GLP-1s Are Breaking Your Food Research Framework

GLP-1 medications are distorting food and beverage research in studies running right now. Liking scores look clean but can stop predicting purchase. Sampling norms quietly average two different consumer populations into one number.

Meanwhile, for GLP-1 users, well-being drivers that hedonic batteries were never built to detect are starting to shape what they buy. Most testing frameworks were designed before a drug could change how a large share of consumers experience appetite and taste.

Key Takeaways

• GLP-1 drugs alter appetite and food preference, so liking scores may no longer predict purchase intent for a growing consumer segment.
• Samples that blend users and non-users can mislead, because GLP-1 users make up a material share of some food and beverage samples.
• For GLP-1 users, well-being drivers such as satiety, portion satisfaction and nutritional value can weigh as heavily as taste.
• Teams that flag GLP-1 status in screeners, as an optional and consented question, can analyze users and non-users as separate segments.

How GLP-1s Change the Way Consumers Experience Food

The liking scores your team collected last quarter may already be misleading. Semaglutide and tirzepatide suppress appetite. A study of semaglutide found it also reduced food cravings and lowered preference for high-fat foods (Blundell et al., 2017). The drugs change when people eat at all. A nine-point liking scale captures none of that.

The disruption runs through four connected mechanisms. Users commonly report that foods that once felt indulgent now taste flat or even aversive. Appetite suppression separates flavor liking from consumption intent, so a respondent can rate a snack's flavor positively and still have no intention of buying it because the product doesn't fit how they now eat. Single-serve and snack formats sized for a pre-GLP-1 appetite can feel excessive. Eating occasions also become less frequent, which thins out moments like the afternoon snack that whole product lines were built around.

Each of these mechanisms challenges a different assumption baked into standard food research. Testing batteries assume taste drives purchase, volume and pack-size tests assume a consistent appetite baseline and usage-and-attitude studies assume the occasions they mapped last year still exist. For any sample where GLP-1 prevalence is material, those assumptions may no longer hold. The liking score looks fine while the business signal is wrong.

The Sampling Norm That No Longer Holds

Most current study designs never identify GLP-1 users as a distinct segment, so their answers get averaged in with everyone else's. That segment is not small. In KFF's November 2025 poll, 12% of US adults said they were currently taking a GLP-1, rising to 22% among adults aged 50 to 64. In a food and beverage sample weighted toward midlife adults, users can make up a meaningful share of respondents. Collapsed into a single aggregate, the overall liking metric carries two populations with structurally different relationships to the product. The average obscures both signals.

Teams that add GLP-1 status to screeners can read users and non-users as separate segments. Because it is health information, it works best as an optional question asked with explicit consent and handled as sensitive data under industry codes and applicable state privacy laws. Used as a behavioral and attitudinal lens as well as a segmentation flag, it changes the consumer insights question itself: preference is now measured within a category where people consume in very different ways.

Measuring Well-Being Drivers Alongside Taste

For GLP-1 users, well-being drivers can weigh as heavily as taste. These include satiety duration, portion appropriateness, perceived nutritional value and how a food fits an eating occasion. A few well-being items added to an existing battery won't capture them. A concept testing framework built around taste misses the job the product does for a consumer whose eating goals have changed.

Open-ended interviews are where this shows up. Asked what satiety means to them or how they judge whether a food fits their current eating pattern, GLP-1 users may answer in functional and well-being terms as often as in terms of taste.

A qualitative feedback analysis framework built on pre-GLP-1 response patterns will under-code those signals unless the researcher updates it. In Enumerate, the auto-generated first coding frame goes through theme review before it is applied. At that step the researcher can rename, remove or replace themes and add manual ones for satiety and portion fit.

Planning a concept test that reads GLP-1 users as their own segment? Book a demo with Enumerate.