BLOG
September 6th 2023
We present: The already tenth edition of the marktforschung.de column "Herbert's World"!
This time it's all about the future - and back again! Where is our journey going - that is, the journey of market research? Who will keep the upper hand in "Human vs. Artificial Intelligence"? And is a fool with a tool actually still a fool?
Exciting questions, exciting answers. Have fun reading!
Goodbye science! Or how much artificial intelligence can market research take?
As is well known, in market research, in addition to the secrecy requirement and the separation requirement, the scientific requirement also applies. Formally, this means that all German institutes that belong to one or more of the established market and social research associations (ADM, ASI, BVM, DGOF) recognize the ICC/ESOMAR International Code of Conduct for Market, Opinion and Social Research and Data Analysis.
This sounds bureaucratic, but it is important: because in doing so, all participating institutes commit themselves to three professional ethical rules, including the aforementioned requirement of scientificity. According to this, all findings must be obtained using "appropriate methods and techniques of empirical research."
The application of all rules are explained in the full Code of Conduct, here again for those interested to read the German Declaration on the ICC.
In addition to the proven quality criteria of market research - objectivity, validity and reliability - it is above all the high-quality empirical research that ensures the very foundation of our industry: a foundation of credibility, trust and professional ethics!
And now, suddenly, artificial intelligence comes into play. And also our very own fear of the unknown to it, because AI is often an opaque and incomprehensible black box for us.
"Artificial intelligence will disrupt every stage of the research process". That's not me saying it, but Steve Philipps, CEO of Zappi, which recently secured $170 million in venture capital from investors to integrate AI into pre-testing. The goal, he says, is to reduce the cost of pre-testing from 20,000 to 2,000 euros. And I personally find that both technically and commercially exciting: how exactly will Zappi implement that technically? And how will the market react to it? And what actually happens when you cannibalize your own business model?
However, I agree 100 percent with Steve Philipps on one point: "AI will disrupt every single phase of the research process!“
In part, this statement already corresponds to reality. Because already today, long-established workflows in market and opinion research are being turned upside down. Ever new applications make almost all processes faster, more efficient and not least more cost-effective for our clients. And they create scope and resources that we can invest in supporting our customers or in new products.
The new Large Language Models (such as chatGPT or Bard) enable the fusion of qualitative and quantitative research. Limiting the amount of data or "too much data" are already concepts and problems of the past. AI-supported content as well as statistical analyses of almost any data set are possible. Yes, actually infinite!
So will the questionnaire become obsolete in the future?
Now the big question: if all these AI tools sweep through our work like a hurricane, isn't the aforementioned scientificity of empiricism also in danger? What is there to the predictions that "fieldwork" will soon be obsolete? That is, that questions will no longer be answered by real people, but will already be "known" by clever algorithms and thus be immediately available as a result for further exploitation by the customer?
Viewed quite soberly: At least for rather simple market research questions, be it about the best packaging design or typical A/B test questions, this scenario already makes sense today and is also possible with a few simple steps. I find it particularly exciting that there are already extensive discussions such as this one around the use of synthetic data.
This in turn means, in conclusion, that AI-generated-synthetic-empirical-data-from-the-black-box will simply be part of everyday life for everyone, virtually free of charge, and thus represent one of the greatest threats in the history of empirical research ever!
The future has already begun!
If I were Doc Brown from Back to the Future, I wouldn't have to travel far to get a glimpse of the future. Multilingual real-time avatars or software-as-a-service (SaaS) applications with hybrid or fully synthetic insights are already available or waiting in the wings. Take a look at Synthetic Users - currently still in BETA version, but already impressive today. AI-based apps are hitting the zeitgeist and, with their unlimited possibilities, are making the traditional products of classic institutions (which, by the way, also include the onliners) look old.
In concrete terms, this means a redefinition of their previous customer journey for the buyers of our services. Market research will suddenly be available 24/7, with results in seconds instead of days or even weeks. What a gain in efficiency and effectiveness! I'm getting a little giddy, but okay....
When do we even need "human intelligence" anymore?
I fervently hope that the "Humans in the Middle" principle will continue to apply to issues involving humans. I hope that whenever we observe, inquire about, or experiment with psychological and social dynamics, opinions, attitudes, desires, and needs, we go back to good old empirical research.
Because the newer, more complex or more sensitive a topic is, the more likely we are to need human rather than artificial intelligence to collect data. In other words, the field is not dead. No, it should please live for a long time!
Let's explore everything!
As a passionate market researcher, this has always been my dream, and now it's coming true: market research with and for everyone, literally! But it won't work without professionals, no matter how good AI is. After all, both data expertise and interpretation skills are our crucial assets that cannot be replaced.
After all, the following still applies: A fool with a tool is still a fool.
Due to the technology-driven quality leaps, the customer experience will change significantly and also shift. Whereas data collection and preparation used to be the main focus, the good market researcher will become a holistic coach. This means accompanying our clients from the initial identification of problems through the implementation of insights to development as a long-term consultant and partner.
If nothing else, these deeper relationships open up new revenue streams that can be lost on the other side through DIY tools. As with disruptive change, this will happen in all industries: The capable will seize the opportunities and become even better. But those who hesitate and cling to outdated processes will go under sooner or later.
KI in market research: more meat grinder or gold mine?
My conclusion: AI is clearly making us better. And in all processes around our products. Let's face reality: almost everything we can imagine today is already possible. But in a few years, what we can't even imagine today will also be possible!
Back to scientific principles: They are the DNA of good, successful market and opinion research. These principles must also be protected in the future and must not be washed away by an AI wave. Field research must therefore never be replaced by synthetic data, at least not when it comes to new and substantial insights about humans.
Rubbish in - Rubbish out.
What good are the best process and decision algorithms if the underlying data only reproduces existing patterns from the past and may even end up feeding on itself? High-quality, valid and resilient data remain the basis for all business decisions. For this, we need contact with real people, be it online, by phone or face-to-face.
We market researchers hopefully retain data sovereignty and data competence, but we take on a new role for our clients so that they do not drown in the vast sea of data: As a "chief understanding officer" or "chief consulting officer," it's more than ever about understanding, consulting, and accompanying.
Herbert Höckel is a managing partner here at moweb research GmbH. He has been a market researcher for more than 25 years. In 2004 he founded moweb GmbH, which he is still the owner today. moweb from Düsseldorf operates internationally and is one of the first German market research institutes specializing in digital processes.