AI - bane or boon for market research?

We live in an exciting time for market research. We are still in the nascent stage of AI unlocking the true potential of data, with ChatGPT and similar chatbots already demonstrating the plethora of possibilities AI will bring. 

ChatGPT burst onto the scene at the end of 2022 and has become the biggest tech headline of 2023. It has already passed law and medical board tests and scored higher than humans on MBA exams. It can write entire articles, draft code and build websites. Educators are already developing apps to detect ChatGPT "cheating" in assignments. The chatbot caused immense panic in the tech world and forced Google to swiftly respond with the launch of Bard in early February. 

Eventually, AI will reshape how researchers work, help us sift through historical data to identify patterns and trends, and predict future consumer behaviour with far more accuracy. However, while AI and machine learning will revolutionise the way we analyse data, they will not replace real human behaviour, responses, and analysis at our industry's core. 

The use of AI and machine learning in market research is nothing new. But never has it been so widely accessible and easy to use. It is now evident and inevitable that AI will be increasingly incorporated into our work practices as market researchers, changing it fundamentally. 

Faster, cheaper, more accessible  

Undeniably, this type of technology will make market research faster and cheaper. Projects that previously took weeks or months may only take hours. AI can analyse millions of voice and text comments in minutes to better understand customer behaviour and come up with relevant follow-up questions. I used ChatGPT recently and asked it to generate two questionnaires, which it did with speed and accuracy. 

Automating time-consuming tasks will enable researchers to focus on higher-value work such as client servicing and data storytelling. In the near future, using triangulation to create a more holistic brand picture will gain significant momentum. Combining sources such as customer sales data, survey data, and footfall will result in a much more comprehensive research outcome. 

AI is not going to be a foolproof solution for market researchers. There needs to be a degree of caution practised here. With unfettered access to tools like ChatGPT and Bard, our industry runs the risk of manufacturing the exact same ideas as we are all using the same AI tool. Where does that leave our original thinking and USP as a data and insights company? 

Another consideration is that AI is only as good as the material it is trained on. ChatGPT, for example, is only trained up to 2021 materials and has "limited knowledge of world and events after 2021". Given the speed with which consumer preferences and markets shift and evolve, this presents a significant hurdle. 

Ultimately, it will be the lazy researcher that looks to AI alone to unlock the power of data. However, utilising it as one of the many weapons in our arsenal is what will serve us well. 

Increased diversity 

Diversity and inclusion are top priorities for organisations today. Inclusive market research involves reaching out to underrepresented groups to learn their specific needs and preferences. It's easy enough to uncover the 'what' in market research, but the 'why' is more challenging because it is often influenced by different cultural and demographic factors. Understanding this enables organisations to develop and refine products and services to break into new markets and maximise sales. 

Using AI removes unconscious human bias from the research process while also effectively capturing information from underrepresented groups that brands are keen to target. From the overall design of a study to how respondents are interviewed and follow-up questions asked, AI has the capacity to be far more objective and culturally considerate than a human. With solid objective research that considers a wider demographic, brands can then create hyper-personalised campaigns that resonate with individual consumers. 

For this to be successful, it’s important that the information used to train AI algorithms is free of bias. On the other hand, it is unfortunately also possible for AI to amplify bias. One famous example is when Amazon used AI for hiring staff. Because Amazon traditionally had far more male than female applicants, the algorithm, which was fed resumé data, learnt to view men as more suitable than women. 

The evolution of the market researcher  

As with most life-changing technology, there are fears that AI will make many roles redundant. But as history has taught us, most of the time jobs simply evolve. 

Labour-intensive tasks such as hand-coding verbatims are currently being managed by market researchers through the creation of classifiers. However, creating classifiers is also a time-consuming, fiddly process. What the likes of ChatGPT can do for us is take over these processes, meaning that entry-level market researchers will be starting their careers with a very different set of foundational skills. Junior and mid-level managers will see their capacity open up because AI will be handling these basic tasks. 

Given this change, the industry needs to consider how we equip those entering the market research industry. If the foundation of data collection and survey writing will be the job of AI, what other fundamental skills will market research graduates need to arm themselves with instead?  

Conclusion  

Seeing how AI transforms market research through 2023 will be interesting, but it is important to keep in mind that AI and machine learning are upgraded tools in our research toolbox. Market researchers and human responses are still what will bring market research to life. 

Ultimately what AI will do is make market research more accessible to companies, and generate more powerful and more timely insights. This will hopefully translate into more clients looking beyond "gut feel" when considering strategic business moves and instead turn to market research to back up or guide their growth plans.