Nielsen Can't See a GLP-1
Three of the biggest CPG moves this month came from operators who can.
Two consumers walk into the same Kroger on a Saturday morning. Same age. Same gender. Same race. Same zip code. Same household income. Same education. Same family structure. By every variable CPG uses to build an innovation brief, zone a trade promotion or allocate a retail-media buy, they are the same consumer.
However, in this case, one of them is on a GLP-1. The other isn’t.
By the time they leave the store, their carts look structurally different. One is buying 44% fewer salty snacks and 58% fewer sweets. One is buying 55% more fresh produce, alongside more yogurt, fresh chicken, protein shakes and protein bars. Demographics cannot see the line between them. Neither can any syndicated consumer data feed most CPG brands are buying (Nielsen, Circana, IRI, Numerator, etc.) because none of them carry a “GLP-1 Status” variable and none of them capture the mindset of a consumer at a particular time when he or she picks up a product or puts it back down.
The three biggest CPG moves of the last few months all came from operators who don’t have to rely on those feeds: On March 3rd, Target’s chief merchandising officer described the company’s 2026 grocery strategy as “emerging, wellness, and owned brands intersect” - a billion dollars of capex pointed at a shelf organized against first-party data from Roundel, Target’s retail media network. On April 15th, Walmart announced a 10,000-SKU redesign of Great Value aimed at the higher-income shoppers who migrated during inflation - planning from Walmart Connect, the largest retail media operation outside Amazon. On April 16th, PepsiCo retired forty years of Gatorade positioning - the $6 billion brand is no longer built for “athletes,” it’s built for “a long flight, going for a walk, nursing a hangover”. This is backed by a 24% growth in its Tasty Rewards first-party data last year, where members spend 60% more per year on PepsiCo products vs. the average U.S. consumer.
Three moves, one advantage. Each operator already owns the data that replaces demographics.
For forty years, demographic planning has been the industry’s default - not because it was accurate, but because it was syndicated. Every brand bought the same feed from Nielsen or Circana, cut by the same age/income/ethnicity grid and built the same kind of plan.
Occasion planning has no syndicated equivalent. You cannot buy a “post-workout protein window × GLP-1 user × Tuesday 4 p.m.” dataset. The data that resolves occasions lives in one place: first-party data around consumer behavior, such as: loyalty programs, DTC subscriptions, retail media clickstreams, basket composition at the household level, etc. It sits with whoever owns the direct relationship with the consumer.
Demographic data was the last data layer brands owned on equal terms with retailers. Occasion data isn’t. This is not a segmentation upgrade, it’s a power transfer.
Coca-Cola took full ownership of Fairlife in January 2020 for $980 million. The brand crossed $1 billion in retail sales in 2022 and is approaching $4 billion in retail value at the end of last year. Coke did not integrate Fairlife into its legacy commercial machinery. It kept the brand running as a standalone operating company with its own P&L, its own commercial team and its own consumer insight flows. The acquisition bought a brand, but the separation protected a data capability.
Nestlé did the same thing with Vital Proteins, which it acquired in 2022 after taking a majority stake in 2020, by placing it inside Nestlé Health Science rather than Nestlé’s food operating company. Unilever did it with Liquid I.V. in October 2020, a DTC-first hydration brand that had done $100 million in its first five years and is now the largest brand inside Unilever's €1.9 billion Health & Wellbeing business - intentionally kept separate from core food and beverage. Three acquisitions, three separate operating structures, three brands that arrived with their own consumer data. The pattern looks like buying occasion-native products on the surface. Underneath, it's buying the capability to plan around occasions in the first place.
The counterexample is Beyond Meat: The company read the GLP-1 demand signal correctly and responded by announcing Beyond Immerse, a protein beverage line explicitly targeted at GLP-1 users. It is currently shipping with 2.8% full-year gross margins and a 15.6% revenue decline. The signal read was right, but the data infrastructure behind it is a small first-party footprint - no loyalty program at scale and a consumer learning loop that depends on retailer POS data arriving on a quarterly lag. Beyond Meat is trying to win an occasion bet against competitors who watch the same consumer in near-real time.
Reading the signal is table stakes. Owning the data to plan against it is the actual capability.
Private label is at record U.S. unit share in food and beverage and most of the coverage attributes the rise to price. And while price is doing real work, the layer underneath that matters is the data. The retailers who own the portfolios (Great Value, Simple Truth, Kirkland, Good & Gather, etc.) have already been planning against their own shopper behavior for a decade. Their merchandising runs against what the shopper actually buys, not what the syndicated panel says the cohort should want to buy. That’s a structural advantage that compounds independently of tariffs, inflation or GLP-1 adoption.
The 2026 planning question that almost every CPG brand is asking (“what’s our GLP-1 strategy?”) is pointed at the wrong layer. While product reformulation is the visible layer, data ownership is the one that decides whether the strategy is executable at all. From the data you own, can you tell which of your SKUs are losing share to GLP-1 users this week? Or whether your shoppers are trading down on price or leaving the category entirely? How about which of your consumers are on a frequency-decline path that ends with them leaving the brand? If the answer is “we’d have to wait for the Circana cut,” then the plan is being built against a worldview that stopped solving for consumer needs three or four years ago.
The consumer changed, but the question isn’t whether CPG notices - It’s which CPG companies notice in real-time, using data they actually own; and which ones are still waiting for the quarterly Nielsen feed to tell them what has already happened.