Is flawed market research worse than no research at all? False conclusions can amplify biases and lead to huge amounts of time and money wasted, and lead to reputation challenges that have lasting consequences. Of the 30,000 new physical products that are launched each year, 95% of them fail. Why are so many brands failing to meet the needs of consumers?
This was the question keeping Frank Pica and Sarah Sanders, co-founders of Native AI, up at night. As brand technology industry veterans, they recognize the many uphill battles that brands face every day. Rising competition, along with higher costs of raw goods and shipping, have squeezed profit margins. Brands need to find ways to work more efficiently, and optimize every dollar spent. Unfortunately, this usually leads to shortcuts that exacerbate blind spots.
As any market research professional can attest, there are many types of bias that plague qualitative research methods. For example, sponsorship bias shows up when consumers answer survey questions in a way that is more flattering to the brand that is sponsoring the survey. Confirmation and culture bias occur when humans interpret open-ended responses through the lens of their own personal experience and cherry-pick answers that support their hypotheses.
Frank and Sarah knew that machine learning was the only way that brands would be able to cut through the noise of the internet. The company they founded, Native AI, is a no-code generative AI solution that has the power to dynamically ingest a limitless number of open-ended text sources from across the internet and synthesize them into digestible insights. Each brand has a unique language model trained specifically from the brand’s own consumer review and feedback data, so brands can be confident that they are engaging with a voice that is truly representative of their customer base.
As a member of the LGBTQ+ community, Sarah recognizes how underrepresented voices are often lost or drowned out. She explains, “less than 10% of consumers in North America participate in surveys or panels, and this group is self-selecting. It is often not representative of the total consumer population, nor is it correlated with those who have the most interest in the brand.” At its core, Native AI is about helping researchers address human bias as much as possible.
Rather than manually comb through the thousands of comments, product reviews, and open text survey responses, resource-constrained brands have been forced to resort to focus groups. This is an unfair disadvantage. Frank wonders, “why shouldn’t every brand and researcher have equal access to information? History tends to repeat itself. Smaller companies that have less access to technology will be at a competitive disadvantage. So accessibility is a big core value of the company. If we can make this technology easy to use for all, then it evens the playing field.”
What’s next for Native AI? With major advances in generative AI, Large Language Models (LLMs) can not only generate descriptive summaries of data, but predictive ones as well. Native AI calls this feature Digital Twins. A Digital Twin is a virtual clone of a business’s customer that can learn from all the digital feedback and purchasing habits of various consumer segments. Using this huge quantity of signals, Native AI can predict how various consumer types will respond to hypothetical scenarios.
Brands and market research firms that use Native AI report seeing significant improvement to their market research budgets and timelines, while simultaneously having increased confidence in the depth of answers. Frank says, “with so much data readily available both online and offline, AI has an enormous impact because you can execute research in a fraction of the time and without the need to recruit humans.” Frank is the first to admit, however, that bias can never be completely eliminated. But for the team at Native AI, that is the north star.