Menu
Close
.svg)
One year ago today on LinkedIN, I published my first in-depth conversation with David Berman, exploring the past, present, and future of pharmaceutical marketing research. A lot has changed since then. Over the past twelve months, the healthcare and biopharma landscape has faced unprecedented turbulence -- from shifting regulatory and pricing policies to the explosive adoption of AI and analytics tools that are redefining how we understand customers and make decisions.
So it felt like the perfect time to reconnect with David -- one of the industry’s most insightful and forward-looking voices -- to get his perspective on how marketing research and insights are evolving in this new era of complexity, speed, and technological transformation.
For those who may not know him, David Berman is the former AVP of U.S. and Global Commercial Capabilities at Merck & Co., now a senior advisor and consultant to a range of healthcare and life science organizations, and a Marketing Research Professor at Fairleigh Dickinson University. He advises leadership teams on commercial insights and analytics, organizational design, digital transformation, brand and corporate strategy, and the integration of emerging technologies such as AI into marketing and decision-making frameworks.
In this follow-up discussion, we explore how pharma’s commitment to insights is evolving, where the balance between analytics and human understanding is shifting, and what the next decade of customer insight might really look like.
Noah Pines: Over the past year, the political and regulatory climate has become increasingly volatile. We’re seeing growing discussions around drug price controls, Medicare negotiations, and potential scrutiny -- even potential elimination of -- direct-to-consumer (DTC) advertising. These macro forces seem to be reshaping the economics of pharmaceutical marketing. From your perspective, how are these shifts influencing how pharma companies allocate their budgets for marketing research, analytics, and insights?
David Berman: You’ve really touched on several of the major concerns dominating the industry right now.
The first is pricing. When top-line revenue is under pressure, there’s inevitably less funding available for ancillary areas like marketing research. It’s a classic case of resource compression -- as margins tighten, discretionary budgets start to shrink.
The second, which I find particularly interesting, is the structural change in distribution and negotiation dynamics. Take, for instance, the rise of direct-to-patient models such as TrumpRx. By bypassing middlemen like PBMs, the system is being rewired. The big question becomes: where does that reclaimed margin go? Does it return to the patient, or does it get absorbed elsewhere in the value chain? That uncertainty is forcing companies to rethink the economics of patient access and engagement.
With profitability under strain, organizations are becoming much more deliberate about value creation -- not just in marketing, but in the insights functions that inform it. Market research teams can no longer afford to execute studies just because “we’ve always done it this way.” Every project now has to pass a litmus test: Will this research directly influence a business decision that creates value -- either by driving revenue or strengthening the asset’s positioning? That kind of strategic selectivity is new, and it’s reshaping the role of insights teams.
If DTC advertising were to disappear altogether -- which, under this Administration isn’t entirely implausible -- the impact would cascade across the ecosystem. Consumer testing, ad pretesting, creative development -- all of that would shrink dramatically. I remember early in my career, before DTC was permitted, when the industry operated under a self-imposed moratorium. When that lifted, we saw an explosion of consumer-focused work. Losing that again would represent a major contraction -- not just for advertisers, but for market researchers as well.
That said, there’s an emerging bright spot. Many companies are doubling down on patient services and support programs. These initiatives blend analytics, patient experience, and value-added engagement, and they’re increasingly becoming the new frontier for commercial innovation. So, while traditional marketing research may be consolidating, new domains of patient-centric insight are rapidly expanding to take its place.
NP: Given all this regulatory turbulence, if DTC advertising were to be curtailed, how might that reshape the way pharma engages patients? Are we looking at a shift toward unbranded communication, social platforms, or something entirely different? And as analytics and AI advance, do you see them empowering insights -- or crowding them out?
DB: It’s a great question because we’re standing right at the intersection of those two shifts — communication and capability.
If DTC gets dialed back, the next phase of engagement will almost certainly be unbranded. We’ve seen this movie before on the HCP side -- physicians grew tired of heavily branded messaging, so unbranded education and disease-awareness content filled the gap. I think the same playbook will apply to patients. There will still be communication, still be outreach, but it will live under the banner of “information” rather than “promotion.” The real challenge is how closely that ecosystem will be monitored, especially when you’re talking about social media, influencers, and public relations-driven engagement.
On the analytics front, the shift in spending is already happening. Many organizations have redirected a meaningful portion of their budgets from traditional insights toward advanced analytics and AI. That doesn’t mean market research is disappearing -- far from it. Every major pharma company I know still has robust research teams, and has not materially scaled back budget. What’s changed is structure and scope: research and analytics are being consolidated into single insights functions, where people are expected to be multidimensional -- part researcher, part analyst, part forecaster, even part competitive intelligence.
This convergence is healthy, but it also introduces new myths -- chief among them, that AI can replace human insight. It can’t. AI can process information at scale, synthesize patterns, even simulate reasoning. But it doesn’t think. It doesn’t understand context, trade-offs, or the emotional drivers that shape real-world behavior. That’s where humans come in — to interpret, to challenge, and to ensure the insight is not just statistically significant but strategically meaningful.
So, yes, analytics is expanding, but the best organizations aren’t letting it crowd out insight. They’re marrying the two -- using AI to accelerate understanding, and people to ensure that understanding translates into business impact.
NP: You’ve mentioned that analytics budgets aren’t necessarily shrinking -- but every model eventually hits its limits. Where do you see the point of diminishing returns, where only direct customer insight can fill the gap? In other words, where is qualitative input still irreplaceable?
DB: That’s a great way to frame it; because there really is a boundary to what analytics and AI can do, at least today. The clearest example is innovation and new product development --places where there’s no historical data for the machine to learn from.
AI performs beautifully when it’s extrapolating from an established dataset. It can model preferences, simulate market shares, even approximate behavioral patterns -- and it’s getting increasingly precise, often within a ±10% range based on some of the latest benchmarks I’ve seen, including one recently published in Harvard Business Review. But when it comes to predicting something truly novel -- a new mechanism of action, an untested delivery format, or a first-in-class therapy -- AI just doesn’t have the experiential base to stand on.
That’s where human insight becomes indispensable. Talking to HCPs, patients, and payers gives you the nuance -- the “why” behind decisions -- that no algorithm can yet infer. Synthetic data can mimic attitudes, but it can’t replicate the complexity of human motivation. Over time, AI will improve, but for now, when you’re operating at the frontier of innovation, it’s still the qualitative insight that reveals what numbers can’t.
NP: We’ve talked a lot about the explosion of interest in AI -- especially in generative applications for research. From your vantage point, what’s legitimate today and what’s still hype? I’m thinking about things like synthetic respondents or AI-augmented interviewing. What’s real, what’s promising, and what’s just noise?
DB: It’s a great distinction to make, because right now there’s a lot of enthusiasm -- and a fair bit of confusion -- about where AI truly adds value.
The real capability today lies in situations where the machine has data -- even a modest amount -- to train on. In those cases, AI can be remarkably effective at handling structured or closed-ended questions. It can supplement human input, automate parts of analysis, and deliver useful directional guidance. In that sense, it’s a powerful accelerator of insight, not a replacement for it.
The hype, however, comes from the idea that synthetic respondents can fully replace qualitative or even large portions of quantitative work. We’re not there yet -- not even close. I’ve seen experiments with AI moderators and synthetic avatars -- and they’re fascinating. Some physicians actually enjoy engaging with them; it feels novel. But you quickly hit a wall: these systems can’t yet sustain deep, nuanced conversations or probe meaningfully beneath surface responses.
What’s ironic is that many people assume AI will dramatically shorten project timelines by removing fieldwork. In practice, it just shifts the time upstream. You spend it training the model, designing lexicons, refining prompts, and QA-ing linguistic accuracy. So, while it’s cheaper, it’s often not faster -- at least not yet.
That said, AI is proving hugely valuable in pre- and post-research tasks -- designing discussion guides, summarizing transcripts, surfacing patterns across large qualitative datasets. Those efficiencies are real. The myth is that it replaces the human element. The reality is that it makes well-trained humans exponentially more productive.
NP: Given today’s pricing pressures and increasingly constrained R&D and commercial budgets, how should companies think about allocating their limited resources across insights, analytics, and newer technologies like AI and data visualization? Where should those dollars really go?
DB: It really depends -- both on the company and where they are in the product life cycle. But if I were sitting in the CMO chair of a commercial organization today, I’d start by cutting back on redundant ad testing. Much of promotional advertising has become over-engineered. At the end of the day, if you have a truly differentiated product -- something with clear clinical value and a strong data story -- it will sell itself.
Now, if you’ve got a “me-too” product, differentiation becomes essential. That’s where I’d focus more on A/B testing and analytics-driven optimization -- running controlled digital experiments, particularly on social channels, to understand what really moves the needle. Consumer brands do this constantly, yet pharma still lags behind in using data to test and refine messaging in real time.
From a broader perspective, I think companies need to be much more intentional about where research adds value. Early-stage market research, especially when integrated with clinical development, is where the real ROI lives. That’s where you define the product’s positioning, the value narrative, the access strategy -- essentially, whether the asset will ultimately be reimbursable and reach patients. Market research and analytics should be intertwined from the very start, not treated as separate silos.
AI and visualization, meanwhile, are becoming critical enablers. Imagine a marketer with an “Alexa for brand strategy,” a system that can integrate internal data with external signals like weather patterns, epidemiological data, or even social sentiment. Think about flu vaccines: combining your sales data with CDC feeds or National Weather Service data could allow you to anticipate outbreaks and pre-position supply. Retailers like Home Depot were doing this years ago: integrating sales, inventory, and weather data to literally move snow shovels to the front of the store ahead of a storm. There’s no reason pharma shouldn’t be operating at that same level of precision.
So yes, invest heavily in AI and data infrastructure, but don’t underinvest in the human insight that gets you to launch. That’s still the bridge between science and strategy. Post-launch, though, there’s room to trim some of the traditional spend — particularly in areas where the value-add isn’t what it once was. The goal isn’t to spend less; it’s to spend smarter.
NP: Post-launch, where do you think companies can pare back without jeopardizing impact? Which parts of the budget no longer carry the same weight?
DB: I’d start with DTC promotional spend, particularly traditional ad testing and production. If policy shifts continue toward tighter restrictions on direct-to-consumer advertising -- or if it’s curtailed altogether -- that’s an obvious place to look for savings. Even if it remains permissible, we’re already seeing diminishing returns from many of those campaigns. Those dollars can be better redirected toward patient engagement programs or analytics-driven optimization efforts that actually shape real-world behavior.
NP: We’ve spent a lot of time on AI and policy, but what other macro trends do you see reshaping the landscape? You’ve mentioned consumerization, digital therapeutics, and precision medicine — what forces do you think will most disrupt how pharma perceives and uses market research over the next three to five years?
DB: I tend to think of the industry’s evolution like a series of signposts on a winding road: you can see where things might turn, but you never quite know which direction the next bend will take.
If the industry becomes more regulated on the commercial side -- in advertising, communication, and engagement -- we’ll inevitably need less traditional market research around message testing and promotional tactics. But that doesn’t mean less insight overall. It just means insight will need to come from different places. We’ll see a shift toward cross-functional intelligence -- R&D, clinical operations, commercial, and even medical affairs will all become insight generators. Social listening, EMR/EHR data, and advanced analytics will feed into a much more integrated view of the business.
Looking ahead, I think the defining trend will be the integration of insights across the enterprise, anchored around one idea: putting the patient at the center. You’re going to see companies reorganize around the patient journey -- not just the prescriber. That means building systems that connect the physician, the nurse practitioner, the physician assistant, and every digital touchpoint that surrounds the patient experience.
We’re already feeling the strain in primary care -- physician shortages, burnout, reduced visit time -- and that creates both a challenge and an opportunity. Patients will increasingly turn elsewhere for information, whether that’s digital platforms, peer networks, or healthcare apps. The smart money is already flowing into technologies that integrate personal health data -- wearables, medical devices, EMRs -- into a single, user-controlled ecosystem. That convergence of data and access will redefine what “insight” even means.
In short, the next five years aren’t just about new tools; they’re about rethinking how we listen. The companies that thrive will be the ones that can connect those signals into something coherent, actionable, and deeply human.
NP: We’ve talked a lot about the adoption of AI and machine learning within pharma itself, but there’s another side to this story: our customers are also changing. Physicians, hospitals, and health systems are increasingly leveraging AI through EHRs, apps, and digital portals. If you were advising a pharma C-suite, what would you tell them to consider about how AI will affect their customers?
DB: I’ve always been something of a purist when it comes to pharma. At its core, what this industry does best -- and should continue to focus on -- is discovering and delivering lifesaving, life-changing therapies. That’s our purpose. My advice to pharma executives would be simple: don’t try to become a tech company. Building apps isn’t your core competency; building cures is.
Now, that’s not to say AI doesn’t belong in pharma. When AI becomes an integral part of a therapeutic solution -- for example, as part of a diagnostic device or treatment platform -- that’s different. That kind of integration belongs in the R&D domain, where it enhances discovery, development, or delivery. Having worked across R&D, clinical, manufacturing, and commercial functions, I can say with confidence that the biggest, most strategic investments should still be directed toward the science itself.
Where I do see AI playing a transformative role is in patient services and engagement. That’s where pharma can, and should, evolve. AI-driven tools can help patients navigate care, access support, or understand their conditions better. But I’d stop short of calling it a “more knowledgeable customer base.” Patients may soon have access to more data, but data isn’t the same as expertise. The real challenge ahead is managing what happens when AI-generated recommendations begin to diverge from what physicians advise.
That’s the challenge we’re walking into: a world where AI is increasingly influencing patient behavior, sometimes in ways that may conflict with clinical judgment. The companies that succeed will be the ones that respect that balance -- supporting informed patients without undermining the trusted physician relationship that remains at the center of healthcare.
NP: When you look at platforms like Epic, it’s remarkable how EHRs are reshaping physician decision-making. I’ve written before about the rise of "anticipatory medicine" -- where systems like Epic will soon have billions of data points that inform not just clinical choices, but the economics behind them.
Do you think pharma really understands how its professional customer base — the HCP — is evolving in this environment? My sense is that we still sometimes operate as if there’s a single, autonomous decision-maker. In reality, physicians are now constrained by time, economics, patient input, and EHR-driven prompts. It’s a very different dynamic than what we saw 10 or even 20 years ago.
DB: You’re absolutely right -- and I think we’re going to see a clear generational divide in how physicians adapt. Younger doctors, those who grew up with an iPad in hand, are far more comfortable relying on technology to guide or even co-author their decisions. They’re native to this kind of augmented decision environment, whereas older physicians are still adapting to it.
At its core, though, this transformation is about productivity. Integrated delivery networks (IDNs) and large health systems are sitting on millions of patient records. They’re using that data to identify what works best: which treatments deliver the strongest outcomes at the lowest cost. Increasingly, they’re feeding that information back to physicians through EHR prompts: “For first-line hypertension, here are the two medications that have shown the best results in our population...”
That changes the whole commercial equation. The locus of influence is shifting -- from the individual physician to the system itself. When clinical pathways are data-driven, pharma’s promotional leverage diminishes. The question becomes: how do manufacturers engage when prescribing decisions are being algorithmically guided?
To their credit, pharma companies are starting to recognize this. Many are forming partnerships with IDNs and health systems to share and co-develop data -- not just to prove value, but to design future assets that fit seamlessly into outcome-based frameworks. That’s where the next wave of competitive differentiation will happen: aligning with the systems, not just selling to the prescribers.
NP: Last year, one of your big calls to action was around the speed of insights -- moving from weeks to hours. How close are we to truly real-time insight in pharma today? And if we’re not there yet, what are the biggest roadblocks?
DB: We’re making progress, but we’re not there yet. The challenges depend a lot on the type of work you’re doing. If you’re running a large primary market research project, you still have to recruit, field, and analyze. Even with AI now supporting everything from questionnaire design to respondent recruiting and report summarization, timelines of four to six weeks are still common.
Part of it is a skills gap -- knowing which tools to use, how to integrate them, and whether to build or buy those capabilities. Interestingly, 30 years ago, we had something similar: a company called FIND/SVP. You could literally call them, ask a question, and get an answer within hours. Today, with AI, we can replicate that kind of immediacy -- and even surpass it. The question is whether organizations are ready to trust those systems.
Another hurdle is data governance. Everyone’s struggling with how to keep client data isolated and compliant while still enabling AI to learn across datasets. My dream scenario would be for the industry to create a shared AI-enabled database -- a secure ecosystem where pharma companies contribute anonymized data to accelerate discovery and innovation. We’re starting to see the early steps toward that, but there’s still a long way to go.
NP: You’ve long advocated for insights professionals to hold a point of view, not just report data. But in an environment that prizes objectivity, how do you coach teams to do that without crossing into bias?
DB: To me, that’s the essence of what separates a great insights leader from a data reporter. Yes, you have to present the objective findings -- that’s table stakes. But your real job is to translate those findings into business action. That means understanding the problem the organization is trying to solve, the internal politics at play, and how to use data to influence the right decisions.
Insights professionals need to be both truth-tellers and navigators. You have to be strong enough to tell a marketing lead, “Don’t spend $3 million on this campaign -- the data doesn’t support it.” And you need to know how to bring in other evidence -- KOL input, secondary analytics, field intelligence -- to strengthen your case.
Training for that takes years. You start by exposing junior researchers to marketing teams early, letting them listen, contribute, make mistakes, and learn the language of the business. There’s no substitute for that kind of apprenticeship. Unfortunately, because people move between companies so often now, that long-term mentorship is harder to sustain. It’s something the industry needs to rediscover.
NP: Given how demanding the environment has become, what are you seeing in terms of recruiting and retaining quality respondents for research? Any innovative approaches out there?
DB: That’s a tough one -- and it’s been a challenge for as long as I can remember. I’d love to see us build something like a social network for panel members -- a “Facebook of physicians,” where participation isn’t just transactional but community-driven. If respondents find real professional value in the network -- learning from peers, sharing cases, building relationships -- they’ll keep coming back.
We also need to acknowledge the limitations of traditional panels. Even when we run studies with 100 cardiologists or oncologists, we have to ask: is that truly representative? Probably not. Panels give us better information, not perfect information. And yes, fraud is an issue. While AI-powered fraud detection is improving, finding engaged, qualified participants remains one of the hardest challenges in pharma research.
NP: Since we last spoke, have you changed your view on how insights and analytics teams should be organized? Any new models or capabilities that feel over-hyped or under-appreciated?
DB: I still think we’re in a bit of the AI hype cycle, but it’s starting to normalize. Companies are beginning to understand what AI is actually good at -- and where it isn’t ready. It’s making a real impact in R&D and early discovery, but on the commercial side, expectations are still catching up to reality.
From an organizational standpoint, I remain convinced that market research should be independent. It should partner closely with marketing, but not report into it. When your boss owns the $3 million campaign you’re advising against, that’s a built-in conflict of interest. Independence protects objectivity — and credibility.
That said, today’s market researchers must be fluent in analytics and technology. I’m teaching at Fairleigh Dickinson University, and every student I meet can code in R or Python. In a few years, that will be the norm. The next generation of insights professionals will be true hybrids: equal parts strategist, technologist, and storyteller.
NP: Let’s finish with a bit of future-casting. Looking ten years out, what does the world of pharma insights look like to you? Is it fully AI-augmented? Self-service? Something else entirely?
DB: If I could design the future, I’d create a universal insights engine -- a platform that ingests data from everywhere: social media, clinical trials, real-world evidence, even satellite data -- and turns it into real-time guidance for marketers and strategists.
But even with that level of automation, you’ll still need humans who can contextualize and interpret. AI will get faster and smarter, but it can’t yet understand the nuances of organizational culture, ethics, or human emotion -- all of which shape how insights are acted upon.
Over time, we’ll need fewer people, but they’ll be more strategic and multidisciplinary: fluent in business, science, analytics, and behavioral insight. My hope is that AI eventually gets good enough to project multiple future scenarios -- not to predict perfectly, but to help us see what’s possible.
Of course, all of this will require massive computing power. The energy costs alone are staggering -- we’re talking about global-scale investments in infrastructure. But for those just starting out in insights, my advice is simple: learn the data, master the tools, and never lose sight of the business problem you’re solving. That’s what will define the insights professional of the future.
NP: More integrated, more multidisciplinary — and hopefully, faster.
DB: Exactly. My dream is for insights to operate in real time -- where decisions are informed instantly, not weeks later. We’re not there yet, but that’s the direction we’re heading. Until then, the human insight function remains not just relevant -- but essential.
