Sensory testing key element in selection of new produce varieties

Many years of development precede the launch of a fruit or vegetable variety in the market. An important part of the developmental stage is taste-testing. “My intent is very simple. I want to bring tasty fruits and vegetables to the population,” says Chow-Ming Lee, Sensory Scientist with Monsanto’s Vegetable Seeds Division. “Simultaneously, my goal is to create less waste as good taste equals less waste,” he added. Melons, tomatoes, watermelons, lettuce, broccoli and cucumbers are all part of Lee’s sensory-testing portfolio. 

He calls himself a hybrid as he is involved with new produce varieties from the Research & Development phase all the way to commercialization. He meets with growers as well as retailers. “The grower side is important as growers need yields and a certain volume to make money. At the same time, the consumer’s need for flavor needs to be met. Bringing varieties to market that consumers appreciate is a balance between growers and retailers,” shared Lee.

Taste-testing is a key element in the variety-selection process. “We work with a flavor-modeling program that is used to understand the components that drive consumer satisfaction in any produce variety,” said Lee. “What do they want in a tomato or melon?” In the early stages of variety development, the same set of samples is measured with instruments. “Once we understand the human liking for a certain produce variety, we have what we call algorithm and start verifying it,” continued Lee. “It’s the algorithm to consumer satisfaction and with instruments we try to mimic the human perception. Models have been developed and tested and we know what consumers prefer and are able to predict their liking.” Once the algorithm has been identified, instruments are able to measure the different varieties. Instruments start with measuring thousands of varieties and based on aspects such as flavor results, yield, plant health and location adaptability, the number of varieties is narrowed down to hundreds, to tens and finally to ten or five varieties that will be launched in the market. 

In the beginning phase, instruments are of key importance as they don’t get tired or saturated. They analyze extremely large numbers of varieties and rule them in or out. Instruments are able to measure the texture, the coloring and even the smell component. They enable identification of produce varieties that are particularly favorable to the consumer or unfavorable. Some components will be identified as positive, others as negative. If the breeder has a way to breed out the negative component, it results in improving the liking of the variety. 

Taste-testing with humans is done in the final stages. “When we are down to 10 varieties, we bring in humans as they improve the chance of success. Let’s take brix as an example: it is an easy and common measure, but it alone doesn’t predict consumer liking. It is a component of consumer liking, but not the only component,” Lee said. “Models are models and are not 100 percent accurate.” The reason to bring in humans in the final stage is that taste-testing with people is costly and in addition, humans only taste about six samples per day. If they get presented more than six samples, the response will be poor. Lee mentioned it is interesting how some people have a much better sense for taste than others. Literature suggests that about 40 percent of the consumers in India cannot perceive bitterness easily. When Lee ran a study in India on cucumber bitterness, he had to segment the people that are sensitive to bitterness from the ones that aren’t and kept the results of the bitterness-sensitive consumers.

Because sense for taste differs around the globe, Monsanto Vegetable Seeds (MVS) works with different algorithm models. “People prefer one model for the whole world, but in some markets taste preferences are completely different,” Lee shared. Just as an example, MVS is using models for the US as well as the Netherlands. “We crisscrossed and had the model from the US trying to predict the Dutch and the model from the Dutch trying to predict the taste of Americans. It was a disaster.” In this case, the US population didn’t pick up on the smell of a certain produce variety whereas the Europeans did and rejected the product. 

Contact:

Chow-Ming Lee

Tel: +1 530-669-6049