Technological advances—such as the proliferation of various generative Artificial Intelligence (AI) and use of language learning models like Chat GPT, plus the increasing sophistication of market research bots—are affecting the market research industry and necessitating an adjustment to best practices and policies when conducting surveys.
As a custom market research firm, our continual aim here at The Farnsworth Group is to keep pace with industry changes, adjust our best practices, and integrate the highest quality controls to stay up to date and produce reliable findings to guide our customers’ important business decisions.
Top Takeaways from the Quirk’s Event
To get new and professional insights about recent developments in the market research industry, several team members from The Farnsworth Group attended The Quirk’s Event in July. This annual in-person conference is designed to support marketing research and insight professionals and provide tools and information to help us drive our practices forward in the 21st century.
Here is a look at some of the emerging and prevalent trends in the market research industry and the measures being taken by our team at The Farnsworth Group to stay on top of changes, continually implement best practices, and ensure the reliability of research findings:
1. The Spread of AI Presents Opportunities and Challenges
The explosion of generative AI such as Chat GPT is sure to have an indelible impact on the way market research is conducted. Industry professionals are experiencing a mixture of emotions and concerns surrounding the benefits and potential ramifications of AI.
A significant portion are worried about the possibilities that AI presents, while a smaller segment is curious and filled with optimism. Professionals at the conference also expressed concern that AI will threaten their job, while others believe it will make their job easier.
In response to the pervasiveness and ever-evolving nature of generative AI, we’re integrating technology platforms and survey design practices that will identify AI and terminate surveys or reject surveys that involve suspected fraud. We are actively improving the standards we use when creating robust survey questions and when searching for “human elements” in data responses to weed out fake respondents.
For example, we are
- looking for grammatical errors in long-form answers, AND grammatical perfection as Chat GPT won’t make spelling mistakes unless prompted;
- analyzing the length of responses to open-ended questions, as people use an average of just nine words, while Chat GPT averages more than 200;
- asking for self-identification or self-detection;
- and looking for more personal answers that reference life experiences.
Also, since AI is programmed to be politically correct, and won’t use swear words, we are considering and testing techniques to exploit this tell of AI to identify real human vs fake respondents.
“Even though AI is the biggest technology disruptor since the iPhone, the human element is still important. AI is a tool, not an oracle. While people’s roles may change, their expertise is still needed to ask it the right questions, verify, and interpret the information.” - Kimber Kreilein, Project Director at The Farnsworth Group
2. Sample Quality is an Industry-wide Challenge
The struggle continues for research firms and in-house insights teams alike, working to ensure quality sample in surveys, as there has been an increase in bad actors—bots or those using AI to quickly supply answers that aren’t thoughtful or genuine—over the past three years. Plus, survey respondents are becoming increasingly sophisticated, which makes it challenging to spot them among a group of legitimate respondents.
Here are 5 real examples showing the process we use to prevent fake responses from affecting research output, insights, and recommendations.
Unfortunately, there is not one simple solution for dealing with this, but we can use a multi-pronged approach. One way we can help address this issue is by designing research surveys that are shorter in length and more simply structured. This will enable suppliers to eject bad actors from all databases quickly and efficiently by frequently sharing information on rejected respondents.
Additionally, we can use server-to-server integration (S2S) to eradicate ghost completions and help mitigate fraud on the customer end. Also, we have to be aware of the survey incentive rate; if it’s not valuable enough, real people won’t invest their time, and it will be difficult to achieve an adequate sample size of genuine respondents.
3. Making Research Tell a Dynamic Story is Critical
Another struggle is that market research can be perceptively dry and unengaging, making it difficult for people to understand and interact with the data. To deal with this challenge, it’s important that research presentations tell a dynamic story and integrate appealing elements to illustrate key themes.
At The Farnsworth Group, our goal is to make sure insights are clear and comprehensible for the variety of stakeholders in your organization and across departments. We also add videos, recordings, word-for-word transcripts, and other visual elements to capture the voice of the customer behind the findings and conclusions.
4. Using Secondary Data as a Valuable Supplement
We do believe that AI can be used to help deliver additional value to clients. One way AI can potentially assist market research professionals is by mining existing data to look for any information and insights that can augment findings in a new or original survey. We are actively investigating the reliability of this approach, usually manual processes for quality assurance during this testing.
While it can’t replace the benefits of primary research, it can supplement that work. For example, any secondary data and internal client data that’s available can be used to cross-check additional findings from primary research.
Selecting the Right Vendor for Market Research
Companies in various industries utilize market research to make important and often costly decisions about product development, increasing revenue, pursuing new markets and acquisitions, and furthering brand recognition. You don’t want to make those decisions lightly or based on faulty research.
The best way to help reduce the risk of making questionable corporate decisions is to partner with a market research company that is constantly evolving and using up-to-date tools to improve the authenticity, reliability, and value of data and insights.
That’s what you can expect with our team at The Farnsworth Group, as we respond to changes in the market research industry, employing accurate data collection practices and taking measures to prevent bad actors from distorting findings.