When does a recall really end?

A closed campaign is not the same as an eliminated hazard

A recall has at least two possible endings: the campaign stops operating, and the affected products stop creating exposure. Those events need not coincide. For products installed in buildings or used for decades, recall management needs a way to preserve knowledge after the original campaign, customer relationship and responsible personnel have changed.

The useful management question is therefore not only “How many units did we replace?” It is also “What do we know about the units that may still be out there?”

The case: a warning renewed 25 years later

On 24 September 2026, CPSC and Tyco Fire Products renewed their warning about unreplaced Central fire sprinklers. The products were manufactured from the mid-1970s to June 2001; the original recall began in July 2001, was modified in 2003 and closed in 2007. The new notice warns that affected units may remain in service and calls for assessment and replacement through qualified fire-protection professionals (U.S. Consumer Product Safety Commission, 2026).

This is a renewed warning about an old recall, not a newly discovered problem in a newly recalled 25-year-old product. That distinction makes the case particularly useful: substantial past remediation can coexist with unresolved exposure much later.

The product may outlive its information trail

Consider a hypothetical building that changes ownership twice after a recall. The installed equipment stays in place, but the person who received the original notice leaves, the maintenance contractor changes and records are moved between systems. The next owner may know the building’s age without knowing which individual components were corrected.

This scenario is not a claim about any particular property in the Tyco case. It illustrates why an original purchaser database and an archived press release may be insufficient as a long-term management approach.

For durable products, the relevant contact network may extend to maintenance providers, inspection services, property managers and subsequent owners. Recall information needs to remain attached to the product or asset, not only to the first customer relationship.

Separate “not found” from “confirmed resolved”

I would review an installed-product inventory using explicit evidence categories: confirmed affected and outstanding; confirmed corrected; confirmed removed; and status unknown. An unanswered letter should not silently become a completed case. Nor should a missing record automatically be presented as proof that a product is unsafe.

The distinction is particularly important when reporting completion rates. A high percentage can look reassuring while hiding a poorly understood denominator. Managers should state which population the percentage covers and what remains uncertain.

A practical starting point is an exception list: which assets have no current location, which have ambiguous identifiers, and which have a reported correction without supporting evidence? Each exception needs an owner and a next step rather than indefinite disappearance into an archive.

Build recall memory into ordinary handovers

As a management proposal, I would make unresolved recall status visible in asset transfers, contractor changes and maintenance handovers. Retain a usable route to the official notice and a record of what was done, when and to which identifiable unit. The appropriate record-retention arrangements will depend on the product, organization and applicable requirements.

Manufacturers can ask how later owners will find historical notices. Law firms can examine the difference between campaign closure and claims about hazard resolution. Insurers can ask what evidence supports an account of older installed equipment and how unresolved items are followed up. None of these questions alone determines legal responsibility, coverage or a premium.

Measure persistence, not just the launch

The framework proposed here is not a measured evaluation of Tyco’s campaign. The reviewed notice does not establish a current nationwide completion rate, and this article does not attribute any particular fire outcome to a specific component.

To assess long-term effectiveness, I would look at the ability to identify remaining units, retrieve correction records and reach the person currently responsible for the asset. Those capabilities may matter long after media attention has disappeared.

Recall closure should describe what the evidence supports. The calendar cannot tell us whether the last affected product has been corrected.

References