As with all things AI, there is a lot of hype about its impact on market research. But how could you take advantage of it in your business to support your brand and marketing efforts? In this post, we’re going to look into how AI could help analyse your existing brand/marketing research, how it can tell if your brand is distinctive, and how you could use it to alongside more traditional consumer research methods.
Using AI to do the hard work
When market research or brand data is expressed in cold, hard numbers, it can be easy(ish) to interpret. Quantitative data can give us plenty of insights about what customers do, where they do it, and when they do it. It can even tell us, to a certain extent, who they are too. But when it comes to quantitative’s tricky qualitative cousin, the job of interpretation is a bit, well, open-ended. That’s because, when consumers express their views in their own words, the data isn’t so black and white: it’s full of nuances, complexity and emotion. It’s human. Processing and interpreting these open-end responses is complicated and therefore time-consuming and expensive. So much so, that many companies will avoid qualitative data if possible.
But avoiding qualitative data is a missed opportunity. Qualitative data can tell you ‘why’ your customers behave the way they do; why they make certain decisions. And when you know the ‘why’, you have the insight you need to develop relevant, on the mark advertising campaigns that truly connect with your brand’s audience. The ‘why’ allows you to tailor your product advertising and messaging to align more closely with your customers’ preferences and needs.
One of AI’s biggest selling points is it can do the time-consuming work for you. Large Language Models (LLMs) can take vast amounts of unstructured data and sort, analyse and provide insights from it in the blink of an eye.
What makes AI even more useful for qualitative data is that it can process natural language. All those honest and emotional responses from your market research can be processed and turned into meaningful insights. Natural Language Processing (NLP) can provide sentiment analysis, emotion detection, and keyword extraction, broken down across numerous audience segments.
In complex and fast-changing modern markets, the companies that will thrive are those who can build strong, meaningful relationships with their customers. Using the analytical power of AI in combination with open-ended human responses can help you understand their motivations, preferences, and needs, quicker than ever before.
Using AI to make your brand distinct
Part of what market research can tell you is how identifiable your brand is. Consumers might be asked what characteristics they recognise from a selection of brands.
But what if you could get a visualisation of how distinct and defined your brand is? With image-generating AI like Dalle-E and Midjourney, you can ask for a generated advert of your brand. If you get a selection of images that are clearly your brand (if a little wonky and, you know, AI-like), then congrats: you have a distinctive brand. But if the generated images look generic or look nothing like how you think your brand looks like, you know something’s wrong. Marketing Week tried this with Guinness and Dall-E, with positive results for the golden harp.
Brands like Heinz have already put a creative spin on this tech to demonstrate how iconic their ketchup bottles are, and how synonymous Heinz is with ketchup.
Using AI to synthesise your audience
Whilst the complexities of processing large amounts of open-ended data is one problem, another is collecting the responses in the first place. It can take weeks or months to collect enough responses to produce an insightful and meaningful report. This is a particular problem in B2B marketing as niche decision-making audiences are notoriously hard to find (and incentivise).
So, how do you collect those responses for a fraction of the time and cost? What if you just, kind of, created it yourself? In simplified terms, that is what’s happening with synthetic data.
Using generative AI and LLMs, companies like Evidenza can produce human-like responses from “AI copies of your customers.” That small, tricky-to-pin-down group of decision-makers in your particular niche could now be at your fingertips, ready to give you human-like responses at the tap of a button.
Human-like is the key term. Evidenza claims that its synthetic data is 88% as accurate as traditional research results (some results with the consulting giant EY put that figure as high as 95%). But how useful is data that is 88% accurate? When making a relatively small decision, perhaps that’s fine. But when you’re making decisions worth millions, small discrepancies in accuracy can result in huge losses.
As one marketer aptly put it: “cats have 90% of the same DNA as humans, but I wouldn’t ask a cat what a CFO thinks.”
Already, there are some red flags being waved with regards to synthetic data, from its legality to biases. Take a look at our opinion piece to understand whether it could be an option for you or not.
A bit of both
Blended models, where traditional, human-sourced responses are supplemented by synthetic samples might be the best and most accurate way forward. For particularly niche audiences in the B2B world, synthetic data could be used to boost sample sizes.
On the other hand, utilising AI’s ability to sort, analyse and provide insights from large amounts of open-ended responses, complemented by traditional research seems like a winning solution.
Here at saintnicks, we’re exploring different offers and methods with our research partners. If you’re interested in market research – with or without the complement of AI – get in touch with our Strategy Director, Mark to talk through how we can help.