How is AI going to impact market research
and insights over the next ten years?

Decorative imagePublished by Ray Poynter, 9 February, 2023

NewMR has conducted a study to assess how market researchers view the future impact of AI and Automation. You can download the report by clicking here. The study repeats most of the questions we asked in 2021, allowing comparisons to be drawn.

The three main points that we can see in the data are:

  1. People expect AI / Automation to have a big impact on quantitative research and the coding of open-ended responses, and a smaller, but still substantial, impact on qualitative research.
  2. People can see the speed, cost and efficiency benefits, but many of them worry that quality could suffer, especially if the human element is lost.
  3. The views expressed in 2021 are remarkably similar to those reported in 2023 – which means the market feels that the impact of AI and Automation as far away as they were two years ago.

My feeling is that that study (the average view of the people we spoke to) understates the likely speed of the impact of AI and Automation.

As part of the analysis we used Word Cloud Plus to kickstart our analysis of the open-ended comments, as shown below. Which highlighted the belief that most of our research participants, felt that the human element is still going to be necessary and there were several potential negatives, mostly relating to quality.

Word Cloud about AI and Automation

Remember, you can use Word Cloud Plus for free to generate your word clouds.

4 thoughts on “How is AI going to impact market research
and insights over the next ten years?

  1. the thing that is amazing me about the sudden surge in interest and disccussion about AI based platforms is :
    1. its old news … there have been very strong AI based platforms for doing MR around for a decade. from colation tools ( as my AI academic friendssay ” the easy part of AI usage ). Including platforms for using open source data to explore, analyse, track and predict people and brand narrative actions as well as emotional response. I know I attended and spoke at a number of indstry conferences around this subject and been asked to speak about the use of existing platform ( and how they had tested as often more accurate than trad quant
    especially for tracking )
    2. the bigger “missing elements” in current discussion was around the simple truth that AI platforms will disrupt so called “secondary research” more than any other form. there are already many platforms that use all the exisiting contentof the internet, of google etc to provide summary and prediction around subjects. having been advocating for decades that “secondary” should actually be “primary” ( as in done first, done better and if done right elimnates the need for other qual and quant ) revolutionises the way we think about research, and indeed how research companies should be thinking of their business models.
    would love to join more open discussion of the real changes AI platforms are and will be making

  2. I think it’ll be a good while before AI is able to approach the nuanceand sophistication of the human brain for understanding human behavior. As a pre-eminently, social species on the planet, our brains have evolved to do that specific work. Meanwhile, full disclosure – I’ve always said that “big data“ may be able to tell businesses, the “what, where, and win“ of consumer behavior, but it’s unlikely that this type of data, no matter how sophisticated its analysis, will ever be able to answer the all important “why” questions. And of course the answers to those “why” questions are critical for understanding how to motivate people – which is in the end, the first order task of marketing.

  3. From the studies I have run, most research is tactical and pretty simple. I agree that AI replacing the real insight stuff is some way off, but I think it will blow a big hole in the 80% which is pretty standard.

  4. Hi Dave, it’s big news now because now it is not just the cognoscenti who get it. One of the big changes will be the use of secondary research first (or usually instead). For example, I just asked it some usage and demographic questions for coffee in the UK. It gave me an opinion (I have not checked how good the estimate was). I then offered it two names for a new coffee product (an instant coffee that has lots of upmarket values and advertising and asked it to assess two names. It assessed them. I then asked it to suggest which I should use if I wanted to target the richest 25% in the UK, and it picked “Coffee Gold”. Let us assume for a moment that everything it said was complete BS – I can still imagine plenty of brand managers spending 5 minutes with a Generative AI too, rather than paying and waiting for research. With a bit of skill you can get ChatGPT to support almost any semi-plausible idea, by tweaking the questions. So, if a brand manager ‘knows’ their idea is right, they are frustrated by research which costs money and takes time and could give the wrong answer – so why not turn to a Generative AI for validation?

Comments are closed.