"The Data Doesn't Lie, But It Often Misleads": A Conversation with Automotive Market Research Veteran Dr. Renata Solberg
Introduction: A Career at the Intersection of Cars and Consumer Insight
Dr. Renata Solberg has spent more than two decades at the forefront of automotive market research, having led consumer insights functions at a major European OEM, founded her own research consultancy specializing in mobility and transportation, and served as an advisor to the European Automobile Manufacturers' Association (ACEA) on consumer adoption of electric vehicles. Her work has influenced product strategy decisions representing tens of billions of euros in capital allocation. We sat down with Dr. Solberg to discuss how the automotive research landscape is changing — and what researchers need to do to keep up.
On the Current State of Automotive Market Research
Let's start with the big picture. The global automotive market is going through an enormous transition. How is that changing the research questions your clients are asking?
The questions have become simultaneously more urgent and more uncertain, which is a difficult combination. Five years ago, the core commercial research question for most OEMs was some variation of: "Which customer segments will be loyal to our brand in the next product cycle, and what features do they prioritize?" That was a complicated question, but it was a fundamentally stable kind of complication. You could answer it with reasonably well-established methods — conjoint studies, brand tracking, customer satisfaction programs.
Now the questions are existential. Will consumers actually adopt battery electric vehicles at the pace that regulatory mandates require? How do you research a consumer experience — charging infrastructure, range anxiety, total cost of ownership — that a large portion of your potential customers have never actually had? How do you build a brand around a transition that your competitors are also navigating? Traditional research methods were not designed for that level of market uncertainty, and honestly, a lot of the industry hasn't caught up yet.
What do you mean when you say traditional methods weren't designed for this level of uncertainty?
Consider how most automotive market research works. You recruit a sample of car intenders — people who plan to purchase a vehicle in the next 12–18 months. You show them concepts, features, configurations. You measure purchase intent. You model the relationship between product attributes and willingness to pay. This is the conjoint-survey-based approach that has been the backbone of automotive product research for decades, and it works reasonably well when you're making incremental decisions — should the new model have a larger infotainment screen? How much will consumers pay for a panoramic sunroof?
"When you're asking consumers to evaluate a fundamentally new technology against their existing mental models — models built around decades of ICE vehicle experience — the data you collect may be systematically biased in ways you cannot detect or correct for."
EV research is different because you're not measuring preference between two familiar options. You're measuring openness to a new behavioral paradigm. Consumers who say they are "very interested" in purchasing an EV in a survey are not necessarily modeling the actual behavioral change that EV ownership requires. They may be expressing an aspiration, or a social identity preference, or a response to perceived social desirability — not a genuine purchase commitment. The gap between stated EV purchase intent and actual EV purchase behavior has been documented repeatedly, and it's larger than in almost any other product category I've worked in.
On Methodology: What's Actually Working
So if traditional stated preference research has limitations in the EV context, what methodological approaches are you finding more reliable?
A few things are genuinely moving the needle. First, extended test drives and real-world immersion studies. There's a well-documented finding — I've seen it replicated across studies in Germany, the US, and South Korea — that consumer attitudes toward EVs shift significantly after even a short period of actual EV use. If you give someone an EV for a week and have them go about their normal driving routine, their expressed concerns about range anxiety and charging infrastructure often decrease substantially. Conversely, some concerns that didn't surface in surveys — the cognitive load of planning charging stops on longer trips, the social awkwardness of explaining why you need to charge at a friend's house — become much more salient.
So immersive, experiential research designs are producing more predictively valid insights than static concept evaluation surveys. The challenge is that they're expensive, logistically complex, and harder to scale. A survey of 2,000 respondents costs a fraction of an extended use study with 150 participants. But the survey gives you precise estimates of imprecise things, while the immersion study gives you genuine behavioral insight.
What about the role of data analytics and passive data in automotive research?
This is where things get genuinely exciting — and genuinely complicated. Modern vehicles generate enormous amounts of behavioral data: how far people drive, when they charge, which features they actually use versus the features they said they wanted in pre-purchase research. Connected car data from platforms like OnStar or from OEM telematics systems can, in principle, give researchers a real-world behavioral validation layer that we've never had before.
The complication, of course, is privacy regulation. GDPR in Europe imposes strict constraints on how vehicle telematics data can be used for research purposes, and various US states are moving in similar directions. The research opportunity is real, but it requires careful governance frameworks and explicit consent architectures to use legitimately. OEMs who are investing in responsible data governance now will have a genuine research advantage in five years.
On the Competitive Intelligence Dimension
How is competitive analysis in the automotive sector evolving? The competitive set looks very different than it did a decade ago.
Completely different. The traditional OEM competitive intelligence framework assumed a relatively stable set of competitors — a handful of global OEMs competing in overlapping segments. Now automotive competitive analysis has to track Tesla, BYD, Rivian, Lucid, and a growing cohort of Chinese EV manufacturers entering Western markets. It also has to track technology companies — Google with Android Automotive, Apple with CarPlay and the rumored Apple Car project — as potential disruptors of the in-vehicle experience.
And honestly, the research community hasn't fully figured out how to do competitive analysis across such a heterogeneous competitive set. The tools that work for benchmarking one legacy OEM against another don't translate cleanly to benchmarking a software-defined vehicle company against a traditional automotive brand. J.D. Power's Initial Quality Study and Vehicle Dependability Study are still important benchmarks, but they measure dimensions of quality that may matter less to a Tesla buyer than a BMW buyer — and research teams need to be explicit about those measurement limitations.
On the Future of Automotive Research
Final question: what skills should automotive market researchers be developing right now?
Three things. First, data engineering literacy. You don't need to be a data scientist, but you need to understand what connected vehicle data, CRM data, and digital behavioral signals can and cannot tell you, and how to integrate them with survey data. Second, behavioral economics fluency. The EV transition is fundamentally a behavior change problem, and researchers who understand reference dependence, status quo bias, and loss aversion will design better studies and interpret results more accurately. Third — and this one surprises people — regulatory and policy literacy. The automotive market of 2030 will be shaped as much by EU Euro 7 standards, US EPA regulations, and Chinese NEV mandates as by consumer preference. Researchers who understand the policy context will provide far more valuable strategic guidance than those who focus purely on consumer attitudes in isolation.
The data doesn't lie. But without the right methodological infrastructure and the right analytical framing, it misleads all the time. That's the permanent challenge — and the permanent opportunity — of this work.