to model at the individual level and be open about how it works.

The conversation also covers how collapsing time to insight is changing research workflows, where digital twins fit alongside real respondents in the next few years, and Samuel's unfiltered take on which players are well-positioned and which ones need to rethink their model now.

Key Takeaways:

• Why the data behind the model matters far more than which LLM is powering it

• What most synthetic vendors are actually doing versus what they claim

• Who wins as synthetic data matures and who is in trouble

• What are the best use cases for Digital Twins & Synthetic Data

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Connect with the CEO of Fairgen, Samuel Cohen:

• LinkedIn

• Fairgen
  

Andrew is in Italy on his honeymoon, a guest backed out, and Brian decided to go solo for the first time in 10 years of podcasting. No co-host, no guardrails, no agenda. Just 30 years of experience and a list of questions he has been sitting on for a while.

None of them has clean answers. Thats kind of the point.

Brian works through six questions the market research industry is not asking loudly enough: whether stacking fraud tools without coordination is quietly introducing a new category of data quality bias, who actually owns the definition of quality when nobody agrees, whether synthetic data is a legitimate solution or a convenient way to avoid the harder problem, where the next generation of researchers is coming from and whether the industry even knows what skills it needs to fill that pipeline, why market research is one of the only professions that shapes billion-dollar decisions without any required accreditation, and whether the M&A wave is actually good for research quality or just good for returns.

These are not gotcha questions. Brian is not here to throw anyone under the bus. But he is willing to say some things out loud that tend to get avoided in favor of AI hype cycles and vendor showcases, and this episode is the result of that.

Key Takeaways:

• Why stacking uncoordinated fraud tools may be creating invisible bias, not solving fraud

• Why the quality definition problem may eventually be settled by procurement departments instead of researchers

• The legitimate use cases for synthetic data and the less legitimate reasons adoption is accelerating

• Why the researcher talent shortage is really two problems bundled into one

• Why market research informs billion-dollar decisions with zero required accreditation

• Why consolidation looks like efficiency at first, and what history tells us happens next

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

In this episode, Brian and Andrew sit down with Dan Entrup, co-founder of AggKnowledge, to talk about one of the most overlooked drivers of the data quality crisis: the respondent experience nobody is actually fixing.

Dan tracked every research outreach he received over 15 months and turned it into a SampleCon presentation. The numbers were brutal. 137 times asked his age. 57 dental requests sent to someone who has never been a dentist. An inbox so flooded with irrelevant surveys that even a motivated expert panelist nearly quit. His point is sharp: bad targeting is not just inefficient, it is actively killing the panels we depend on.

The conversation gets into what AggKnowledge is building to solve it upstream, why free-text profiling and stale data are quietly sabotaging research operations at scale, and why identity verification and workplace verification are not the same thing. There is also a candid take on the M&A and funding landscape, who is well-positioned, where the yellow flags are, and why human knowledge only gets more valuable as AI takes over everything else.

Key Takeaways:

• Why bad targeting is a fraud accelerant, not just a waste of budget

• What 15 months of tracking his own respondent experience revealed about how broken outreach really is

• Why the gap between identity verification and workplace verification is where embellishment hides

• How upstream profiling data can reduce fraud before a screener is ever sent

• Why companies beating on revenue are still seeing their stock drop, and what that signals for the research industry M&A

If you loved the episode, have comments, or want to appear on the show, connect with us down below!

Connect with us:

• LinkedIn

• YouTube

• ROI Rocket

Connect with Dan Entrup:

• LinkedIn

• AggKnowledge

• It's Pronounced Data (Newsletter)