Protein. Low-sugar. Plant-based. Artisanal. No added sugar. Indulgent. Clean label. Every one of them tested “important” in a rating scale, which is exactly the problem: rating scales let people call everything important, so you end up with a cluttered pack and no real hierarchy.
MaxDiff (best-worst scaling) fixes this by refusing to let respondents off the hook. Instead of rating each claim, they see a small set, usually 4-5 claims at a time, and have to pick the one that matters most and the one that matters least. Repeat that 8-10 times with different combinations and every claim gets compared against enough others to produce a real, ranked score of relative importance, not just a stack of similar-looking averages.
A few things worth getting right when you set this up in Helix:
One honest limitation: the worst-choice step takes real attention, and on a phone that second tap can feel like a chore after screen six or seven. We keep claim counts and exhibit numbers conservative for exactly this reason.
How are you deciding which claims make the final cut on pack right now, rating scales, gut feel, or something more structured?
Practical guides on questionnaire design, qualitative analysis and market entry. One email a month. Unsubscribe anytime.
The numbers: Saudi Arabia accounted for 39% of the region’s canned-food import value in 2024, worth roughly $1.3B, with canned vegetables (tomatoes, beans, peas, olives) forming the volume core of that trade. It’s also…
That's a quiet shift with real consequences for anyone selling packaged food into Canada. US rice has long held the majority of the broader Canadian rice market, but it's losing relative ground as South Asian…
That gap is why percent agreement alone shouldn't be the gatekeeper for coding quality, whether your coders are two humans, two AI runs, or one of each. The workflow we recommend before any codeframe goes into…
