In this two-part series, we explore how flavor may be understood, perceived, and valued in the future, based on insights gained from speaking with industry experts. Part I is more focused on the food system, while Part II delves into new flavor profiling technologies employing big data and AI.

The fact that artificial intelligence and data-driven methods are being harnessed to upgrade kitchens, create new products, and generate supply chain solutions is already old news in the food industry. Innovations in digital flavor profiling, however, have yet to gain widespread traction. Hopefully, that will soon change, as companies like Gastrograph and FlavorWiki are seeking to alter how companies – and by extension, consumers – view flavor.

We at the Future Food Institute spoke to Jason Cohen, founder of Gastrograph, and Daniel Protz, founder of FlavorWiki, to gain more insight into how these digital profiling pioneers are using analytics to reshape the flavor profiling landscape.

Experiencing flavors

We asked Cohen how the role of flavor or taste might change as food becomes increasingly experiential. Cohen thinks the role of flavor itself may not undergo a transformation; rather, the shift lies in the role of data changing how products are developed. “People have always wanted good tasting food, and will continue to purchase the products that taste best to them.” However, as novel techniques are “making it possible to predict perception and preference for the first time,” product developers need to become progressively more aware of how they can take data and predictions into consideration to develop competitive products.

In the past, companies could develop “generally acceptable products,” but as the number of companies and products grow, on both local and global levels, increasingly niche products are required to remain competitive. Cohen firmly believes that companies failing to use modern techniques and data to predict the changing consumer preference of each consumer cohort will be at a disadvantage. Due to increased competition in both old and new markets, companies are reinvesting in the competitive attributes that will bring consumers back to their brands – which, in Food and Beverage (F&B), is flavor.

Personalizing flavors 

Technological advancements extend to personalization. As flavor may be experienced differently by different people, it’s only expected that the flavor becomes something more individualized. Sensory panels that most companies currently use to taste-test products are unrepresentative of general market preference. Gastrograph uses AI and predictive models to identify how different demographics and regions experience and enjoy different flavor profiles. They then utilize that information to optimize new product development, product adaptation, and portfolio management services to companies, taking targeted and cognitive marketing into account to ensure the right products are developed for target consumers.

FlavorWiki, too, is developing flavor-mapping technology that quantifies individual taste perception and preference. Food retailers and producers can then use this data to personalize product development, improve product innovation, and deliver a more engaging food experience. Their founder and CEO, Daniel Protz, believes consumer research is ripe to undergo a transformation.

The goal of consumer research to determine what flavors, textures, and fragrances will be popular with a target market. Companies typically have trained tasters to give feedback on products. It is generally very expensive to conduct this type of research; so much so that the large majority of consumer research in sensory science is done by a few large conglomerates like Nestle and Unilever.

Increasing accessibility

Protz believes “somebody should figure out a way to make this type of research more accessible to smaller companies, to make it faster, to make it more agile,” so that new companies creating products like meat or dairy replacements can develop foods that are both innovative and liked by consumers. FlavorWiki’s methodology enables the profiling of food products by untrained consumers, using an algorithm and data capture methodology developed by a Princeton statistician. A person can taste a product and use the application to profile it, allowing FlavorWiki to obtain profiles from anywhere in the world.

Their profiling technique is more robust than existing consumer research as they can collect a lot of data at a lower cost. They gather information like time of day, age, demographics, where you consumed the product, and what you consumed before the product as well. Outputs may include: what is the level of sweetness, or bitterness, et cetera, in a juice product? Is that product accepted by the group of people the company is testing it with? How much sweeter does it need to be to be accepted? Is that statistically relevant? What is the difference between products A and B, and which do people prefer, and why?

A “flavor profile” can be created for each person, showing their preferences for different flavors. People’s taste preferences can be grouped into taste archetypes, or personas; the archetypes of people who live and work in East Asia will be different from those who grew up in the American midwest. Archetypes can change depending on age and exposure to new foods, and any one person can fall into different archetypes for different types of foods.

In the past, food companies have mass-produced a single product and marketed that product to all types of consumers. With taste archetype data, companies can better target particular archetypes, and consumers can be matched with food products based on their archetypes. It will be easier to tailor products to specific geographic locations. Companies can still mass-produce, albeit on a slightly smaller scale, catering to specific archetypes of a desired market.

Ultimately, FlavorWiki hopes to construct a “taster community platform” where groups of people with certain profiles are given both new products to try before they are released and incentives to provide feedback on products currently on the market. Likely first adopters may be sent food based on on criteria like “vegan” or “dairy free” to review.

The resulting scenario

The future we imagine is one in which digital flavor profiling technologies empower consumers as well as producers, providing us a greater voice in the development of food products. Cohen notes that successful F&B products can have long-term impacts on the future flavors of other goods. If product development becomes more consumer-centric, we may see changes in the way food trends disseminate. In addition, profiling technologies seem to enable consumers to be better arbiters of flavor. Users claim FlavorWiki causes them to think more about what they’re eating; as people become better at recognizing flavor notes, they can gain more appreciation for the food they consume. Due to heightened awareness of flavor on the consumer side, and awareness of predictive technologies and big data on the producer side, flavor may be viewed as an increasingly personalized experience to tap into.

SHARE
Previous articleWill BSH’s PAI Usher In The Era of the Kitchen Projection Interface?
Next articleTarget Cancels Indiegogo For Smart Home Replenishment Platform
Chiara is fascinated by food as a means to impact bodies, minds, and environment. She has studied international business in three different countries, and is an alumni of the Food Innovation Program and US Director at the Future Food Institute. Based in California, she is also a Research Scholar at Food Science and Technology at UC Davis, working on building the first comprehensive Internet of Food to enable food care through food systems semantics. She is a Researcher at the Barilla Center Food Nutrition Foundation, a Research Affiliate at Institute For The Future, Board Member at Maker Faire and a Global Shaper at the World Economic Forum. She is passionate about social entrepreneurship and impact investing, and aims to leave her mark on society.

1 COMMENT

  1. I’m curious about how personalization, taken far enough, would be balanced with novelty. With people who are into food, part of the joy of eating diversely (sometimes while travelling, sometimes by experimenting locally) is being exposed to flavors you didn’t know you liked. (Thinking specifically of KitKats in Japan, but there’s plenty of other examples.)

    In theory, of course, an excellent algorithm would take all of this into account and throw in some novelty as it customizes your experience, but algorithmic quality, coverage and accountability will be issues to overcome.

    Depending on how broadly they’re used, these algorithms can, at some point, define our eating experience! Compared to issues with algorithms determining jail sentences, solidifying political divisions, and many of their other current challenges, this particular topic seems relatively unimportant. At the same time, given that food is much more than just sustenance for billions of people in the world, it’s not a discussion that should be ignored.

    Algorithms can create silos, and it would be unfortunate if, instead of opening up the world of food, flavor algorithms unexpectedly restricted it.

Leave a Reply