by Danny Sheahan

11 April 2023

It’s coming up on thirty years since I designed, administered and reported on my first survey for a paying client.  It was a customer satisfaction survey, and trying to maximise response rates back then meant including stamped addressed envelopes along with the printed questionnaire in a physical mailer to potential respondents drawn from the client’s customer base.

In the three decades since then a lot has changed.  The most radical change in survey research methodology has been the shift from paper-based surveys to online distribution.  While there is a time and a place for other modes – paper and telephone still have their roles – online has become the dominant survey mode.  This was first enabled by the rise of the internet, and then accelerated further by the adoption of smartphones, putting a computer into everyone’s hands.  This whole shift took time and there was a long transition period between paper-dominant to digital-dominant.

Recent months have seen the beginnings of another revolution which will impact not only market research but every aspect of our professional and daily lives.  Generative Artificial Intelligence (AI) has arrived and it wants to eat our lunch – or does it?

Like millions of others over the past couple of months I’ve been experimenting with and exploring the functionality of ChatGPT and other Generative AI applications.

As a market researcher I’m particularly interested in how generative AI will impact the market research profession and the opportunities that it offers.

First of all there is no doubt that this is a truly revolutionary development.  As with anything that is revolutionary there are some opportunities and threats that are clear.  There will be other opportunities and threats that are not necessarily clear right now but will quickly come to light in the coming months and years I am sure.

Among the many, many opportunities and risks that are emerging, here are just three of each that are immediately apparent in the market research field:

Opportunity #1 – Speed to data exploration and reporting – the summarizing abilities of ChatGPT are impressive and this is just the tip of the iceberg I feel.  Visual summary of quantitative data with the ability to slice and dice via natural language text or voice prompts can’t be far behind.  All of the parts are there and they will soon be joined up.  Access to the ChatGPT API and other APIs will accelerate this.

Opportunity #2 – Speed to insight – insights are still generated by human expertise, but we can get to that point more quickly with the almost instant completion of mundane tasks in the market research survey workflow.

Opportunity #3 – training AI to deliver insights – a tougher task but with enough data this is possible.  Is thinking “outside of the box” within the capabilities of Generative AI?  By definition generative AI is trained on existing information, so true innovation may not be strictly achievable just yet.  Creative extrapolation is possible though, and this can be a useful tool to prompt further data exploration by the researcher which may not otherwise have been on the radar.

Risk #1 – early implementations of using ChatGPT to design a survey questionnaire yield poorly designed questionnaires. This will improve over time, but for now I expect to see a spike in poorly designed questionnaires proliferating.  This can only and inevitably lead to risk #2…

Risk #2 – confidently wrong in analysis and conclusions – obviously there is potential here for catastrophic business decisions.  At Delve Research we are testing historical data (where we have carried out traditional analysis) on generative AI systems.  The speed of analysis is incredible.  So are some of the conclusions.  Again this is something that will improve over time, but for the time being the AI hallucination risk is something we should all be wary of.

Risk #3 – enabling smarter bots for survey fraud.  Survey fraud is already a major issue facing the market research industry.  Generative AI will only add to this with bad actors utilising the technology to build armies of fraudulent respondents.  The response of the market research industry in the face of this needs to be robust and rapid to preserve its credibility.

All of this is, of course, rapidly evolving and very exciting.  At Delve Research we are investing and upskilling in this area.  If you would like to share your thoughts on this topic or if you would like to find out more about how we are using generative AI in market research please get in touch here or email me directly danny@delve-research.com

A few notes:

This article was not written by ChatGPT.  Out of curiosity I did ask ChatGPT to paraphrase the article and you can see that version here.  Judge for yourself who is the better writer. ChatGPT will not be offended. I may be!

The image for this article was generated by Midjourney using my prompts

© Delve Research 2023