If you run supply chain and distribution for a multi-unit restaurant company, the RFID case you take to your board should be built on evidence a third party carries — because a vendor's claim about a vendor's product is not evidence, and a deadline argument is falsifiable in one search. There are exactly four artifacts on the public record worth citing. Here they are, dated, quoted, and scoped. The discipline is the content.
The four artifacts
| Artifact | Source and date | What it establishes | Exact scope |
|---|---|---|---|
| A1 | MIT Technology Review Insights + GS1 US, "Unlocking supply chain resiliency," 2023-10-30 | A McDonald's ↔ Golden State Foods RFID pilot, fresh beef, manufacturer to restaurant. Sue Fangmann, US Supply Chain Services Director at McDonald's, on the record: "create a golden digital thread of traceability." | Pilot only. The record does not support "McDonald's US is rolling out RFID," and neither do we. |
| A2 | RFID Journal, "Chipotle and Suppliers Partner to Roll Out RFID Nationwide," 2023-12-19 | A nationwide rollout: roughly 3,200 US restaurants, about 330 product categories, expanding to 19 distribution centers. Vendors named in the piece: Mojix (software), Zebra (handheld readers), Avery Dennison Smartrac (tags). Mojix's president on the record that the program is "driven by the compliance requirements of FSMA 204." | A third party carries the regulation-to-spend causal claim. It is theirs, not ours. |
| A3 | McDonald's China + Cainiao (Alibaba), 2024-11-28 | "One Box, One Code" — an RFID digital ID per package, described as the "first large-scale application of this technology in the domestic restaurant industry," with receiving and inventory efficiency up 30%. | China only. The +30% figure never travels to the US business, in our copy or yours. |
| A4 | GS1 US case study, Golden State Foods, 2024-02-09 | Case-level RAIN RFID at the Opelika, AL protein plant: pre-tagged corrugate, GS1 TDS 2 encoding of production date, 100% encode success. The customer appears only as "a leading QSR customer." No vendors are named at all. | The best-documented supplier posture on public record — and a study of how quietly this segment talks. |
What each does not support
Every artifact above invites an overclaim, and the right board deck refuses each one in print.
A1 invites "McDonald's is doing it." The record supports a pilot with one supplier on one product line, plus a named executive's stated ambition. Cite the ambition; never inflate the scope.
A2 invites "the industry is moving." It supports one chain moving, at real scale, with a named software vendor attributing the spend to FSMA 204. That attribution is the single most useful sentence in the set — precisely because someone else said it.
A3 invites importing the +30%. The number is real, dated, and attributed — and it belongs to McDonald's China and Cainiao, in a Chinese operating context. Use it as the shape of the ROI, never as your forecast.
A4 invites "suppliers are ready." It supports one supplier achieving a 100% encode rate for one customer's program. What it actually proves is subtler and more valuable: when a QSR customer specifies case-level encoding, a serious supplier can execute it completely.
The ROI shape the record supports
Read together, the artifacts support an operational case, in this order: receiving labor and inventory accuracy first (that is what A3 measured and A2 funded), recall readiness as the floor underneath (that is what A2's own vendor attributes the spend to). Compliance is the floor, not the pitch. The full cost stack behind that case — tags, readers, and where the money actually goes — is worked through in the restaurant-chain cost and ROI read, under the RFID food-traceability pillar.
The reason third-party-carried claims beat vendor claims in your deck is the same reason a traceability record should be verifiable without joining anyone's network: evidence that depends on the party who benefits from it is not evidence. Hold your data layer to the standard you hold your citations to.
Run this review on any vendor's case-study page
The method transfers. For each claim on the page, ask: who carries it — the vendor or a third party? Is it dated? Is the scope stated, and does the headline match the scope? Does a number from one geography or one pilot travel, unlabeled, into a general promise? Most case-study pages fail in the first two questions. The four artifacts above are what passing looks like — which is why they are the only four we use.
When your rollout's "where does the data live" question is ready for an answer, get started. The interview branches for restaurant and distribution operators, and it ends with us knowing enough about your DC network to be worth your time.