A conversion event tells you that someone clicked. A receipt tells you they bought four items, spent $31, picked up wine and steak alongside your product, did it at 7:40 on a Thursday evening, and paid full price for everything in the cart.
Those are not two versions of the same data point. One is a record that an action happened. The other is a description of a person.
For CPG brands the difference matters more than it does almost anywhere else, because the purchase happens somewhere you cannot see. The US Census Bureau put e-commerce at 17.1% of US retail sales in the second quarter of 2026, which means roughly 83 cents of every retail dollar still moves outside the channel your analytics can follow. In food and beverage the gap is wider still, with EMARKETER reporting that 91% of food and beverage sales take place in physical stores.
A verified receipt is not one data point. It is a structured record with at least nine usable fields, and none of them require a survey or a panel to obtain:
Circana's basket work shows what this kind of data yields once you have enough of it. A premium sparkling water brand found that baskets containing its product averaged $5 more than baskets without it. A cereal brand found milk and bananas appearing together in 70% of trips containing its product. A pet food brand found online orders adding $12 more to the basket than in-store trips did.
Those are the kinds of facts that reposition a brand with a retailer, and they are also the kinds of facts a click never produces.
Stack enough receipts and distinct shoppers separate out. Three patterns show up consistently.
The occasion buyer. Four items, $31 basket, wine and steak and seafood, Thursday at 7:40 pm. Individual, premium, lifestyle-oriented. Probably hosting or treating themselves.
The household shopper. Twenty-eight items, $147 basket, kid snacks and staples and bulk, Saturday at 9:15 am. A planned grocery run, repeat-capable, and the volume customer most CPG brands are actually trying to acquire.
The deal-only shopper. One item, $12.99 basket, heavy discount reliance, Tuesday at 2:05 pm. Incentive-driven, useful for hitting a volume number, risky to build a business on.
All three redeemed the same offer off the same video. Every conversion pixel in the world would record them identically.
Beyond the archetype, the receipt supports a second layer of inference. Buying occasion, meaning why this product in this moment. Market positioning, meaning whether the rest of the cart says premium or value. Trial versus loyalty. Household versus individual, which is the foundational CPG segmentation. Price sensitivity, meaning whether the rebate drove the purchase or merely enabled a purchase that was going to happen. Planned versus impulse. And substitution versus incremental, meaning whether you replaced a rival in that cart or added genuinely new behavior.
The deal-only shopper is where receipt data earns its keep, because the cost of not seeing that segment is measured in years rather than weeks.
The canonical work here is Michael Lewis in the Journal of Marketing Research, who found that customers acquired with a 35% discount were worth roughly half the long-term value of customers acquired without one. Anderson and Simester, running three large-scale field experiments, found something more uncomfortable still: deep promotions increase future purchases among first-time customers but reduce them among established ones, and one of the four mechanisms they identify is increased deal sensitivity. You can train your own best customers to wait for the discount.
Both papers are twenty-plus years old and drawn from catalog and online grocery contexts rather than modern omnichannel CPG, so pair them with current behavior rather than resting on them. NIQ's 2026 work found 28% of consumers choosing whatever brand is on sale, against 62% who say trusting the brand they buy is very important. Those are two real and differently valuable populations, and the receipt is what tells you which one your campaign is pulling.
The tension shows up in outcome data too. In Circana's independent causal analysis of 48 CPG digital promotion campaigns, competitive switchers produced the highest incremental lift at 32.9% while existing brand buyers delivered the highest incremental return on ad spend at $5.05. Highest lift and highest return come from different shoppers. Without basket-level visibility you cannot tell which one you bought.
The obvious objection is that retailers already have this. They do, for their own stores, and that turns out to be the limitation rather than the solution.
FMI's 2026 shopper research found that US shoppers visit 5.4 separate grocery banners per month, rising to 6.7 among Gen Z. No single retailer's data can see a shopper's actual behavior, because most of it happens at other retailers.
The infrastructure meant to bridge this is not there yet either. EMARKETER reports that 66% of organizations use data clean rooms in some capacity while 39% struggle to derive actionable insights from them, and only 48% of US retail media networks offer clean room capabilities at all. On the in-store side, 86% of grocers say their media channels remain completely siloed or only partially integrated, and in-store retail media accounted for just 3.3% of non-Amazon retail media spend in 2025 despite 91% of food and beverage sales happening in store.
The structural problem underneath all of this is old and unresolved: CPG brands sell through retailers, so they have never had a direct relationship with the shopper. Everything else follows from that.
It is also why the offline attribution gap is so large. The cleanest causal estimate remains Johnson, Lewis and Reiley's field experiment across three million users, which found that 84% of the sales lift generated by an online display campaign came from offline purchases. That study is from 2017, and its finding is that the overwhelming majority of what your media does is happening where you cannot measure it.
The most common mistake is treating receipt insight as a targeting input. It is a creative input first.
Receipt insights rarely change bids or knobs. They change the story you tell and the people you ask to tell it.
Four moves, in order of value:
The funnel work is supportive rather than primary. New ad sets aligned to discovered segments, budget biased toward higher-quality receipt cohorts, receipt-based lookalikes once volume allows, and geographic emphasis shifted when purchases and leads diverge.
A receipt is proof that a sale happened, which is the reason most brands want one. The more useful thing is that it is also a description of the person who made it.
Basket size, co-purchases, competitor presence and discount density turn a transaction into a portrait of a customer, and a portrait of a customer into better creative. That loop is available to any brand willing to ask shoppers to upload proof of purchase, and it does not depend on a retailer choosing to share.
The full framework, including the nine raw receipt inputs, the eight level-two inferences and how receipt findings feed back into the weekly campaign decision, is available here.
.svg){kind=icon}
.svg){kind=icon}
.svg){kind=icon}
.svg){kind=icon}
.svg)
.svg)
.svg)
