Our artificial intelligence consumer insights tool is helping inform ahead-of-trend product development for food and beverage brands
KerryDigest Fast Facts:
- In the increasingly competitive food and beverage industry, tracking trends in real time can give brands an advantage.
- The Kerry Trendspotter™ artificial intelligence tool, developed in partnership with IBM Watson, captures consumer research in real time by tracking social media channels.
- Developing products with the aid of Kerry Trendspotter can speed up innovation and better predict market success.
Get KerryDigest articles delivered to your inbox
KerryDigest Full Scoop:
When a large snack manufacturer asked us for help ideating its next big hit, the team knew Kerry Trendspotter™ data would ensure the product hit all the consumer sweet spots.
By “listening” to the social media conversations of tens of thousands of food and beverage micro-influencers, the Trendspotter artificial intelligence (AI) tool, which was developed by Kerry in partnership with IBM Watson, is able to sort through the chatter and identify emerging flavour, format and product trends long before they register in typical consumer research.
This access to consumer sentiment in real time, without the limitations of surveys and focus groups, gives our researchers an unfiltered view of the spread of new food and beverage trends. To date, the product has been used on dozens of projects for brands around the world, ranging from monitoring holiday LTO drink trends to guiding brands on the most in-demand functional health ingredients during COVID-19.
Utilising AI to eliminate gaps in market research
Kerry Trendspotter is built on the hypothesis that conversations on social media provides a lens into the ever-changing consumer point-of-view.
“Consumers are increasingly reliant on their social media feed when searching for products and new experiences,” says Avinash Lal, Market Research and Consumer Insights Director for Kerry Asia Pacific, Middle East and Africa, who helped develop Trendspotter. “With Trendspotter, we gain near perfect view on what is happening in the consumer landscape in real time and have the ability to filter the results based on how likely a food and beverage trend is going to result in widespread consumer demand.”
This AI-enhanced view allows our researchers to create data-backed predictions about future food and beverage trends, which gives our customers insights to build consumer-preferred attributes and differentiation into their products. This low-touch data collection has been especially critical during COVID-19, with focus groups largely off the table and consumer preferences seemingly changing overnight.
Adding accuracy and speed to innovation
With trends shifting more quickly than ever, launching the right product at the right time is key to market success, and something only a fraction of products achieve. (Nielsen data suggests as much as 85% of new products fail.)
By simplifying data collection, Trendspotter accelerates the production cycle and pinpoints consumer needs, increasing the chance a product launch will score big amongst consumers.
“Before Trendspotter, our work on new product development relied on traditional research, which is often lengthy and costly and can still fail to accurately predict consumer needs and trends,” says Lal. “Now that we have real-time access to insights from millions of posts from Instagram micro influencers, we have insight into trends that are not yet mainstream.”
Because consumer views are being shaped by the same social media posts Kerry Trendspotter examines, our chefs and food scientists are able to start experimenting with new experiences and products before consumers notice a new trend is forming. By the time an emerging flavour, ingredient or format is officially a trend, customers with access to Kerry Trendspotter insights could have a product ready for launch.
Tracking the evolution of Trendspotter
Initially, Kerry Trendspotter was used to identify already trending flavors, ingredients and products—to get a real-time pulse on widespread consumer preference. But increasingly, food brands are looking at the tool and its predictive scores to help inform their innovation pipeline for the next two to three years.
As is implied by its classification as an AI tool, Trendspotter is learning as it operates. With inputs from the research team, including direction on which engagements to save or ignore, the initial build evolved into a tool that was able to independently spot trends and track their genesis. That knowledge, plus improved natural language processing skills, now allows Trendspotter to catalogue early trend markers, such as enthusiastic language about unusual flavours and ingredients. When these mentions hit a certain score, they can be segmented based on how close the trend is, to going mainstream.
For brands looking to launch a product in the next 9 to 12 months, the key is to focus on trends that are just entering consumer consciousness. When working on a 12- to 24-month timeline, Trendspotter is looking for small but significant early interest.
AI as part of an integrative approach
Despite its one-of-a-kind functionality, Trendspotter does not operate or work in isolation. New product development still begins with having an understanding of a customer’s business, brand, objectives and challenges and analysing syndicated and proprietary consumer research to decide on priorities and focus.
Only then does Trendspotter come into the picture, says Lal. After the trending ingredients, flavours and products with the highest chance of breakout are identified, our culinary and research and development teams build new concepts with a high probability of success. These are tested and the winning idea is reworked for production at scale.
When working with the snack company from Vietnam, Trendspotter helped to shorten the insight collection process from 3 months to 1 week, setting a new benchmark for speed with the customer and halving the typical production cycle. The research highlighted an emerging interest in a new western flavour direction, which was gaining relevance amongst local consumers. Trendspotter further identified complementary local flavours, which the team used to inspire 6 concepts to present to the customer.
A unique flavour blend was ultimately chosen, and when the product launched it was the area’s first instance of a snack with that seasoning to enter the market. The product was received with great fanfare, and within six months of launch, the customer doubled its production. They also contacted us to begin work on a further line extension as well as a brand new product concept, adding a significant vote of confidence to the predictive powers of Trendspotter as a way to quickly uncover and act on new trends.
This article first appeared in the March edition of Food and Drink Technology Magazine.