Ozempic: A surprising side effect

How GLP-1s are changing food purchases and what we eat.

Post summary

• New GLP-1 drugs are booming in popularity—12.5 percent of Americans have tried them. The figure is expected to grow by roughly 350 percent by 2035.

• The drugs work on the brain to lead people to eat significantly less.

• One curious side effect of the drugs: They’re changing how much food we buy and even the types of food we buy.

• That’s according to a fascinating and rigorous new study from researchers at Cornell.

• We spoke to the researchers to learn how GLP-1s are changing shopping, the food industry, and their users.

• You’ll learn the details of their results and what it means for the future of food.

Housekeeping

• This post, like all Monday posts, is free to all subscribers. If you want full access to all Two Percent material, our community, and our New Year Challenge, become a Member here:

• New partner alert! Two Percent is now partnered with Function Health. Function gave me critical insights into my health that my routine annual physical didn’t. I was able to take my Function data to my doctor and make educated decisions on my own terms. Function was also 32% cheaper than going through my insurance for the same blood work. If you’re interested in getting blood work done, I recommend two things: work with a doctor to analyze your results and read this post to understand the upsides and downsides of direct-to-consumer blood work in the absence of symptoms. Use this link to learn more about Function and to skip the waiting list.

Audio/podcast version

• Listen on your preferred podcast player here.

The post

I recently found some fascinating research from two scientists at Cornell University named Sylvia Hristakeva and Jura Liaukonyte.

Their study investigated how GLP-1s like Ozempic and Wegovy are changing their users¹.

But the paper didn’t examine changes you might expect, like the drugs’ physical side effects or how they impact our health or waistlines.

Instead, their study looked at how GLP-1s are changing our wallets.

Hristakeva and Liaukonyte are both economists who study consumer food purchasing behavior. They also teamed up with Leo Feler, an economist with Numerator and thought leader in the GLP-1 space.

As of May, 1 in 8 Americans had tried a GLP-1. Right now, about 7 million Americans are on one the medications. But by 2035 that figure is projected to hit 24 million.

We use GLP-1s to control diabetes, lose weight, and more. The drugs seem to “work” in the brain by reducing hunger and slowing down how fast your stomach empties, leading people to eat less.

And they’re working better than anything we’ve ever had. For example, some obese people are losing as much as 15 percent of their body weight.

The lead author, Hristakeva, was traveling, but I had a long and fun conversation with Liaukonyte about their work. She told me this: “I think the statistic that is incredibly compelling is the first sentence in the paper. ‘After more than a decade of steadily rising obesity rates, 2023 marked the first recorded decline.’ We cannot causally link that decrease to GLP-1s, but chances are it’s due to GLP-1s.”

As economists, Hristakeva and Liaukonyte wanted to know how the drugs are changing our food purchases—how much food we buy, the types of food we buy, and more.

This has massive implications for the future of food, the food industry, and what we’ll see on grocery store shelves.

How they got their data

The two teamed up with Feler at Numerator, a company that tracks shoppers and recently started surveying them about GLP-1 use.

Numerator incentivizes 150,000 households in the U.S. to upload their shopping receipts to a portal.

They get the date and time of every trip, the name of the retailer, the store type (e.g., grocery stores, convenience stores, fast food stores, food delivery apps), the total amount spent on an item, and the total quantity of items purchased. They also know the shopper’s age, income, and more.

In 2023, Numerator also started asking all the households they track if anyone in the home was taking a GLP-1, when they started the drug, how long they stayed on it, etc.

Using Numerator data had a big upside for the study: They capture homes that use GLP-1s off-label and people who pay out of pocket, so it gives a truly representative sample of people on the drugs. For example, if you only used insurance data, you’d miss a huge swath of people on the drugs.

Who they included in the study

Hristakeva and Liaukonyte looked at a representative sample of Americans and what they bought at the supermarket both before and after they started taking GLP-1s.

The important part, Liaukonyte explained, is that they included control households. This means they compared households where someone adopted a GLP-1 to similar households where no one used a GLP-1.

• Why they did that: You need to have control households so your data on households where a person adopted GLP-1s aren’t skewed by forces other than the drugs.

  • Otherwise, inflation or economic shifts could lead people to buy more of one food and less of another unrelated to GLP-1s. For example, people often buy more peanut butter during economic downturns. If you don’t include control homes, changes you see could be for reasons other than GLP-1 drugs.

Two other smart decisions they made:

1. The study participants also had to have been on GLP-1s for at least six months, so the data was consistent.

2. They also separated people taking GLP-1s for diabetes from those taking the drugs for weight loss.

   1. That’s because a person who takes a GLP-1 for diabetes may make different purchases than someone taking it for weight loss. For example, a person taking the drug for weight loss might be more incentivized to buy healthier foods.

The final sample included roughly 23,000 people, 3,000 of whom started taking a GLP-1 between January 2023 and July 2024.

How food purchasing changed

After they ran all the data, they found:

• The average home with at least one GLP-1 user reduced their grocery spending by approximately 6 percent.

• Higher income houses (making more than $125k a year) with at least one GLP-1 user reduced their spending by roughly 9 percent.

How that translates to dollars

Liaukonyte told me, “You might think, okay, about a six percent or nine percent drop in spending is a modest number—but it’s actually not because it’s at the household level.”

Translation: The drop wasn’t just for a single person in the house—it was for everyone in the house. And that’s because one person in a household usually does the shopping. If that person buys differently, it impacts the entire family.

And that translates to real dollars.

• The average household spent $416 less each year.

• The average higher income household spent $690 less each year.

What foods people bought less of and more of

This is what I thought was most fascinating.

“One of the things that was kind of surprising to us is that the spending reduction was kind of across the board, across most categories,” Liaukonyte said. “It wasn’t like people switched from some unhealthy category of food to more healthy categories—it was an almost across-the-board reduction in spending.”

Five things stuck out at me:

• Junk food had the biggest drops
  ”Decreases in spending were really driven by the categories of ultraprocessed, sweetened or savory foods,” Liaukonyte said. “The biggest decline, around 11 percent, was in chips and other savory snacks category like pretzels, popcorn, that type of stuff.”

• But many foods we might consider healthy also dropped
  For example, shoppers on GLP-1s also bought less eggs, meat, milk, seafood, and dried fruit, and nuts.

• There were some differences in purchases among people taking GLP-1s for weight loss versus diabetes
  For example, people taking the drugs for weight loss bought less candy and alcohol, but diabetes users bought about the same amount. This suggests weight loss users may be more likely to make “healthier” choices once on the drug.

• Only two categories out of 40 increased
  Fresh produce and yogurt rose. Anecdotally, many GLP-1 users say they crave more fruits and vegetables than they did before they got on the drugs. As for yogurt—who knows? Perhaps people are looking for easy-to-eat forms of protein. This theory is consistent with the fact that meat sticks and nutrition bars rose slightly, but the change wasn’t significant.

• One family member on GLP-1s could make the entire family healthier on accident.
  Recall that when a family’s primary shopper buys differently, it impacts everyone in the home. In theory, if a GLP-1 shopper starts buying more fruits and vegetables and less junk food, everyone else in the home has more access to healthy foods and less access to junk.

Here’s what happened to the 40 categories of food:

Any dot to the left of the dark vertical line represents a drop in spending.

Who takes GLP-1s and why

Hristakeva and Liaukonyte found:

• Weight-loss users had higher incomes. They were disproportionately concentrated in higher-income households, with the largest share in those earning $150,000 or more. That could be because GLP-1s are available for weight loss through companies like Weight Watchers and Noom.

• Diabetes control users were more evenly distributed across income brackets.

• Weight loss users were younger. They tended to be younger and middle-aged, with the largest share between ages 34 and 64.

• Diabetes users were older. They were mostly concentrated among groups older than 55, which is consistent with the fact that more diabetics are over 55.

How the food industry is responding to GLP-1s

Hristakeva and Liaukonyte wrote:

This transformation presents both challenges and opportunities for the food industry as GLP-1 medications reshape consumer food purchasing patterns. Companies that rely heavily on calorie-dense, processed, or indulgent foods are likely to face declining demand and need to reassess their product portfolios to remain competitive.

Liaukonyte said, “I think there will be changes that help the GLP-1 user navigate the grocery aisles and help them find the items. For example, I saw that Healthy Choice is now coming up with a label that they added that says ‘GLP-1 friendly’ on their frozen items that are protein heavy.”

Industry will indeed respond.

As the study and many anecdotes suggest, one interesting effect of the drugs is that they seem to increase cravings for fresh fruits and vegetables. But those foods have a short-shelf life and aren’t as profitable.

The food industry is currently trying to formulate profitable packaged foods that appeal to GLP-1 users. As the study found, we saw some of the biggest drops in items like salty snacks, baked items, and frozen packaged meals.

Food labs are currently experimenting with versions of those items that appeal to the taste buds of people on GLP-1s.

We’ll also likely see more processed products with higher protein, given that protein may mitigate some of the muscle loss that can happen as people lose weight.²

Have fun, don’t die,

-Michael

1. They titled the study “The No-Hunger Games: How GLP-1 Medication Adoption is Changing Consumer Food Purchases” … I love it when academics have a sense of humor.

2. I’m not convinced that muscle loss happens at a higher rate in GLP-1 users than it does with similar levels of weight loss via diet alone.