For the director of customer programs at a co-manufacturer — the seat where account retention lives, whose budget logic is cost-to-serve, and where every customer's traceability decision lands as an operational obligation nobody negotiated with you.
You have no Sunrise 2027 date, and you have every Sunrise 2027 date. Both halves are exactly true, and the whole discipline of running customer programs at a co-manufacturer is refusing to blur them.
No date, because GS1's programme is aimed at retail point-of-sale — lanes reading 2D marks by end of December 2027 — and you run lines, not lanes. Nothing in Sunrise obliges a co-manufacturer to do anything. Every date, because the parties who do hold clocks sit exactly one hop from you: the brand owner deciding its mark in its artwork cycle, and the retailer refreshing its lanes. Their decisions do not arrive at your plant as programmes. They arrive as mandates — a customer letter with specifications and a date — and the mandate is the only shape in which the entire 2D transition will ever reach you.
That is the structural fact this brief turns on: your exposure is mandate-mediated. You cannot see it coming by watching GS1; you see it coming by reading your customers. And it changes what preparation means — a party with its own clock prepares by complying; a party one hop removed prepares by building capability that answers whichever mandate arrives next.
The anatomy of a mandate, from the one on public record
There is exactly one artifact in the citable record that shows this obligation from the inside, and it rewards close reading: the GS1 US case study of Golden State Foods (February 9, 2024). A protein plant in Opelika, Alabama, tagging at case level with RAIN RFID on pre-tagged corrugate, encoding per GS1's Tag Data Standard (TDS 2) with production date encoded, achieving 100% encode success — for a customer the study names only as "a leading QSR customer." No vendors named, either. That anonymity is itself a fact about how this seat talks: the value of the programme accrues to the customer's name, and the supplier's excellence appears in public with the serial numbers filed off.
Decompose what that customer must have specified, clause by clause, and map each clause to the line capability that answered it:
| What the mandate specified | The capability that answers it |
|---|---|
| Case-level identification (not pallet, not lot-on-paper) | Item identification at the grain the customer's receiving reads — the grain decision is theirs, the execution is yours |
| RAIN RFID as the carrier | Tag application on your line — solved here as pre-tagged corrugate, moving the marking upstream into packaging procurement |
| TDS 2 encoding, production date included | An encoding step that writes instance data (this case, this production run, this date) — not a static label file |
| An encode-success expectation | Verification on the line, measured — "100% encode success" is a measured rate, which means someone built the measuring |
| Data the customer's programme can consume | Emission: what leaves the plant besides the truck — events, ASNs, the paperwork that joins to what their dock reads |
Now the observation that pays your seat: every row is a capability, and only the first two rows are customer-specific. Encoding instance data, verifying at line speed, emitting consumable records — those answer any customer's mandate. Which carrier, which grain, which date field — those are parameters. A plant that builds the rows once holds parameters per account; a plant that treats each mandate as a project rebuilds the rows per customer, at project prices, forever.
Capability once, mandates many
This is the emit-once architecture, stated in account language: your plant emits one conformant record of what it does — cases produced, encoded, verified, shipped — and each customer receives a view scoped to their products. The alternative you are being sold by every customer's chosen platform is join-their-network: one portal, one integration, one audit per account, with your switching costs compounding by customer count. One record with per-customer views inverts that: the marginal customer costs a grant, not an integration. The record itself is standard vocabulary — EPCIS 2.0 events, with the shipping side joined to the documents you already emit, because CBV 2.0 carries po and desadv as core vocabulary, so the ASN-to-events join is spec machinery, not custom work. And every event carries an attested who — the observer on the line distinct from the account warranting the capture — which is the difference between a record your customer's auditor reads and one they re-audit.
The platform behind this brief is that record layer; its stateless doors are live at epcis.dev today, and every open gap between the capability described and the capability hosted is a named entry in the open P0 ledger, where we tell you the truth about it.
Pricing it like the seat you sit in
Your budget logic is cost-to-serve, so refuse both wrong numbers. The internal-ROI fiction — "traceability will pay for itself in plant efficiency" — is not how this spend works and pretending otherwise gets the programme cut in year two. The panic number — whatever the incumbent platform quotes per customer connection — is the join-their-network tax, and it scales with exactly the thing you want to grow. The defensible number prices capability against the book of business it retains: what is the Opelika-shaped capability worth per account that would otherwise churn to a co-man who has it, or stay and be served at mandate-project prices? That is a number your CFO can audit, it is per-account, and it is the only frame in which "100% encode success" reads as what it is — an account-retention asset with a measured value.
Before the customer tables it
The mandate always arrives eventually; the only variable is which side of the table introduces it. At your next account review, table it yourself: "Here is the identification, encoding, verification, and emission capability we run; here is what we can switch on for your programme, and when." One page. The customer whose own artwork-cycle and lane clocks are running hears operational readiness from a supplier one hop removed — which is rare enough to be memorable.
What that page should claim for your plant, and in what order the capability gets real, is a read we produce per situation at get started: your address first under the one-message promise, then questions that branch on the mandates already in hand, the customer formats you maintain, and what your MES or WMS emits, ending in a written read that locks when it ends.