Preference and Perception

Perception and Preference: The Key Ingredients to a Successful Product

If you don’t understand the relationship between perception and preference, you’re hindering your ability to make products your target audience will enjoy.

An average of 30,000 new CPG products are launched every year. It’s no secret that around 70% of those products fail. But what separates those failures from the successes you see on convenience store shelves? To put it simply: a product consumers like. People keep buying products they like, and they don’t re-buy ones they don’t like.

To create food and drinks your target demographic will love, you have to understand the way they experience flavor. Most product taste testing only looks at consumer preference. It doesn’t collect information on the variety of flavors different consumers perceive. It ignores the fact that we all taste differently.

If you don’t understand the relationship between perception and preference, you’re hindering your ability to make products your target audience will enjoy.

Perception vs. Preference

Perception is what you taste. In CPG food and beverages, perception is the flavors, aromas, and textures the consumer notices. It's the consumers' sensory experience of the product. Preference is how much they like it.

Your preference is based on the features you notice. For example, some people love salt and vinegar chips because they perceive the tang of the vinegar. You might:

  • Dislike the chips because you perceive the same thing and don't like the flavor. Same perception but a different preference.
  • Like the chips because you perceive something different — you identify a sweetness that appeals to you. A different perception but the same preference.
  • Or maybe you dislike the chips because you perceive something different — perhaps you perceive the vinegar as extremely sour. Different perception and a different preference.

In traditional new product development testing, professional tasters are used for the perception part. They identify what flavors are in a product. Then, testers ask the consumers to rate the flavors that were picked out by the professional tasters.

Expert tasters are referenced trained. They do calibration training where they learn the “standard” of what a flavor is. For example, let’s say Smucker’s blackberry jam is the standard flavor for blackberry. The professional tasters would evaluate the blackberry-ness of all products in reference to that specific blackberry jam flavor.

The problem is that an expert might not identify a product as having a blackberry flavor because it doesn’t match the Smucker’s jam reference. But a consumer might say, “This tastes like blackberry,” because their standard is different. That means the results of a typical product tasting are missing key information — the flavors consumers notice when they taste a product.

Why Perception and Preference Matter

If you deal with perception and preference separately, you’re cutting out the voice of the consumer — your product development is blind. In traditional NPD testing, you don’t ask consumers what they perceive, only what they prefer. So you can’t accurately identify why they like or dislike a product.

Ryan Ahn, VP of Innovation and Application at Gastrograph AI, emphasizes the importance of gathering data based on consumers’ perceptions as well as their preferences: “I would say you cannot truly understand preferences unless you understand perception. You can measure preferences without perception, but you can't understand it unless you know the experience the consumer is having.”

Let’s say you’re testing out a new coffee. The expert tasters define the flavors in the coffee: bitter, roasted, caramel, and chocolate. Then you ask consumers to rate how much they like the flavors identified by the professionals. Unfortunately, one consumer tastes a horrible cardboard flavor. They might explain this flavor with one that’s listed — maybe they translate “cardboard” to “bitterness” and tell you they don’t like the bitterness of the product. Or, they don’t communicate it at all, so you have no idea that your product tastes like cardboard to some consumers.

fallon gif

A Cardboard Mocha Frappuccino

This could also happen the other way around. Maybe the consumer tastes a dark fruit note they love. Because you’re not asking consumers about the flavors they perceive, you’d have no idea about this dark fruit note. You might end up removing this flavor because you didn’t realize it was a key reason why consumers enjoyed the product.

“If you don't ask [what consumers taste], you're not capturing something that's predictive of preference.” — Ryan Ahn

If you look at what different consumers perceive, you can understand why they like or dislike certain products. Then you can improve your products or create new ones to target those preferences. In this case, you might notice consumers are tasting something you hadn’t planned. Now you know you need to fix it before the drink launches. Or maybe the consumers mentioned the dark fruit notes and you can increase them to make the product more likable.

An understanding of how consumers perceive flavor is also useful for improving packaging and branding. If you include flavors you know consumers will notice on the packaging, you can better set consumer expectations. You'll be able to cue them into the attributes they're most likely to enjoy. It also allows you to pick out flavors to differentiate your product from other competitors on the market. You’ll be putting your consumers and the way they experience flavor at the heart of your decision-making.

How Gastrograph AI Treats Perception and Preference

Gastrograph AI conducts tastings to find out which flavors different demographics notice and how much they like the flavors they taste. With this information, our AI models learn to identify the impact perception has on preference. Once the models have been trained on enough data, they understand how a demographic perceives the flavors they taste and which flavor combinations they prefer. Then we can predict how people will respond to novel flavor combinations.

It’s similar to how you can go into a store and predict which flowers your mother would choose. Years of living together mean you understand how she perceives different flowers, and you know what she likes. She’ll steer clear of carnations (because she sees them as old-fashioned and boring) and probably go for lilies (elegant and exotic) or tulips (colorful and happy). Now imagine the store starts selling a new type of flower. You don't know what your mom will think of it because she hasn’t seen it yet. But you can make a pretty good guess because you know her so well.

The AI system makes predictions about what consumers will like. And the system’s predictions are a lot more reliable than your guess. The predictions are based on a lot more data — Gastrograph AI has the world's largest sensory database — and it’s not going to make human errors. This data is used to train an AI model to spot patterns and make predictions. We collect tasting information from consumers all around the world. The consumers taste a wide set of products so the model can learn how they respond to diverse flavor, aroma, and texture combinations.

Lots of factors influence the way people perceive flavor. When someone tastes for the system, we collect their demographic and environmental information like age, location, and smoking habits. With this data, the models learn how various factors impact perception.

Request a Demo

undefined-Apr-07-2022-05-56-34-68-PMTasters fill out information that might affect how they experience flavor.



All of the data comes together to create a robust prediction algorithm that can be used across demographics. Let’s say you want to test out a new coffee in Paris, different from the cardboard-and-dark-fruit one before. Gastrograph AI has collected data from French consumers living in Paris, so we know how they perceive different flavors, but they haven’t tasted this coffee. This coffee was profiled in New York.

First, you need to make sure you have an accurate understanding of how a French consumer would perceive this coffee. Perhaps they’ll notice different flavors than the people who tasted it before. Using the AI models, we can translate the flavors the New Yorkers recognized to discover what flavors Parisians would taste.

Then, we can take the new coffee and run it through the system. The AI models have processed a lot of information about the way Parisians perceive and prefer flavors, so the system knows them pretty well. The models can predict how much the French consumers would like the coffee and what they would like about it. We’ll be able to identify how you can improve the product to make it more likable to a Parisian consumer. That sets your new coffee offering up for success when you launch it.

Let Consumer Preference Lead Product Development

Instead of developing a product and then bringing consumers in afterward, use consumers as your North Star. Let their preferences drive your product development from the very beginning. Find out what flavor, texture, and aroma combinations your target demographic will enjoy, then create products just for them. Request a demo of the Gastrograph AI system to help you perfectly align your product to your consumer.

Similar posts

Stay up to date with Gastrograph AI

Be the first to know about company updates and industry-related news, straight from our internal subject matter experts.