Product Safety
The Control Exists. Does It Still Work?
A legacy recall still in service, certified safety footwear failing market checks, and a new cyber reporting clock all expose the same weakness: controls must continue to work.
A 25-year-old recall, continuing-production failures and a new 24-hour cyber reporting clock expose the same governance problem: a control can exist without continuing to work.
THE READINESS DIRECTIVE
PRODUCT SAFETY BRIEFING
Issue 01 | Week commencing 21 September 2026
Executive intelligence on product risk, regulation and consequence.
The Control Exists. Does It Still Work?
A control is easy to point to.
A certificate.
A recall notice.
A test report.
A maintenance programme.
A cybersecurity procedure.
The harder question is whether that control still performs when the product has left the factory, entered the installed base, changed through production or software evolution, or generated information somewhere inside an organisation that now has hours - not weeks - to respond.
Three recent developments make that distinction unusually visible.
They come from different sectors and jurisdictions, but they ask the same leadership question:
Does the control still work when it matters?
1 | WHEN RECALL CLOSURE IS NOT RISK CLOSURE
On 24 September 2026, the U.S. Consumer Product Safety Commission and Tyco Fire Products issued a renewed warning concerning previously recalled Central Sprinkler fire sprinkler heads containing O-ring seals.
The story began a quarter of a century ago.
Central Sprinkler Company, an affiliate of Tyco Fire Products, announced a voluntary replacement programme with CPSC in July 2001 covering approximately 35 million Central sprinkler heads, together with about 167,000 O-ring sprinkler heads sold by Gem Sprinkler Company and Star Sprinkler. The programme was subsequently modified in 2003.
Corrosion, minerals, salts and other contaminants in water could affect the rubber O-ring seals over time and, under some circumstances, prevent a sprinkler head from activating during a fire.
The replacement programme ultimately closed on 31 August 2007 after millions of sprinkler heads had been replaced.
Yet on 24 September 2026 - 25 years after the original recall - CPSC and Tyco returned to the issue.
Their message was not that a new defect had been discovered. It was that the old one may still exist in the field.
CPSC and Tyco say a significant number of unremediated recalled sprinkler heads may remain installed in homes, residential facilities and commercial buildings across the United States. Owners and responsible organisations are being urged to identify and replace affected sprinklers.
The 24 September announcement also refers to the July 2025 fire at Gabriel House, an assisted-living facility in Fall River, Massachusetts. Ten residents died and dozens were injured. CPSC and Tyco state that dozens of recalled sprinkler heads at the facility had not been replaced despite having been identified during a prior fire-safety inspection.
CPSC does not state that the recalled sprinkler heads caused the fatalities.
In preliminary findings issued on 22 July 2025, Massachusetts investigators classified the ignition cause as undetermined, with two possible accidental sources identified. That finding concerns how the fire started, not how the sprinkler system performed.
For this briefing, the governance significance is the continuing presence of recalled equipment in the installed base long after the corrective-action programme had closed.
The Readiness Directive (TRD) Perspective Perspective
This is an installed-base governance problem.
The control was real: a recall was launched, a replacement programme operated for years, remediation took place, and inspection and maintenance requirements remained relevant.
Yet the 2026 warning exists because risk appears to have remained beyond the life of the programme itself.
For executives, the more demanding question is:
Can we prove that the remedy reached the end of the installed base?
A closed recall file is evidence that a process ended.
It is not necessarily evidence that every affected product disappeared.
Executive questions
For historic recalls or field actions, can we distinguish programme closure from confirmed installed-base remediation?
Do maintenance, service and inspection systems retain product-recall intelligence for the full service life of long-lived assets?