The numbers from 2025 are difficult to look at squarely. A total of 3,295 product recalls were recorded across five major U.S. industries — one of the highest annual totals in the past decade. The volume of defective units recalled rose 26% year over year, from 681 million units in 2024 to 858 million units in 2025. There were 517 FDA food recalls — the highest count in nine years. Pharmaceutical recalls surged 140% by unit volume compared to 2024. Medical device recalls exceeded 490 million units, representing the second-highest total in 20 years.

These are not abstractions. Each recall event represents a product that reached or nearly reached people who trusted it to be safe, effective, and correctly manufactured. Each one traces back to a failure that happened somewhere inside a quality system — a process that deviated, a control that failed to catch the deviation, a person who didn’t know what they should have known before they touched that product.

And in the overwhelming majority of cases, the failure was preventable. Not theoretically preventable with perfect systems or unlimited resources. Preventable with the quality infrastructure most regulated organizations already claim to have — but aren’t fully using.

The Pattern Behind the Numbers

Quality failures don’t appear randomly across industries. They cluster around predictable failure modes that regulators, auditors, and quality professionals have documented for decades.

A peer-reviewed retrospective analysis of FDA pharmaceutical recalls from 2012 to 2023 found that the most common reasons for recall were sterility issues and inadequate compliance with current Good Manufacturing Practices. The cGMP recalls specifically exposed five primary types of violations: process control issues, inadequate storage practices, manufacturing problems, nitroso-amine impurities, and stability concerns. In 2024, the majority of recalled pharmaceutical units were due to cGMP deviations — the same root category that drove recalls a decade earlier.

The pattern in food and beverage is equally persistent. FSMA was designed to shift the industry from reactive to preventive — from responding to contamination events to preventing them through documented hazard analysis and verified preventive controls. Yet the number of FDA food recalls reached its highest level in nine years in 2025. USDA food recalls increased 20% in the same year. These aren’t industries that lack regulatory frameworks. They’re industries whose quality systems aren’t consistently connecting documented requirements to the people responsible for executing them.

In automotive, the complexity of supplier networks means that a quality failure at a Tier 2 or Tier 3 supplier can propagate through the supply chain to affect millions of vehicles before anyone identifies the source. Despite declining recall frequency in 2025, the automotive sector regularly contends with high-unit-impact recalls that trace back to inadequate component qualification, insufficient supplier training requirements, or processes that changed without controlled retraining of the personnel operating them.

In aerospace and defense, the consequence of quality failure is measured differently than in other industries — not in units recalled but in safety incidents, airworthiness concerns, and potential loss of life. AS9100 Rev D and NADCAP requirements exist precisely because the cost of failure in this sector is categorically different from a packaging defect or a labeling error. Yet aerospace quality teams face the same structural pressures as every other regulated industry: complex processes, high personnel turnover in critical roles, and quality systems that can drift out of synchronization with the workforce responsible for operating them.

Across cannabis and hemp, the regulatory environment is tightening rapidly. The FDA continues to issue warning letters to companies for cGMP failures. State licensing bodies are conducting third-party audits with increasing rigor. An industry that is still maturing its quality infrastructure faces the same core requirement as its more established counterparts: documented processes, trained personnel, and evidence that both are functioning together.

The specific industry, the specific regulation, and the specific product all differ. The failure pattern repeats: processes that aren’t consistently followed, training that isn’t consistently verified, and quality systems that treat these two elements as parallel tracks rather than integrated ones.

The Cost That Doesn’t Appear on Quality Budgets

Most regulated organizations track the cost of quality failures that reach the surface — recall expenses, regulatory fines, warning letter remediation, and consent decree oversight. These costs are high. In the UK alone, 50% of organizations surveyed by ETQ reported that their most recent recall cost between £8 million and £39.9 million. U.S. consent decree remediation programs have historically run into hundreds of millions of dollars for affected manufacturers.

But the full cost of preventable quality failures extends well beyond what gets recorded in a quality event log.

For Fortune Global 500 companies, unplanned downtime from quality and production failures now costs an estimated $1.4 trillion annually — a 62% increase from 2019 to 2024, according to Siemens’ True Cost of Downtime 2024 report. Unplanned downtime can consume up to 20% of a manufacturing facility’s productive capacity. These costs compound through supply chain relationships: a single plant shutdown can disrupt deliveries to multiple customers within hours, triggering contract penalties and relationship damage that outlasts the original quality event.

The hidden cost categories that rarely appear in quality budget discussions include:

Rework and scrap

A product that fails inspection before leaving the facility doesn’t generate a recall, but it consumes materials, labor, and production time that directly affects unit economics. Organizations that track rework rates frequently discover that the same root causes driving their most visible quality failures are also driving recurring rework patterns — often at a higher total cost than the external failures that prompted regulatory attention.

Batch rejection and production loss

In pharmaceutical and biotech manufacturing, a batch rejection triggered by a process deviation or documentation failure represents not just the cost of the rejected material but the full production slot consumed in manufacturing it. In a facility operating near capacity, a rejected batch is also a loss of revenue from product that could have been produced in the same slot.

Supplier disruption

For manufacturers operating under IATF 16949 in automotive, AS9100 in aerospace, or ISO 13485 in medical devices, a supplier quality failure can trigger a supplier corrective action request, a source approval suspension, or removal from the approved supplier list. Each of these outcomes carries operational consequences that extend far beyond the specific quality event.

Talent and management bandwidth

Quality failures consume organizational attention at a multiplied rate. An investigator reviewing a warning letter cited the example of a compounding pharmacy where FDA reviewers examined 23 variance and out-of-specification investigation reports from a seven-month window. None of them followed the firm’s own standard operating procedure for investigating OOS results, and nineteen of the 23 remained open at the time of inspection. The organizational bandwidth consumed managing unresolved quality events — investigations that don’t close, CAPAs that accumulate, escalations that absorb leadership time — is a direct cost of a quality system that isn’t working.

Regulatory remediation overhead

FDA warning letter responses require 15-day initial responses and documented corrective action plans. Warning letters that escalate to consent decrees typically require hiring third-party consultants, establishing independent oversight mechanisms, and demonstrating systemic remediation across all affected sites and processes. These activities consume resources that could otherwise advance product development, process improvement, or capacity expansion.

According to a survey of 750 senior-level quality professionals across the U.S., U.K., and Germany conducted by ETQ in 2024, 73% of organizations had experienced a product recall in the last five years, and 48% noted an increase in recalls over that period. These aren’t unusual events happening to poorly run companies. They are the normal experience of regulated manufacturing organizations operating with quality systems that have structural gaps.

The System Gap That Makes Failures Preventable

There is a specific structural gap that underlies the majority of preventable quality failures across regulated industries. It is not a shortage of quality procedures. Most organizations have extensive procedure libraries. It is not a lack of regulatory understanding. Quality professionals across these industries understand their obligations. It is not, fundamentally, a lack of effort or commitment.

The gap is between what quality systems document and what workforces actually do — and it persists because the systems that manage quality documents and the systems that manage employee training don’t consistently communicate with each other.

When a pharmaceutical manufacturer revises a critical SOP, the document control system registers the effective date and archives the previous version. What happens next is where failures begin. Someone must recognize that the revision requires retraining, determine which employees are affected, create training assignments, track completion, verify competency, and confirm that the people who make that product every day have actually absorbed the change. In organizations running disconnected quality and training systems, that chain of steps depends on human memory and manual execution at every stage.

The analysis of pharmaceutical recalls over a decade found that quality management systems must oversee the manufacturing process, quality controls, personnel training, and documentation together to prevent the cGMP violations that drive recalls. “Together” is the operative word. These elements function as a system, or they fail individually.

Without clear, documented, and enforced SOPs for setup, operation, and changeovers, organizations introduce variability. Each operator runs a process slightly differently, leading to inconsistent performance and unpredictable quality outcomes. But SOPs without verified training are effectively aspirational documents. They describe what should happen. Without a systematic connection to the people who execute the process, there is no reliable mechanism ensuring that what should happen does happen.

This gap manifests across every regulated vertical, just through different regulatory lenses.

In pharmaceutical and biotech facilities, cGMP deviations that trace to process control failures often trace further to inadequate assurance that the personnel performing controlled processes were qualified to do so under the current version of the applicable procedure. Insufficient staff training has contributed to poor documentation practices and failure to meet quality control standards in documented regulatory cases, directly contributing to product recalls and production halts.

In food and beverage facilities, FSMA’s preventive controls framework requires that qualified individuals implement and verify preventive controls. But qualification is not a one-time event. When a food safety plan changes, when a new hazard is identified, or when a critical control point is added or modified, the personnel responsible for executing those controls need current training on the current requirements. Facilities whose training records can’t demonstrate currency with their food safety plan version have a compliance gap that FSMA inspectors and GFSI auditors will find.

In cannabis and hemp operations, GMP audit frameworks specifically examine whether staff training programs are functioning. Processors who document training without verifying competency — who record that training happened without confirming that the people trained can actually perform the relevant tasks correctly — have procedural compliance without operational reliability.

In automotive and aerospace, supplier quality failures often trace to process changes that weren’t accompanied by systematic retraining of the personnel affected by those changes. An IATF 16949 corrective action request for a supplier defect frequently discovers that the process specification had been updated, the quality system recorded the update, and the operators running the process had never been trained on what had changed. The system that managed the process change and the system that managed workforce training occupied separate spaces with no automatic connection between them.

In CDMO and CRO environments, the same problem takes a different form. Client study personnel may rotate across multiple programs, each with different SOPs, different quality requirements, and different competency standards. Without systematic management of who is qualified to work on which program under which version of which procedure, CDMOs face recurring instances of unqualified personnel performing restricted activities — not through negligence, but through the absence of a system that would have caught the gap before the work was done.

What Systematic Prevention Actually Looks Like

The organizations that maintain the lowest rates of preventable quality failures share a common characteristic that has nothing to do with the size of their quality teams or the sophistication of their procedures. They manage quality events and workforce competency in the same integrated system — where a change to a controlled document automatically creates a training requirement for every affected role, where a quality event automatically links to the competency status of the personnel involved, and where an auditor can trace the complete chain from document version to trained personnel to competency verification in a single report.

This is not a theoretical architecture. It is the practical difference between a quality system and a quality management system — between a collection of documented processes and an integrated framework in which those processes connect to each other and to the people responsible for executing them.

The practical impact of this architecture shows up in several specific ways.

Procedure changes reach the workforce.

When a document control system and a training management system share the same database, a document approval triggers training assignments automatically for every role mapped to that document. The effective date of the new procedure and the training requirement for affected employees are the same event, handled in the same transaction. No manual trigger, no coordinator dependency, no gap between when the procedure changes and when the workforce is trained on the change.

Quality events generate competency reviews.

When a deviation, nonconformance, or customer complaint enters the quality system, an integrated platform can immediately display the qualification status of the personnel involved in the affected process — which version of the relevant procedure they were trained on, when that training was completed, and whether competency was verified. This information, available at the initiation of an investigation rather than after a manual records pull, changes what investigators look for in root cause analysis and what corrective actions they specify.

Training currency is always visible.e

In a connected quality and training system, the training matrix for any process reflects the current procedure version, the required training modules, and the completion and competency status of every employee assigned to that process. A supervisor preparing for a production run can confirm that every operator on the line is qualified to the current SOP before the line starts. A quality manager preparing for an audit can generate a current qualification status report in minutes. An auditor can examine the same report and see a complete, coherent picture of workforce competency at any point in time.

Audit readiness is continuous, not episodic.

The manufacturing operations that manage quality events and training in disconnected systems spend significant pre-inspection time manually reconciling records — pulling training completion data from a separate LMS, cross-referencing names against quality event records, and building composite reports that neither system can generate independently. Organizations running integrated systems produce those reports on demand because the records were never separated in the first place.

The Industry-Specific Stakes

The financial and operational consequences of preventable quality failures vary by industry, but the direction is consistent across all of them: failures that could have been prevented by connecting quality management and training management cost significantly more than the systems required to prevent them.

For pharmaceutical manufacturers, a single consent decree can impose remediation costs exceeding $100 million, halt distribution of products representing hundreds of millions in annual revenue, and require third-party quality system oversight for years. The most common root cause of pharmaceutical enforcement actions — cGMP deviations tied to inadequate personnel training and qualification — is directly addressable through systematic quality-training integration.

For food and beverage companies, a contamination recall carries direct costs of product destruction, consumer notification, retailer penalties, and regulatory remediation — alongside reputational damage that affects brand equity long after the products are off shelves. The 2025 data showing 517 FDA food recalls at a nine-year high reflects an industry where the regulatory standard has risen, but quality infrastructure in many facilities has not kept pace.

For cannabis and hemp operations, the cost of a failed GMP audit is not just regulatory — it is commercial. Licensing delays, lost retail placement, and damage to customer confidence in a market where operators are still building credibility represent costs that are often existential for smaller operators. The facilities that invest in documented, verified training programs before their first GMP audit have a materially lower risk profile than those that assemble documentation in response to audit findings.

For automotive suppliers, a production shutdown triggered by a quality hold can generate customer penalties, expedited freight charges, and recovery costs that dwarf the original quality investment. IATF 16949’s emphasis on zero-defect supplier performance creates a supply chain environment where quality failures are not absorbed — they are charged back to the source, with contractual and commercial consequences that create sustained pressure on supplier quality economics.

For aerospace manufacturers, the cost calculation is different from all other industries because the consequences of failure in service include safety risk to human life. The rigor of AS9100 and NADCAP requirements reflects this. Organizations that invest in systematic qualification management — ensuring that every person performing a safety-critical process is currently trained on the current procedure and has demonstrated competency — are investing in both regulatory compliance and the operational integrity that the nature of aerospace products demands.

The Measurement Problem

One reason preventable quality failures continue at scale is that their full cost is systematically undermeasured. Organizations track the costs that surface through formal reporting mechanisms — recalls, regulatory actions, and major production incidents. The distributed costs — recurring rework, slow-burning deviation rates, inspection preparation overhead, supplier management friction — accumulate without being connected to the systemic gap that drives them.

Quality professionals who want to understand their organization’s true exposure to preventable failure costs should look at three metrics that typically go unmeasured.

Training currency at the time of quality events

For the last twelve quality events that resulted in deviations, CAPAs, or formal complaints: what was the training status of the personnel involved? Were they trained on the current version of the relevant procedure? Had their competency been verified? How long had it been since their last training on that process? Organizations that can answer this question quickly — because the information is in one integrated system — consistently discover patterns in the relationship between training currency and quality event frequency. Organizations that cannot answer it quickly are operating without intelligence about one of their primary quality risk factors.

Procedure change to training completion time

For the last ten controlled document revisions: how long elapsed between the effective date of the new procedure version and the date when all affected personnel completed training on the change? Organizations where this interval is measured in days operate differently — and with different quality outcomes — than organizations where it is measured in weeks or is simply unknown.

Recurring deviation root cause patterns

What proportion of quality events in the past 18 months share root causes with events from the prior 18 months? A rising recurrence rate in deviation root causes is the clearest leading indicator of a quality system that is generating documentation without driving improvement. Organizations with high recurrence rates almost always have a structural gap between quality event management and corrective training response — and that gap is the most cost-effective place to intervene. eLeaP’s guide on the cost of poor quality provides a framework for quantifying this exposure before it surfaces in a recall or regulatory action.

The Structural Fix

The recall data, the enforcement statistics, and the production cost research all point to the same conclusion. Regulated industries are not losing money to random quality failures. They are losing money to predictable failures at the intersection of process management and workforce competency — the point where what the quality system documents and what the workforce actually does come apart.

The structural fix is not more procedures. Organizations in every regulated industry already have extensive procedure libraries. The fix is not more training. Organizations that track training completions without verifying competency or connecting training to quality events are already generating training records that don’t prevent failures.

The fix is integration — a quality system architecture in which document control, training management, deviation tracking, CAPA management, and competency verification share the same data layer, so that a change in one automatically propagates to the others, and so that an auditor, a manager, or a quality professional can see the complete, current state of both quality events and workforce competency in a single view.

That architecture is available. The organizations that have built it are not immune to quality events — no organization is. But they catch quality signals earlier, connect them to competency gaps faster, and close the loop between corrective action and verified performance improvement in a fraction of the time it takes organizations still managing these functions in disconnected systems.

The 26% increase in defective units recalled in 2025 is not an inevitability. It is the aggregate result of quality systems that are partially functioning — that document requirements without reliably ensuring those requirements reach the people who execute them. The organizations that close that gap do not contribute to next year’s recall statistics. An eQMS that integrates document control, training, CAPA, and competency verification in a single system is how that gap closes.

eLeaP’s unified QMS and LMS platform connects quality management and training management in a single validated system — so that procedure changes, quality events, and corrective actions automatically link to workforce competency across pharmaceutical, biotechnology, medical devices, CDMOs and CROs, food and beverage, cannabis and hemp, automotive, aerospace, and manufacturing operations. Learn more at eleapsoftware.com.