The Near-Term Impact of AI on Market Research

Would we even be researchers if we weren’t thinking deeply about how AI is going to impact some of the work we do over the next few years? It is impossible to ignore the avalanche of AI news that hits our inboxes every morning. Although we’re still in the early stages of AI product development, it’s undeniable that AI will soon take over key aspects of research and marketing. Lately, I’ve been thinking a lot about how AI will integrate into segmentation, and I’ve already seen early AI-driven segmentation solutions. This is just the beginning.

Existing AI Segmentation Applications

AI’s capacity to quickly process massive datasets makes it easy to envision scenarios where AI solves real problems for both marketing and core research functions. For example, AI tools can efficiently parse customer service channels to identify key consumer questions, needs, and pain points.

This, in and of itself, is considered a segmentation by some (needs-based segmentation based on real-time chat requests, anyone?).

While there is value in such tools, we need to ensure they don’t replace the foundational segmentation processes that provide a much more holistic view of the category and address strategic business questions. 

Where AI Segmentation Falls Short

When considering AI’s role in segmentations, let’s focus on three key questions for an actionable consumer segmentation:

• What is the total addressable market (TAM) for my category?
• Which customers am I trying to win?
• Which strategies are likely to win those customers?

To answer these questions effectively, our data must meet certain criteria, and this is going to be a pivotal consideration in the traditional vs. AI debate over the coming years. 

To get a clear picture of the TAM and identify key customers to target, we need to have a holistic view of the category. However, most AI-accessible datasets will only show a partial category view, often oversampling certain audiences and biasing the final solution. The current demand for quick, cost-efficient research solutions will drive some competitors to accept these data biases, introducing and amplifying a shared market distortion across areas of the category. To counteract this, we need to focus on the key needs of a foundational segmentation. The dataset must meet the following criteria:

1. Represent the entire market accurately
2. Contain detailed information about potential consumers to allow us to model why they are making the choices they are making
3. Provide the data needed to model how different strategies will be evaluated by potential customers

No amount of complex algorithms can compensate for a dataset that falls short in three areas. This will be one of the biggest challenges we face over the next few years. The allure of being an early adopter of AI segmentation among your peers will lead some competitors to look past critical flaws in method.

Don’t Sacrifice Quality & Utility for Novelty & Efficiency

The last rule – framework quality – is one that merits an expanded discussion. A good segmentation framework creates a vital, shared understanding of the category and consumers (or businesses in B2B contexts), enabling the organization to collaborate on strategies to attract and convert consumers. When done correctly, segmentation allows marketing, product, sales, innovation, and R&D teams to work from a common understanding of the category. To accomplish this, the framework must be reliable, repeatable, and easy to communicate. While AI solutions may eventually meet these criteria, they are not reliable enough to be utilized for research as foundational as segmentation  

Research Data Lessons from the Past

The current landscape is reminiscent of the 2010 – 2015 period when the proliferation of internal datasets shifted the inputs driving brand strategies. The rapid adoption of brand-created, internal datasets was so quick that many researchers became focused exclusively on on-platform work. The allure of the huge data sets caused researchers and marketers to take their focus off the category as a whole and they missed big opportunities as a result.  While a valuable resource, this singular focus on existing customers neglected the truism that growth happens by converting non-users into users. Overemphasizing current users at the expense of non-users resulted in some very big, preventable strategic missteps.

How Murphy Research will Embrace AI in Segmentation

At Murphy Research, we will undoubtedly experiment with AI-driven segmentation over the next few years. We are already exploring a few use cases where AI can streamline the process. As we explore these new methods, we will keep in mind a few key principles:

1. The dataset must meet our three data quality criteria outlined above.
2. The model must be clearly repeatable in different markets and easily understandable.
3. Most importantly, the priority must always be the quality of the framework, even if that means using a simpler method.

Conclusion

As AI continues to evolve, its impact on market research, particularly in segmentation, will be significant. While AI offers exciting possibilities, it’s crucial to approach its adoption with caution, ensuring that the quality and integrity of research remain paramount. At Murphy Research, we’re committed to exploring AI’s potential while upholding the rigorous standards that our clients trust.

Interested in learning more about how AI can complement your market research strategies? Contact us today to discuss how we can help you navigate the evolving landscape of AI-driven research and make sure your business stays ahead of the curve.