Givaudan Used Our Technology in Developing New Flavours

The challenge: Givaudan was trying to figure out which flavors were liked by the most people

Givaudan has a super efficient way to let panelists evaluate flavors. However, in sensory science if you mixed the same ingredients in the same proportions, but different volumes - the liking results can differ wildly! E.g. I may love a little bit of vanilla, but with twice as much I might hate it.

So they devised the following experiment:

  • They took seven ingredients to create a target flavor composition.
  • They synthesized a total of 40 flavors varying ingredient volumes.
  • They created a focus group of 70 people to evaluate the flavors.
  • The subjects would rate them 1 to 9.
  • The total data set contained 70x40 evaluations of eight columns - seven ingredients and the main Key Performance Indicator - the liking score.

This method of collecting data was extremely time consuming and expensive. Here are the problems with traditional testing:

  • It's expensive to find people and bring them to the lab.
  • 3+ weeks to find people.
  • Thousands of dollars to accommodate the 70+ test subjects.
  • A normal human can only evaluate about 15 flavors before people lose the ability to distinguish between them.
  • The subjects are from the target population of consumers, but they are not experts, so there is a lot of “noise” in the data.
  • Different people have wildly varying “likes”.

Every flavor company needs to repeatedly do these experiment to gather data on hundreds of flavor compositions. As you can imagine, the time and expense to do this testing is enormous.

Even after all this testing, the company was looking for better ways to outsmart the competition and push new flavors to market faster.

This is where the technology behind DataStories comes in.

Givaudan wanted to see if there was a faster way to devise popular flavors. The process was very easy:

  • We sign a non-disclosure agreement.
  • You hand us a spreadsheet with experimental data.
  • You decide what are the main things you want to know from the data:
    1. "Which ingredients drive liking?"
    2. "Which optimal combination of ingredients will make a flavor that many people like a lot?"

DataStories delivers back:

  • A full report.
  • An interactive presentation.
  • Robust predictive models for each subject.
  • Segments of people driven by the same ingredients (in the same direction).
  • Groups of people segmented by propensity to like things ("easy to please" and the groups who are "hard to please")

What can the company do with this information:

Instead of doing 70-person focus groups every week, and not knowing exactly how to optimally use ALL the data, they can:

BEFORE

AFTER

Spend 6 months of the year conducting research in focus groups. Spend only 2 weeks conducting research in focus groups. (92% time savings)
Spend several months of the year figuring out which ingredients drive "liking" Figure out which ingredients drive "liking" by the next morning
Search for 40-80 people to evaluate a new flavor. Plug the flavor composition into the "predictive models" to get the scores with confidence intervals. Zero test subjects required.
Evaluate 13-16 flavors per session per person. Evaluate 1,000,000,000+ flavors per session.

How we can help YOU:

So if you are a flavorist, don’t worry that the tools you currently have cannot do this type of advanced modelling.

  • We can apply this advanced advanced modelling to your problem.
  • We will handle all the data science, so you don’t have to.
  • We will find the correlations hiding inside your data.
  • We can predict flavors for 20x less money, and 6x less time.
  • Give us a call to see what we can do for you: +32 47 638 84 97.

Want DataStories to bring you similar results?

Request a free online demo and see us in action!

Resources:

  • Katya Vladislavleva, Kalyan Veeramachaneni, Matt Burland, Jason Parcon, Una-May O'Reilly, – Knowledge mining with genetic programming methods for variable selection in flavor design. In Juergen Branke et al. (Editors),GECCO'2010: Proceedings of the 12th annual conference on Genetic and evolutionary computation, pages 941-948, Portland, Oregon, USA, 2010.
  • Kalyan Veeramachaneni, Katya Vladislavleva, Matt Burland, Jason Parcon, Una-May O'Reilly. – Evolutionary Optimization of Flavors. In Juergen Branke et al. (Editors), GECCO'2010: Proceedings of the 12th annual conference on Genetic and evolutionary computation, pages 1291-1298, Portland, Oregon, USA, 2010.
  • Kalyan Veeramachaneni, Ekaterina Vladislavleva, Una-May O’Reilly – Knowledge mining sensory evaluation data: genetic programming, statistical techniques, and swarm optimization Journal on Genetic Programming and Evolvable Machines, March 2012, Volume 13, Issue 1, p.103-133
  • Katya Vladislavleva, Kalyan Veeramachaneni, Una-May O'Reilly – Learning a Lot from Only a Little: Genetic Programming for PanelSegmentation on Sparse Sensory Evaluation Data. In A.I.Esparcia-Alcazar et al. (Editors), Proceedings of the 13th European Conference on Genetic Programming, EuroGP2010, Lecture Notes on Computer Science, Volume 6021, pages 244-255, Istanbul, 2010, Springer.