Bad Habits, Bad Research: The Case for Formal Education in Market Research

The Insight250 spotlights and celebrates, annually, 250 of the world’s premier leaders and innovators in market research, consumer insights, and data-driven marketing. The awards have created renewed excitement across the industry whilst strengthening the connectivity of the market research community. Winners of the 2025 Insight250 were announced last September - you can see the full list of Winners, and those from previous years, at Insight250.com. The 2026 Insight250 nominations are currently in review.

With so many exceptional professionals named to the Insight250, we regularly tap into their expertise and unique perspectives on a range of topics. This regular series does just that: inquiring about the expert perspectives of many of these individuals in a series of short topical features. 

With insights advancing at an incredible pace and the value of insights ever increasing, I sat down with Insight250 Winner Michaela Mora. Michaela is the Program Director, MS in Marketing Research and an Assistant Professor of Practice at University of Texas at Arlington. She is also the Founder and Chief Research Officer for Relevant Insights, a full-service market research and UX research firm.

Crispin: You came into market research yourself from a background in psychology and public relations, working across agencies in Cuba, Sweden and the US before founding Relevant Insights in 2007. How much of your own formation was learning by doing, and at what point did you conclude that the industry had a structural education problem rather than just an individual one?

MM: In all my years in this field, most market researchers I’ve worked alongside never formally studied marketing research. Some came from psychology or sociology, which are disciplines that sit at the very roots of our field and give you a solid methodological foundation, but they often arrived without a business lens. I know this firsthand because that was me. My first degree was in social psychology before I studied marketing, PR and advertising at the School of Communications at Stockholm University, where I heard about marketing research for the first time.

When I started at Research International in Sweden, I was still learning on the fly. Others I met came with marketing or communications degrees and had the business fluency but were essentially winging the methodology. Over time, experience fills some of those gaps, but slowly and usually with a few bad habits baked in along the way.

I realized that even after two degrees and nearly ten years of practice, I had blind spots. The truth is that your experience only stretches as far as the organizations you’ve worked in will allow. For me, the fastest path forward wasn’t another job, but specialized education in marketing research. While learning by doing is always ongoing, the MSMR program offered a significant leap forward in my career development. I’m not the only one who experienced this.

The two decades I’ve since spent working directly with clients have only confirmed this education issue in our industry. Time and again, I’ve seen many client-research teams struggle with what should be the foundation of our work: defining a business problem clearly, translating it into a research question, and then choosing the method that actually fits the purpose. Instead, many jump straight to requesting a data collection method (a survey, a focus group, a tracking study, etc.) without fully understanding its limitations, or whether the data it produces will even answer the questions they started with.

Crispin: You are a graduate of the MSMR programme at UT Arlington and now its Programme Director. What did completing that degree formally change about the way you thought and worked, even after years of practice?

MM: That program changed the trajectory of my career because it rewired how I think about research. I was exposed to a bigger toolkit and, more importantly, I understood when and why to use each tool in the context of business problems. One of the most liberating realizations for me was that no method is perfect. Every approach carries its own caveats. Once I accepted that, my job as a researcher became about designing research systematically, in a way that mitigates risk and accounts for the limitations I know I’m bringing into the room.

It also showed me something I hadn’t fully appreciated before: qualitative and quantitative research complement each other. When you need both breadth and depth, neither alone is enough. The numbers tell you what’s happening, while the human stories tell you why. Together, they can create insights that actually hold up. I learned to get involved at every phase of the research process, not just the design stage or the final presentation, but everything in between. That hands-on engagement is an important way you protect data quality. It’s how I make sure that what lands on a stakeholder’s desk is useful, not just technically defensible.

Finally, the most practical lesson of all was to learn how to navigate the moment when budgets shrink and timelines compress. I learned to adjust, knowing where compromises would cost quality and what that meant for the insights I could honestly claim to deliver. I believe that insights that mislead are worse than no insights at all.

Crispin: The MSMR at UTA is one of very few dedicated postgraduate qualifications in marketing research anywhere. Why does the industry have so few of them, and why has the case for formal education been so difficult to make stick?

MM: For decades, marketing research scaled within agencies and corporate research departments through an apprenticeship model. People learned by doing under senior researchers, on live projects, and with real deadlines. That model worked well enough that a formal educational pipeline never became a prerequisite the way it did in, say, medicine or law. The habit of growing your own talent became deeply embedded. As we all know, the field sits between disciplines, so no one owns it. It draws from psychology, sociology, statistics, economics, ethnography, and business, but it belongs fully to none of them. That interdisciplinary nature means universities have no natural academic home for it. It falls awkwardly between business schools, communication, and social sciences schools, and neither has historically championed it with much conviction. Without a clear institutional owner, dedicated programs simply don’t get built.

Also, the industry never made formal credentials a hiring requirement. Unlike professions that gatekeep entry through qualifications, marketing research never did. Agencies and clients hired for curiosity, analytical instincts, and communication skills, and then trained people themselves. When credentials don’t move the needle in hiring decisions, the incentive to pursue them weakens, and the incentive for universities to create them weakens further still.

Organizations like ESOMAR, the Insights Association, and the MRS in the UK have long advocated for professionalization, but unlike bar associations or medical boards, they have no power to restrict practice. Anyone can call themselves a market researcher. Without that gatekeeping function, certifications and qualifications remain optional enhancements rather than entry tickets.

The rise of DIY research has also made the case harder to make. The democratization of survey tools, cheap data, and now AI has created an illusion that research is essentially accessible to anyone with a platform subscription. That perception, however flawed, actively undermines the argument that deep, specialized training matters. Why fund a postgraduate program for a skill that a bootcamp claims to cover in four hours?

Finally, research quality is difficult to measure directly. A well-designed study that prevents a bad decision is invisible since the disaster simply doesn’t happen. That makes ROI arguments for formal education frustratingly abstract, especially when the alternative, like hiring someone cheaper and training them on the job, appears to work fine on the surface.

The deeper irony is that the industry suffers visibly from exactly the knowledge gaps that formal education would address. The case for it isn’t hard to make intellectually. It’s hard to make structurally because the system has organized itself around not needing it.

The creation of the MSMR program at UT Arlington was driven by the clear vision of two individuals at a time in the late 80s when Marketing Science was becoming established and professionals were adopting quantitative methods and statistical modeling for marketing problems.

Jerry Thomas, CEO of Decision Analyst, and the late Carl McDaniel, who chaired the Marketing Department at UT Arlington in the late 1980s, recognized that the industry had a talent problem it wasn’t solving and that the apprenticeship model, however entrenched, was too slow and too expensive to be the only answer. Training someone from scratch takes years and significant resources, and there’s no guarantee they’ll develop the right instincts even then. The establishment of the first MMR program at UGA in the late 70s, known for cultivating outstanding researchers, affirmed the advantages of employing formally trained marketing professionals.

What the industry needed was a faster, more deliberate path that produced people who arrived already understanding how to translate business problems into research questions and research findings into decisions with an advanced methodological foundation. The MS in Marketing Research program at UTA College of Business offered another opportunity to fill that gap. For those of us who have gone through it, the impact has been lasting.

Crispin: You describe many professionals entering the field by chance. What are the specific bad habits that get acquired in the process? When you look at research being produced today by agencies, in-house teams and AI-assisted platforms, what are the telltale signs that the people behind it lack grounding in the fundamentals?

MM: One of the most telling signs of weak fundamentals is a research brief that jumps straight to objectives and methodology without ever clearly articulating the business problem. What decision is being made? Who is making it? What would change their course of action? Researchers without formal training or a lot of experience in this field often treat the brief as a formality before the real work begins. But if the problem isn’t defined precisely, everything that follows is built on sand.

Another telltale sign is when researchers often fixate on sample size numbers without interrogating how they were recruited, what the response rate was, or whether the sample actually represents the population the client cares about. A common question I get is, ‘How large should my sample be to be representative?’ This stems from confusing the accuracy of an estimate with the representativeness of the sample itself and a lack of understanding of sampling methods. The basics are missing here.

Also, when researchers are trained in one tradition and not the other, which is common, they tend to treat the two as separate work streams rather than complementary lenses. Qual findings don’t inform the survey design. Survey results don’t get explored deeper through follow-up in-depth interviews. The opportunity to triangulate, to use one method’s strengths to compensate for another’s weaknesses, gets left on the table.

Worse, they become beholden to the law of the hammer. If the only tool you know is a hammer, every problem looks like a nail. A researcher who came up through qualitative work reaches for focus groups or IDIs regardless of whether the question actually requires depth or scale. A quantitative researcher defaults to a survey even when the topic is unexplored, and a structured instrument will only produce the illusion of measurement. The method stops being a response to the problem and becomes a reflex, chosen out of comfort and familiarity rather than fit for purpose.

When you look at research coming out of agencies, in-house teams, and AI-assisted platforms, the fingerprints of these habits are everywhere. Research from agencies often shows the pressure of commercialization and many use templated approaches sold as solutions before the problem is fully understood. Although short deadlines and small budgets are drivers of this problem as well, the research design and resulting insights are markedly worse when the fundamentals of those approaches are weak.

In-house teams, meanwhile, frequently suffer from proximity bias. They’re so close to the business that they often unconsciously design research to validate decisions already made rather than test them. Research often becomes a political instrument rather than a decision-making one.

AI-assisted platforms have introduced a new and particularly insidious version of these problems. They make it trivially easy to generate surveys, analyze open-ended responses, and produce summary reports, which means the volume of research has exploded while the quality of thinking behind it has not kept pace. The interface looks professional and the outputs look rigorous, but if the person operating the platform doesn’t understand sampling, questionnaire design, or the difference between correlation and causation, the AI simply automates and scales their blind spots.

Garbage in, garbage out, just faster and with better fonts. A telltale sign of a lack of grounding in research fundamentals across all three is often research that answers the question asked rather than the question that mattered. It takes solid grounding in the fundamentals to know the difference.

Crispin: The industry has been talking about data quality for years, but you describe the current situation as a crisis. What has changed? Is this a problem that has materially worsened, or one we are finally being forced to acknowledge?

MM: Data quality problems in marketing research are not new. Satisficing, speeding, straight-lining, and professional respondents gaming survey systems have existed for as long as panel-based research has existed. The industry has tolerated a slow deterioration in data quality for years because the economics rewarded speed and price over rigor, and because the consequences were diffuse and hard to attribute directly. What has changed is not the existence of the problem, but its scale and speed.

A generation ago, many client-side research buyers had agency experience. They understood the process. They could ask hard questions about methodology, sample composition, and data quality, and suppliers knew they would. That cohort has largely retired or moved on. Their replacements grew up in a world of dashboards, automated reporting, and research-as-a-service, in which untrained researchers learned on the go, aided by technology, without supervision from senior researchers. They’re often more comfortable interpreting outputs than interrogating inputs.

The arrival of self-serve survey platforms in the late 90s was very valuable in many respects. I was among the first users of SurveyMonkey and SurveyGizmo, so I speak from experience. It put research capabilities in the hands of people who previously had none. But it also removed the layer of professional judgment that used to sit between a business question and a fielded study. Now a product manager with a Qualtrics subscription and a SurveyMonkey Audience panel can have 500 responses by tomorrow morning. Often there is no time or budget to test the questionnaire for bias or to check whether the sample is appropriate. Often internal clients don’t have the skills to clean the data before the deck gets built. The volume of research being produced has multiplied, but the quality control infrastructure has not scaled with it.

Online panels were supposed to democratize access to research participants. In many ways they did, but the business model created bad incentives from the start. Panel companies compete on price and speed, which means margins are thin, and thin margins mean corners get cut on recruitment, verification, and engagement. Panels became polluted with respondents who are there purely for the incentive, clicking through surveys as fast as possible.

The research industry kept buying from these panels because the alternative was slower and more expensive, and clients kept demanding faster and cheaper. The argument can be made that the affordability of samples today is illusory, given that the associated expenses have been moved to the backend. A $5 per completed survey paid to the panel ends up costing the client $30 or more when they add time and needed tools if they decide to clean the dataset. Many don’t do it, so this cost is invisible to them.

Some estimates suggest fraudulent responses can account for a third or more of the

data collected. The tools for detecting this have improved, but they’re in a constant arms race with the tools for evading detection, and many research buyers have no visibility into what their data supplier is actually doing to protect them.

Generative AI has now introduced a threat that the industry was not prepared for: synthetic respondents based on data that clients can’t verify the quality of, adding another layer of distance between the raw signal and the business decision.

Crispin: The perennial tradeoff between quality, cost and speed is well documented, and speed and cost reliably win. But that pressure has always existed. Why is a deficit in research fundamentals a more significant contributing factor now than it was twenty years ago?

MM: Twenty years ago, the trade-off was real but contained. When I started out, the infrastructure of research had built-in friction, and that friction was actually protective. Data processing was a distinct phase with distinct people who caught problems before they reached the analyst. The process itself slowed things down enough that errors had more opportunities to surface, and the people involved at each stage- interviewers, coders, data processors and moderators - brought craft knowledge that acted as a distributed quality control system even when nobody was explicitly thinking about quality control.

The perennial trade-off between quality, cost, and speed was always going to favor speed and cost, but when the process had structural depth, that trade-off had a floor. You could compromise on speed and cost, and still land somewhere defensible on quality, because the system caught the worst mistakes for you. That floor has dropped now. And without grounding in the fundamentals, many researchers today don’t even know they’ve fallen through it.

Most of the layer of human infrastructure has been stripped out in many organizations. Today, a researcher with a platform subscription and a panel account can move from brief to data in 48 hours with little human intervention. Every guardrail that used to slow things down and catch mistakes has been removed in the name of efficiency. What’s left is the judgment of the person running the study. If that person doesn’t have strong fundamentals, there is nothing else standing between a flawed design and a business decision. The speed that platforms advertise as a feature has eliminated the checkpoints that used to compensate for incomplete knowledge.

On top of that, the nature of the decisions that research is now being asked to support has grown more complex, not less. Today, research is being asked to keep pace with rapidly shifting markets, inform agile product development cycles, and feed into algorithms and models that amplify whatever signal or noise they receive. The stakes in getting the fundamentals wrong have risen at exactly the moment when the people doing the work are, in many cases, less equipped to get them right.

Crispin: Your career spans client-side roles at Blockbuster Online and Match.com, agency work across three continents, and now academia. Where in that chain, client, agency, platform or analyst, does the deficit in research fundamentals do the most damage to the quality of the eventual output?

MM: Early in my career, I would have said the agency because that’s where the research was actually designed and executed, and a methodologically weak agency can undermine an entire study regardless of how well the client briefed it. And that’s still true. But after two decades of working directly with clients across industries, I’ve come to believe the most consequential deficit sits on the client side, at the briefing stage, before a single question has been written or a supplier has been contacted.

If the client doesn’t know how to define the business problem precisely, if they can’t articulate what decision they’re actually trying to make, what they already know, what would change their thinking, or there are internal political forces, bruised egos or simply hubris, driving the research in a particular direction, then the agency is working from a flawed foundation from the first conversation.

A strong agency can push back, probe, and help reframe a weak brief. But in my experience, that only happens when the agency has both the expertise to recognize the problem and the commercial confidence to slow things down and have that conversation with the client. Both conditions are rarer than they should be. What I’ve seen far more often is a weak brief met with a compliant proposal. The agency gives the client what they asked for rather than what they need, because questioning the brief risks the relationship or the budget. That dynamic of a client who doesn’t know what they don’t know, and a supplier who doesn’t push back, is where research quality dies most quietly and most completely.

The client side is not monolithic, of course. In my experience, the most dangerous position in the entire research chain is the mid-level client-side research buyer, experienced enough to operate independently, not senior enough to have accumulated the hard lessons, (sometimes simply under the grip of the Dunning-Kruger Effect) and under enough internal pressure to deliver quickly that there’s no incentive to slow down and interrogate the design. Senior research leaders often have the instincts, even when they lack formal training. Junior researchers know they don’t know things and tend to ask questions. It’s the middle layer that moves fastest with the most confidence and the least oversight.

Researchers with deficient research fundamentals tend to rely more on technology that can normalize poor practices at scale. When a platform is designed to move a user from problem to fielded survey in the fewest possible clicks, the interface itself is making methodological decisions that the researcher should be making consciously. The platform’s design philosophy quietly becomes the research methodology, and the probability of noticing declines because the output looks clean and professional.

A weak analyst can take reasonably good data and drain it of meaning; however, if the brief was weak, the design was flawed, and the data is compromised, even a skilled analyst is working with too little to produce something useful.

So, if I had to place the deficit where it does the single greatest damage, I would put it at the client-side problem definition stage because a well-defined problem with a clear decision at stake can survive a lot of downstream imperfections. Research built on a well-defined problem, even if imperfectly executed, tends to land somewhere useful. Research built on a vague or misdirected problem, executed flawlessly, produces irrelevant or misleading answers with great confidence. And in my experience, that second scenario is far more common.

Crispin: There is a lively industry debate about AI-generated respondents and synthetic data panels. From the perspective of someone who teaches research fundamentals, how do you assess that debate? Is it a tool that demands more methodological rigour from its users, or one that papers over the cracks?

MM: This is a debate that I follow closely. I find it interesting, but the legitimate use cases are narrower than the industry wants to admit, and in the hands of people without strong fundamentals, it is one of the most effective crack-papering tools ever invented.

From where I stand, as someone who teaches fundamentals, the deeper issue is not the technology but the conditions under which it’s being adopted. A researcher with a strong grounding can engage with it critically: interrogating whether a synthetic panel represents a population or merely models one, whether AI-generated attitudinal responses measure anything real, whether a vendor’s validity claims hold up. A researcher without that grounding can’t do any of those things.

Synthetic data demands more methodological rigor from its users than traditional research, precisely because the outputs look superficially convincing. In a world where fundamentals were widely held, this technology could find its appropriate niche. In the world that actually exists, where fundamentals are patchy, speed and cost pressure is intense, and some clients often can’t distinguish good research from bad, it is more likely to accelerate the crisis than resolve it.

Crispin: You hold both an MSMR and a UXMC and have spent years working at the intersection of market research and UX research. Do both disciplines share the same educational deficit, or does each have its own distinct version of the problem?

MM: Both disciplines share the same structural problem regarding education. Most practitioners in both camps have learned on the job and carry the gaps shaped by where each field came from and what it was designed to do.

Marketing research grew out of social science and statistics. Its foundational methods, such as surveys, sampling, experimental design, and focus groups, have decades of methodological literature behind them. The educational deficit here is largely about methodology, if practitioners don’t have education in those disciplines, or business translation if they understand the methods but can’t connect them to business decisions.

UX research has a different origin story and a different version of the problem. The field grew rapidly inside technology companies, populated by human factors specialists, anthropologists, and interaction designers who brought depth in human behavior but arrived without a shared methodological framework. There was no equivalent of the survey research canon to anchor practice. I initially approached it thinking it was something totally different research-wise, since they use different terminology. I soon realized that the value for me was learning about the principles of interaction design, which helped me communicate better with digital product managers and designers. Regarding research methods, I found it to be too narrow and too focused on a few qualitative research methods and techniques (usability testing, contextual inquiry, “user interviews,” etc.), which are variations of in-depth interviews and ethnographic observation.

Over time, UX research has evolved into digital product research, concerned with how people interact with digital interfaces, often at the level of micro-interactions (navigation flows, onboarding friction, feature discoverability, etc.). That granular focus is valuable, but it comes with a significant blind spot.

A lot of UX research is conducted in isolation from the broader marketing context that frames the product. Questions about whether the market actually needs the product, how it should be priced, how it will be distributed, and how it will be promoted, which are the foundational pillars of marketing, are outside the frame. For years, UX research focused on optimizing the user experience with less regard for market needs or business outcomes.

Fortunately, a growing number of UX researchers are now engaged in needs research, which they refer to as ‘generative’ or ‘discovery’ research. Marketing researchers know this type of study as ‘exploratory research’ using qualitative research methods.

Given its origin story, UX researchers often have strong instincts about human behavior and skills in qualitative observation, but many lack the statistical grounding to know when quantitative methods are needed and how to execute them credibly. The pressure inside technology companies to move fast has created a culture where ‘research’ sometimes means a handful of interviews conducted the week before a product decision, with findings that carry more confidence than the evidence warrants.

I see UX research as a specialty within the marketing research family of approaches, but market researchers and UX researchers have a hard time talking to each other. They sit in different parts of the organization, use different terminologies, and sometimes actively compete for budget and influence. There are a lot of misconceptions about marketing research in the UX community, and most marketing researchers failed to be curious about UX for a long time, thinking that they didn’t have the skills to do this type of research.

The educational deficit in both cases is ultimately the same deficit. Limited by the established practices in the companies they work for and the tools at hand, many research practitioners in both camps are unaware of a more extensive research toolkit or how to effectively apply it to business problems. They often fail to recognize the limits of the methods they know or reach across disciplinary lines when the problem demands it.

What I bring from holding both credentials is an understanding of each tradition’s blind spots and a conviction that the most meaningful research happens when you refuse to stay inside either one.

Crispin: You have argued publicly that personas stripped of demographic context are a methodological mistake and that much UX research misunderstands what market research can contribute. Is that a training failure, a disciplinary silo problem, or something else?

MM: It’s all three, and they reinforce each other in ways that make the problem stubbornly resistant to easy fixes. Personas as a tool were conceived primarily as a communication device to give product teams a shared, humanized mental model of the user they were designing for. That origin shaped everything about how personas developed inside the UX practice. They became narrative artifacts: named, storied, visually represented, emotionally resonant.

And in that communicative function, they work reasonably well. The problem is that somewhere along the way, personas stopped being a communication tool and started being treated as a research output. I have seen personas created by marketing, sales, and UX teams based on internal assumptions driven by user scenarios with no data or idea of how representative or large those use case scenarios were.

Stripping demographic context from personas became fashionable partly as a response to legitimate concerns about bias in design. I’d argue that you can design inclusively and still need to know who your market is, how large each segment is, and what differentiates them in ways that matter commercially.

Personas have also been a popular artifact in marketing research to help communicate segmentation profiles to marketers and internal stakeholders. The segmentation frameworks we use in market research combine behavioral, attitudinal, psychographic, and demographic dimensions because each layer adds something the others can’t provide. We ground those personas in actual data, not on assumptions. We know that demographics are weak differentiators of segments, but they are useful in connecting a segment to media consumption patterns, purchasing power, household structure, and life stage in ways that make the segmentation actionable for pricing, promotion, and distribution decisions. Without that grounding, a persona floats free of commercial reality.

I also see this in the context of how UX research has developed its professional identity in partial opposition to what it saw as the limitations of traditional market research. There are a lot of misconceptions about marketing research, which has become associated, in parts of the UX world, with a kind of reductive, survey-heavy approach that misses what really matters about human experience. This has led to criticism that has hardened into a disciplinary posture that is sometimes more about differentiation than accuracy.

This characterization is unfair and outdated, but it persists, and it creates a resistance to learning from market research that has no methodological justification. To close the divide, we need training that crosses disciplinary lines, organizational structures that facilitate the dialogue between UX and market researchers, and a degree of intellectual humility on both sides that professional identities don’t always make easy.

I’ve seen it work when those conditions exist. I started doing UX research before it was a thing. I worked in integrated teams at Match.com and Blockbuster Online that simply did research. The research that comes out of those collaborations is consistently better than either discipline produces alone.

Until the training catches up with the reality of what good product research actually requires, personas without demographic grounding will keep producing beautifully designed products that may or may not reach their intended market.

Crispin: Industry bodies, certification schemes and short courses have all attempted to raise the floor on research literacy. Why has none of that proved sufficient, and what can a rigorous postgraduate degree do that those interventions cannot?

MM: The short answer is that certifications and short courses treat symptoms, while a rigorous postgraduate degree addresses the underlying condition. Industry bodies and certification programs have done useful work in codifying standards and creating a shared vocabulary. The certifications I’ve encountered seem to offer only superficial knowledge, often from short courses. While these might touch upon fundamental concepts, they don’t allow for the supervised practice necessary to correct errors and misunderstandings.

Short courses can transfer knowledge effectively. A two-day workshop on questionnaire design can improve someone’s survey writing, but knowledge transfer and capability development are not the same thing. Knowing that leading questions introduce bias is not the same as having the internalized judgment to recognize one in your own work, under deadline pressure, when a client is pushing for confirmation of what they already believe. That judgment doesn’t develop in two days.

There are four key things a full MSMR degree, like ours, offers that short programs don’t:

1. It builds integrated knowledge rather than modular knowledge. Short courses leave researchers with a limited set of tools that might or might not talk to each other. A postgraduate program builds the connective tissue, helping us to understand when and why to combine methods, and how each compensates for the other’s weaknesses. That integration is precisely what the MSMR gave me.

2. It develops systematic thinking by being required to justify methodological choices, defend design decisions, and engage with existing literature. This builds a mental discipline that becomes reflexive. That changes not just what you know but how you think.

3. In a structured academic program, flawed thinking gets challenged by faculty, by peers, by literature that has already worked through the problems you’re encountering for the first time. That corrective feedback loop doesn’t exist in on- the-job learning or in short courses at the same intensity, if at all.

4. It teaches the ability to move fluently between a business problem, a research problem, a methodological solution, and a strategic recommendation. That full chain of reasoning is what a well-designed program like the MSMR was explicitly built to develop.

The challenge is that postgraduate education requires institutional commitment from universities, employers willing to value the credential, and individuals willing to invest the time and money. Until employers systematically reward formal research education in hiring and promotion decisions, the incentive to pursue it remains weaker than it should be. The degree can do what short courses cannot; employers and aspiring researchers need to recognize the difference.

Crispin: You co-founded the Multicultural Insights Collective and run a widely read research blog. Both are forms of continuing education at scale. What is the relationship between that kind of public thought leadership and the formal institutional education you now oversee?

MM: The blog was my initial foray into educating about research fundamentals. It was designed to help those who use and purchase research understand how their decisions impact research quality. It became something useful to practitioners who had no other access to this kind of structured reflection on research practice.

However, I recognize the limitations of thought leadership as an educational intervention. The blog reaches people who are already curious and already engaged. It can shift thinking at the margins and occasionally spark a reorientation in how someone approaches their work. However, it can’t build the integrated, systematic capability that formal education enables.

A blog post, however good, cannot replicate the experience of having your methodology challenged by a faculty member who has spent years studying it. Thought leadership widens the conversation. Formal education deepens the practitioner.

The Multicultural Insights Collective emerged from a specific moment, the social justice reckoning of 2020, when a group of research agencies, including my own, Relevant Insights, came together to help brands understand what diversity meant for their customers.

What began as a response to a cultural moment revealed something with permanent methodological implications as we talked to different cultural segments. I realized diversity and rigor are not separate issues in our field.

When the people designing research don’t reflect the populations being researched, experiential blindness gets built into the work at its most fundamental level in how problems get framed, which questions get asked, which methods get chosen, and how findings get interpreted. That is a quality problem.

Multicultural competence isn’t an add-on to research rigor. It is research rigor, properly understood. This is why I’m very proud of our MSMR program offering a multicultural research methods course, teaching students not just that diverse populations require adapted approaches, but how to execute those adaptations with the same systematic discipline the program brings to every other methodological question.

The case I have been making throughout my blog posts, work with clients, and now teaching is that research done well is not a commodity. The difference between rigorous research and the appearance of rigorous research is consequential for the decisions that get made, the resources that get allocated, and ultimately for the people whose lives those decisions affect.

For me, the link is education. Education can do a lot to preserve the quality of research.

Crispin:

Michaela, thank you for such a thorough and clear-eyed conversation. Your point that the deficit does the most damage at the client side, before a single question has even been written, is one I think many in our industry would rather not confront, but it rings true. And your line about AI simply automating and scaling existing blind spots, garbage in, garbage out, just faster and with better fonts, is one I suspect will stay with a lot of readers, myself included. What comes through in everything you have said is that rigour is not a box to tick but a discipline to be built, whether that is through formal education, multicultural competence, or simply the willingness to ask what decision we are actually trying to improve. It has been a genuine pleasure, and I have no doubt this conversation will prompt some overdue reflection across the industry. Thank you again.