When a product team has two or three recipes and one spot on the shelf, opinions in the test kitchen are not enough. The people who buy the category have to taste them, blind, under the same conditions and in a fair order. A central location taste test does that in a day or two, in a focus facility with tasting booths, a prep kitchen and staff who serve every sample the same way.
The method
Category buyers come to the facility in small sessions. Each person tastes every recipe, one at a time, from cups marked only with a three-digit code. The order rotates from person to person, so every recipe is tasted first, second and third equally often. This is called sequential monadic testing. After each sample, people rate it on a tablet, rinse with water and wait two minutes before the next one.
The question it answers
Which recipe should move forward, and what would make it better? Liking scores answer the first part. Just-about-right questions and penalty analysis answer the second, by showing which attributes miss, in which direction, and how much each miss costs.
How we run it
We screen for people who buy the category often and set quotas so loyal buyers and competitor buyers are both in the room. The ballot uses the standard 9-point liking scale for overall, flavor and texture, plus just-about-right scales for the attributes a developer can change, such as saltiness or crunch. At the end, each person picks a favorite. In parallel, a small panel trained on reference samples rates how strong each attribute is, so consumer opinions can be tied to what is physically different about each recipe. The kitchen controls portion, temperature and the time from package to booth, because a chip from a bag opened an hour ago is a different chip.
What you get
A scorecard of liking scores with significance testing, just-about-right results, a penalty analysis that ranks the misses worth fixing, an attribute profile showing how the recipes differ, preference share by buyer group, and a clear recommendation: which recipe to advance and what to adjust first.
Where it goes wrong
Small samples and fixed serving orders cause most of the trouble, since the first sample of the day often scores higher just because it came first. Teams also chase every just-about-right miss, when only misses that affect a meaningful share of people and cost real liking deserve a fix. Palates tire quickly with strong flavors, so we limit the number of samples per session. And a few bites in a booth is a small dose. When repeat eating matters, follow the taste test with an in-home use test.
The sample below follows a fictional kettle chip brand testing two lower-sodium recipes against its current sea salt chip. Click the tabs to move from the design to the ballot, the attribute profile, the penalty analysis and the readout. Every brand, person and number is invented.
Illustrative only. All brands, products, tasters and figures are fictional. Open the sample full screen.