BLOG
17. September 2026
Herbert’s World #23
We present issue No. 23 of the Mafo.de column “Herbert’s World”!
Our CEO, Herbert Höckel, reports on his unique experience as moderator of a fascinating debate at the recent ESOMAR 2026 Congress in Valencia: The main focus there was on speed versus rigour (as a sort of “AI industry vision”, so to speak), as well as on who should be asking the quality questions NOW in order to define what good AI-supported research will mean over the next ten years.
Enjoy reading!
Relevance or Rigour? What the ESOMAR Congress 2026 revealed to me!
By Herbert Höckel
It was a complete stalemate. And that was the most honest answer our industry could have given this year.
At the 2026 Esomar Congress in Valencia, I chaired a debate that will stay with me for a long time to come – not just because of the arguments themselves, but because of the silence that hung in the room after the vote. The question on the table was: ‘In the age of AI, does the market research industry need to become faster, even if that means accepting more errors?’
Two teams, two compelling positions, and a room full of experts with strong opinions. The blue team, led by Adriana Rocha and Tim Bock, defended the importance of rigour. Their key argument: only through methodological rigour can we secure the foundation of our credibility. The yellow team, led by Mathilde Beljarts and Adrian Terron, countered this and stood by the importance of relevance. After all, anyone who is too slow (and cautious) in the market will no longer be in demand and will therefore become irrelevant.
The audience voted. It ended in a dead heat, 50 per cent to 50 per cent.
No winners. No losers. But one question remains: What now??
I must admit: deep down, I had been hoping for a winner. For a result that would signal the direction we should take. Instead, we have been given something else, perhaps even something more valuable: a perfect draw, which only at first glance reveals the disagreement between two camps, but which thereby raises fundamental questions that our industry has had to grapple with ever since the popularisation of ubiquitous AI models.
The uncomfortable truth behind the draw
When a specialist audience – comprising experienced researchers, decision-makers from the insights industry, research organisations, suppliers and even clients – responds to a fundamental question about their own professional identity with a deadlock, this is not an indication of indecisiveness. Rather, it is a precise measurement showing that the industry is, in fact, divided. However, not into two opposing camps, but into two attitudes that both co-exist within the same person.
For the debate has shown one thing just as clearly as the deadlock: almost everyone in the room could have signed up to both sides. Everyone is aware that AI-generated insights are sometimes too hasty, too superficial and insufficiently validated. And at the same time, we know that a three-week project which is only delivered the day AFTER the client’s board meeting is of no use to anyone. This tension is real and (unfortunately) cannot be resolved any time soon. So it must be endured – on a daily basis, in specific assignments, in client meetings and in the choice of methodology.
What impressed me most about this debate was the uncompromising honesty on both sides. We didn’t pretend to have answers that we don’t have. That’s unusual for our sector. And it’s the right start.
The restless decade and what Kelly Beaver (Ipsos) has to say about it
This moment forms part of a bigger picture. Kelly Beaver – who has been Global CEO of Ipsos since this summer – described it aptly in her keynote speech in Valencia: We are in a phase where AI is no longer new, yet is still not fully understood. Whilst we are gaining a better and better understanding of what it can do, we do not yet know what we want to achieve with it, nor to what extent it will dominate our processes in the long term.
In this context, Beaver highlighted something that I consider to be crucial:
The market research industry’s ability to provide clear and precise answers to our clients’ critical queries regarding data quality is not just one of many skills. It is our core competence. Anyone who does not know – and cannot explain – where the data comes from, how reliable the basis is, or how much confidence can be placed in a decision has failed in their actual task, no matter how elegant the dashboard looks.
At a time when literally anyone can generate insights with a single click, the ability to champion data quality is becoming a key differentiator – if not a reason for existence.
Or to put it more bluntly: if speed is ultimately one of THE key factors for success, then trust is THE key currency for sustainable customer relationships. Speed is therefore necessary, but trust remains the sufficient condition. In practice, this should mean supplementing a quick 80 per cent solution with the important information on which parts of the results should be treated with caution.
Trust is not a virtue. It is an economic factor.
Here, I would like to recall an idea that was a recurring theme in most of the discussions at the 2024 Congress in Athens and which is most clearly articulated in Philipp Kristian’s book The Trust Economy. Its central thesis is: ‘Trust leads to innovation. Distrust leads to disruption.’
At first glance, this sounds like a management cliché. However, it is a structural observation with far-reaching consequences – particularly for our sector.
After all, what happens if clients no longer trust the insights because they have been produced by an algorithm in a ‘black box’ and even the institution itself can no longer vouch for them? What happens if, in a client’s eyes, research shifts from being a foundation of evidence to an optional extra? From a ‘conditio sine qua non’ to a ‘nice-to-have’, when time and budget (occasionally) still allow it?
Disruption rarely begins with a better competitor. It begins with a gradual loss of relevance. And you usually lose relevance not because of one big mistake, but because of many small moments when you weren’t there, weren’t quick enough or weren’t clear enough.
That is the very crux of this debate: it is not a question of speed versus thoroughness, but rather that we must define as quickly as possible what constitutes good market research today and, above all, tomorrow. Because if we fail to do so, customers will automatically turn even more to the platforms that are already making full use of their speed advantage.
What I have learnt as a presenter and what worries me
At this point, here is another idea, drawn from my plan for preparing for the discussion. I had prepared questions for any ‘difficult moments’ that might arise between the parties; one of them was: ‘Hands up – which of you has, in the last twelve months, delivered an AI-generated result to a client that they did not fully trust because the pressure of the deadline outweighed the certainty?’
I didn’t end up asking that question. Not because it would have been rude or inappropriate, but simply because I was afraid of the answer and of the number of (honest) hands that might have gone up. Well, but what does that say about the state of our industry??
What worries me about this is the following: we discuss rigour versus relevance as if it were a philosophical question. Yet it has long since become an operational one. Most companies in our sector still do not have clear standards for when AI-generated or AI-assisted output is deemed ready for delivery. There are no internal quality checks for synthetic data. No clear governance on who bears responsibility if something goes wrong. So we’re already talking about attitude, whilst we don’t even have any functioning processes yet.
This is not a judgement – it is a critical assessment of the current situation. And it explains why the audience voted 50/50. Not because both positions are equally valid, but because most people in the room sensed that we are still only at the beginning of this debate. In far too many areas, we still lack the expertise, control and certainty when it comes to AI. So we have not yet decided who we want to be!
Why now is exactly the right time?
There is a window of opportunity. It will not remain open indefinitely. Gartner would say we are approaching the ‘Plateau of Productivity’ – that point in the Hype Cycle where expectations either translate into real value creation or they do not. Those who set the standards now, who ask the right questions about quality rather than accepting AI as a black box, and who define what good AI-supported research means – they will shape the face of our industry for the next ten years.
One thing is clear: if WE don’t do it, others will. The management consultancies. The technology platforms. The VC-funded research AI start-ups, which don’t have to take methodological tradition into account because they have none.
That is why I see the Esomar AI Alliance, in conjunction with the Esomar Professional Standards Committee, not merely as an effective debating club, but as a strategic necessity. But not to force consensus, because Valencia has shown that consensus cannot be the goal. Rather, it is to sharpen the questions, highlight the pain points, test the positions and develop the courage to tolerate uncomfortable answers.
The 50/50 result was an accurate diagnosis and thus both a wake-up call and an invitation to continue this important work.
In conclusion: a word of thanks and a request
I am grateful to Adriana, Tim, Mathilde and Adrian – for their willingness to take a clear stand and to embrace a challenge. To the audience in Valencia, who contributed their thoughts, voted and asked questions. To Esomar, which creates and defends this space.
I have a request for everyone reading this column: keep this debate going. Within your organisations, with your clients, with your teams. Not as an abstract discussion of principles, but as a concrete operational question: what is our standard for quality? Now and in the future? Who decides this? And what do we stand for?
The sector that answers these questions for itself will be the one that people trust. It will be the one that stays ahead of the (technological) wave and is able to keep pace with future developments and then integrate them with confidence.
Trust leads to innovation. Distrust leads to disruption.
The choice is ours!
Herbert Höckel ist geschäftsführender Gesellschafter hier bei bei der moweb research GmbH. Seit mehr als 25 Jahren ist er Marktforscher. 2004 gründete er die moweb GmbH, welche er bis heute als Inhaber führt. Die moweb aus Düsseldorf ist international tätig und eines der ersten deutschen, auf digitale Verfahren spezialisierte Marktforschungsinstitute.