Methodology

How this map is built, and — just as important — what it cannot tell you.

The idea

Store brands rarely make their own food. A packer runs the line; the retailer's label goes on the bag. Nobody publishes that arrangement. But when there's a recall, the FDA names the recalling firm — the company that actually made the product — and often the store brand it was sold under. Line those records up and a private-label supply map falls out of public data.

This v1 covers Taylor Farms (parent: Taylor Fresh Foods, Inc.), one of the largest fresh-cut produce and prepared-salad packers in North America. It is built from 83 recall records in the openFDA food enforcement dataset, 2012–2025.

How a retailer gets identified

For each recall we read the product description, distribution pattern, and reason, and apply three signals in order:

A known store brand or retailer name in the text. Retailers hide behind house brands, so we keep a dictionary that un-hides them: Marketside → Walmart, Lucerne and Signature Cafe → Albertsons/Safeway, Private Selection and Fresh Foods Market → Kroger, and so on.
A "Distributed by…" clause, when no store brand is named outright.
A "food service" flag — the product went to restaurants or institutions, and the record does not name the customer.

If none of those fire, the product is Taylor Farms' own brand, or genuinely unidentifiable from the record. The raw FDA text is stored next to every extraction, so any row can be re-checked, and the whole pipeline is re-runnable against a fresh pull. The full dataset is downloadable.

Absence of a link is never proof of no relationship. We only see a retailer–packer pairing when it happens to surface in a recall. A brand that isn't on this map has no recall evidence tying it to Taylor Farms — that is all it means. It is not evidence that they don't work together.

Food service hides the chain. When a packer supplies a restaurant chain, the recall names the packer but usually not the buyer. The clearest example is in this very dataset: the 2024–2025 E. coli O157:H7 recalls of raw yellow onions "intended for food service" (11 food-service recalls appear here in total). Those onions were widely reported and confirmed by regulators to be behind a McDonald's Quarter Pounder outbreak — but you cannot learn that from the FDA record itself, which never names McDonald's. Food-service rows on this map show "customer not named" for exactly this reason. It is happening again as of this writing: the July 2026 cyclospora outbreak tied to shredded iceberg lettuce at Taco Bell surfaced Taylor Farms as the supplier only because the CDC and FDA named it — see the note at the top of the supply map.

Reading a Class rating

The FDA assigns every recall a class. Class I means a reasonable probability of serious harm or death (Listeria, Salmonella, E. coli, undeclared allergens). Class II means temporary or reversible harm. Class III is unlikely to cause harm. Of the 83 recalls here, 39 are Class I. We surface those counts as a rough, checkable safety signal — not a verdict. Recalls are a normal part of food safety, and every one of these has been terminated. A terminated recall still proves the retailer–packer relationship existed; that's what we're mapping.

One firm, many names

Taylor Farms files recalls under about 21 regional names ("Taylor Farms Texas, Inc.", "Taylor Farms Northwest, LLC", and so on). We collapse all of them to one canonical firm with a hand-maintained alias list plus a prefix rule, so the parent company is the single join key. The facility region is kept as detail, shown in the "Packed by" column.

Sources & verification

openFDA food enforcement API — recall records, public domain.
Each row's evidence link queries that API for the exact record: …/enforcement.json?search=recall_number:"<num>", so anyone can re-verify a claim against the source.

This project is transparency about who makes what. It is not an allegation of wrongdoing against any company named here.