European product safety is no longer defined only by whether an item passes a technical test before reaching a shelf. The current system increasingly depends on risk assessment, traceability, chemical controls, market surveillance, online-platform governance, corrective action and cross-border information exchange.
The General Product Safety Regulation provides the overarching framework for non-food consumer products that are not fully governed by more specific EU legislation. Safety Gate adds the operational surveillance layer, allowing national authorities to circulate information about products presenting serious risks.
The statistics show why the system must be viewed by product and risk type. Cosmetics account for 36% of 2025 Safety Gate alerts, while toys represent 16% and electrical appliances and equipment 11%. Chemical risks alone account for 53% of alerts, far ahead of injury and choking risks.
This report follows EU product safety from the legal framework through alert volumes, product categories, chemical hazards, online commerce, national enforcement, international cooperation and operational response. It then converts those signals into a benchmark index and a practical 90-day implementation model.
Executive EU Safety Benchmarks
The numbers defining Europe's current product-safety environment
The EU product-safety system operates at a scale that makes continuous surveillance a central part of compliance. Safety Gate recorded 4,671 alerts in 2025, up from 4,137 in 2024, 3,412 in 2023 and 2,117 in 2022. The latest total is equivalent to close to 100 alerts in an average week, and the network remains available across 365 operating days with seven-day coverage.
Detection is only one layer. Authorities reported 5,794 follow-up actions in 2025, more than the number of initial alerts. The annual reporting also describes more than 4,500 additional measures linked to follow-up activity, including recalls, withdrawal from sale, border intervention, sales bans and consumer warnings.
National authorities have participated in the modern Safety Gate network since 2003, but the surrounding framework has become more demanding. Online marketplaces now have defined product-safety responsibilities, the Consumer Safety Gateway gives consumers another reporting route, and the GPSR creates working-day deadlines for important actions.
|
Executive readout: EU safety performance should be evaluated as a complete surveillance system. Alert counts describe detection, while corrective actions, response deadlines, product categories and national enforcement reveal whether identified risks are actually controlled. |
Why EU Safety Standards Require a System-Based Benchmark
EU safety cannot be represented by a CE mark, a single regulation or a recall count. Product design determines whether a hazard is engineered out at the beginning. Risk assessment asks whether foreseeable uses and misuses create unacceptable exposure. Chemical compliance determines whether restricted or newly prohibited substances remain in materials and formulations.
This is why prevention, detection and correction need separate measures. Prevention covers design, testing and materials. Detection covers complaints, surveillance, consumer reporting and Safety Gate notifications. Correction covers listing removal, withdrawal, recall, border measures and verification that unsafe products do not reappear.
|
System readout: Strong EU product safety combines compliant design with traceability, active surveillance, rapid alerts and effective corrective action. A product that meets initial requirements but cannot be traced or removed efficiently remains a weak safety outcome. |
The General Product Safety Regulation and the New Compliance Baseline
How the GPSR changes the operating framework
Regulation (EU) 2023/988 became applicable on 13 December 2024, replacing the older general product-safety framework for many consumer products not fully covered by sector-specific legislation. The date is important because it marks the point at which new obligations became operational, but the wider implementation program continues beyond 2024 through formal reviews, system interconnection work and assessment of digital enforcement mechanisms.
Delegated powers associated with the regulation began from 12 June 2023. Parliament and Council scrutiny of delegated acts operates with a three-month objection period that can be extended by two months. Later milestones include reporting on Safety Gate interconnection in 2026, assessment related to online illegal-content removal in 2027 and a broader evaluation by 2029.
The practical change is equally important. The GPSR strengthens the expectation that products sold online are subject to the same safety logic as those sold through physical channels. It also reinforces the need for an identifiable responsible economic operator, clearer product information, structured incident handling and rapid marketplace response.
|
GPSR readout: December 2024 marks the start of application, not the end of implementation. Marketplace enforcement, system interconnection and formal evaluation continue through later regulatory milestones. |
Safety Gate Alert Growth Across Europe
From 2,117 alerts to 4,671 in three years
Safety Gate notifications have expanded sharply. Alerts increased from 2,117 in 2022 to 3,412 in 2023, then to 4,137 in 2024 and 4,671 in 2025. The 2025 level is 2,554 alerts above the 2022 baseline, meaning annual notification activity more than doubled across the three-year interval.
Several mechanisms can contribute to that rise. Authorities may be testing more products, national teams may be using Safety Gate more actively, online surveillance can expose unsafe offers that previously escaped notice, and enforcement may concentrate on product groups with known chemical or mechanical problems.

Figure 1. Safety Gate alert volumes rose sharply from 2022 through 2025, showing a substantially more active surveillance and notification environment.
|
Alert-growth readout: Rising alert volumes indicate a substantially more active safety-surveillance environment. The most useful interpretation combines notification growth with category, risk and national enforcement data. |
Alerts Versus Corrective Follow-Up Actions
Detection is only the first stage of safety enforcement
An alert identifies a serious risk and circulates it through the network; a follow-up action shows what authorities do after that information is available. In 2025, the system recorded 4,671 alerts and 5,794 follow-up actions. The fact that follow-ups exceed alerts is structurally important because dangerous products do not respect national borders.
For companies, the difference between alert and action changes the internal workflow. The first question is whether the notified product matches anything in the portfolio. The second is whether the same supplier, material, formulation or component appears in related products.
|
Enforcement readout: A strong safety system is measured by what follows an alert. Europe recorded more follow-up actions than initial alerts, illustrating how one notification can activate multiple national responses. |
Product Categories Most Frequently Triggering Safety Alerts
Cosmetics, toys and electrical equipment dominate notification volume
The 2025 product mix is highly concentrated. Cosmetics account for 36% of Safety Gate alerts, toys for 16% and electrical appliances and equipment for 11%. Together, these three groups represent 63% of reported alerts. The remaining 37% is distributed across a broad range of consumer goods, meaning the top categories deserve specialized control programs without allowing other sectors to fall outside routine monitoring.
Cosmetics lead because product safety can depend on formulation, restricted substances, contamination, labeling and direct exposure to skin or other parts of the body. A seemingly small formulation difference can determine whether a product is lawful, and chemical restrictions can make old inventory problematic after a regulatory change.
Toys combine vulnerable users with several possible hazard pathways. Chemical exposure can exist alongside choking, small-part, mechanical or labeling risks, and age grading changes what is considered foreseeable use. Electrical appliances and equipment present another pattern: electric shock, overheating, fire, batteries, insulation, chargers and construction quality can all drive serious-risk findings.

Figure 2. Cosmetics, toys and electrical products make up almost two-thirds of Safety Gate alerts, but each category requires a different technical-control strategy.
|
Product readout: Three product groups account for almost two-thirds of reported alerts, but their risk mechanisms differ enough that one uniform compliance process is unlikely to be sufficient. |
Chemical Risk as Europe's Leading Safety Hazard
Why more than half of alerts involve chemical risk
Chemical risk is the largest hazard category in the 2025 statistics, representing 53% of alerts. Injury risk accounts for 14% and choking 9%, while other risk types make up the remainder. Chemical risk alone therefore exceeds the combined share of injury and choking by a wide margin.
Chemical hazards can appear in many product groups. Cosmetics may contain prohibited fragrance ingredients or other restricted substances. Toys and plastics may create exposure through materials, coatings or softeners. Jewelry and accessories can raise issues around metals or skin contact.
The challenge is partly temporal. Mechanical dimensions rarely change because a law suddenly restricts a screw length that was acceptable yesterday, but chemical rules can make a previously used substance unacceptable or require tighter concentration controls. This creates risk in supplier specifications, legacy stock and imported goods.

Figure 3. Chemical risk dominates the 2025 Safety Gate risk mix, accounting for more than half of alerts and exceeding the next two listed risk types combined.
|
Risk readout: The EU safety challenge is increasingly chemical as well as mechanical. More than half of Safety Gate alerts identify chemical risk, making substance control central to product-safety management. |
BMHCA, Lilial and Restricted Cosmetic Chemicals
The cosmetic statistics show how one chemical issue can dominate enforcement. BMHCA, commonly known as Lilial, represented 77% of cosmetic chemical-risk alerts in the selected 2025 evidence and was associated with 1,278 alerts. The concentration demonstrates that once a substance is prohibited, old formulations or non-compliant imported products can continue to generate significant enforcement activity if supply chains do not update consistently.
TPO provides a second example. The dataset records 60 alerts after the September 2025 cosmetic ban period. A smaller number does not make the issue trivial; instead it illustrates how enforcement can respond quickly after a new restriction takes effect.
|
Chemical readout: Chemical compliance is not static. Once a substance is restricted, legacy stock, imported products and outdated formulations can continue generating alerts unless supply chains update quickly. |
Toys and Child-Focused Product Safety
Toys represent 16% of Safety Gate alerts, making them the second-largest product category in the 2025 mix. Their safety challenge is distinctive because the intended users are children and because several hazard mechanisms can coexist in a single item. A product can meet one chemical limit while still creating a choking hazard, or it can be mechanically robust while using an unsuitable material or carrying an incorrect age recommendation.
Follow-up statistics show a stronger corrective-action role as well. Toys represented 12% of follow-up actions in 2025 compared with 7% in 2024, an increase of 5 percentage points in the product-group share. That movement indicates more authority response once toy-related risks are circulated through the network.
Strong toy compliance joins chemical screening with mechanical testing, small-parts assessment, labeling, instructions, age grading and traceability. Supplier control matters because products with similar appearance can use different materials or components across batches. Marketplace surveillance matters because visually identical toys may be offered by sellers outside the authorized chain.
|
Toy readout: Toys combine vulnerable users with multiple risk mechanisms. Their growing share of follow-up actions makes post-market surveillance as important as pre-market testing. |
Electrical Appliances and Equipment Safety
Electrical appliances and equipment account for 11% of 2025 Safety Gate alerts. Their risks can include electric shock, overheating, fire, battery failure, unsuitable chargers, weak insulation, wiring faults and construction defects. These hazards tend to be technical and component-dependent, meaning a product that looks identical from the outside can have materially different safety performance depending on internal design or supplier substitutions.
Corrective follow-up also increased in relative importance. Electrical appliances and equipment represented 8% of follow-up actions in 2025 compared with 5% in 2024, a rise of 3 percentage points. The change indicates that electrical-product alerts increasingly trigger action after circulation, whether through withdrawal, recall, listing removal or additional national checks.
For manufacturers and importers, component control is essential. A tested reference sample is not enough if later production substitutes a power supply, battery cell, plug, cable or insulation material. Change-control procedures should therefore require technical review whenever a safety-critical component changes, even if the model name and packaging remain the same.
|
Electrical readout: Electrical products account for a smaller alert share than cosmetics or toys, but their growing follow-up share indicates increased authority response after hazardous products are identified. |
Motor Vehicles and Corrective-Action Intensity
Motor vehicles present the clearest example of why alerts and follow-ups should be analyzed separately. Vehicles account for 57% of follow-up actions in 2025, far above toys at 12% and electrical appliances and equipment at 8%. This does not mean vehicles create 57% of initial Safety Gate alerts.
Several features explain that pattern. Vehicles are high-value durable goods with traceable identification numbers, manufacturer-led recall systems and broad cross-border distribution. A single defect can therefore affect many units in many countries, and once the issue is confirmed, authorities and manufacturers can generate multiple national follow-up records tied to the same underlying problem.
The lesson applies beyond automotive products. Corrective-action intensity depends on distribution breadth, installed base, traceability and the ability to locate customers. A product category with strong identifiers may produce more recorded follow-up because action is easier to execute and document.

Figure 4. Motor vehicles dominate the composition of follow-up actions, illustrating the difference between initial risk notification and the later corrective-action footprint.
|
Follow-up mix readout: Motor vehicles dominate corrective follow-up even though cosmetics dominate initial notifications. This demonstrates why alert composition and enforcement composition should be analyzed separately. |
Online Marketplaces and Digital Product Safety
Safety standards now extend directly into e-commerce infrastructure
Product safety now depends on digital infrastructure because a substantial share of consumer goods is discovered and purchased online. The Product Safety Pledge+ includes 12 marketplace providers and 19 voluntary commitments in the current evidence set. These commitments sit alongside formal GPSR obligations and illustrate the growing expectation that platforms should monitor dangerous products, cooperate with authorities and make it harder for removed items to reappear under slightly altered listings.
The Consumer Safety Gateway adds another signal. More than 100 notifications were submitted in its first year, providing a route for consumers to report products they believe are unsafe. This does not replace technical investigation, but it broadens the detection network by adding evidence from real use and purchase channels.
|
Digital readout: Product safety increasingly depends on how rapidly platforms can identify, trace and remove unsafe listings, not only whether physical products were tested before sale. |
GPSR Marketplace Response Deadlines
Why product-safety response is measured in working days
The GPSR converts product-safety response into short operating deadlines. Online marketplaces may have as little as 2 working days to act on authority removal orders, while product-safety notices are expected to be processed within 3 working days. Several authority and network notifications operate on a 4-working-day timetable.
The deadlines change the design of compliance systems. Product and seller identifiers must already be searchable. Contact points need to be current. Legal, marketplace, quality and customer-service teams need a defined escalation route. Evidence of decisions and actions must be retained so the organization can demonstrate what happened within the required period.
Response-time testing is therefore useful before an incident. A mock removal order can reveal whether teams can identify every affected listing in hours, whether weekend coverage exists, whether outsourced marketplace operations know who owns safety decisions and whether the company can document removal across multiple platforms.

Figure 5. Key GPSR and Safety Gate processes operate on short working-day deadlines, making rapid data retrieval and escalation essential compliance capabilities.
|
Deadline readout: EU digital-product enforcement works on a two-to-four-working-day response cycle. Companies that need weeks to identify a seller, batch or listing are structurally misaligned with the regulatory timetable. |
Consumer Safety Gateway and Reporting Infrastructure
Direct consumer reporting strengthens product surveillance by adding evidence that may not yet appear in manufacturer complaint systems or national inspection programs. The Consumer Safety Gateway received more than 100 notifications in its first year. That volume is small compared with the thousands of formal Safety Gate alerts, but its importance lies in the source of the information: people encountering products in real use.
Consumer reports can provide product names, purchase channels, seller information and descriptions of unsafe behavior. When these fields are detailed, they can support faster triage and help authorities decide whether further testing or enforcement is warranted. When identifiers are weak, the report becomes harder to connect to a specific economic operator.
Manufacturers and retailers should treat their own complaint systems in the same way. Free-text customer service data can be categorized into safety signals such as overheating, breakage, small parts, skin reactions or unexpected chemical odor. Repeated low-frequency signals may justify investigation before a regulator publishes an alert.
|
Consumer-reporting readout: Safety intelligence no longer flows only from manufacturers to authorities. Consumer reporting adds an additional detection channel that can reveal risks after products enter real use. |
Coordinated Activities on the Safety of Products
Testing and surveillance beyond individual national actions
European product surveillance is strengthened through coordinated activities that allow authorities to test products and develop common enforcement methods. The cumulative evidence includes 30 product-specific sampling and testing activities and 18 horizontal market-surveillance activities, creating a combined base of 48 coordinated initiatives.
Product-specific projects focus on defined groups of goods, allowing authorities to compare test results and identify recurring non-compliance. Horizontal projects address issues that cut across categories, such as online surveillance, risk assessment or enforcement practice. These activities improve institutional consistency even when the products under review differ materially.
For companies, coordinated projects offer useful intelligence because they reveal what authorities are choosing to examine and how tests are structured. A business that sells into several Member States should not wait for a direct inspection before learning from recurring findings in its category.
|
CASP readout: Coordinated EU testing creates a shared evidence base. Product-specific checks identify immediate failures, while horizontal activities improve surveillance methods across categories. |
Product Safety Award and Safety Innovation
The safety framework also contains a positive innovation dimension. The 2025 EU Product Safety Award attracted 33 applications from 18 countries. The initiative began in 2019 and is organized every 2 years, creating a recurring platform for recognizing approaches that go beyond minimum compliance.
Recognition also balances the report's enforcement statistics. Alerts and recalls show where systems fail; award participation shows that firms and organizations are actively developing better practices. A mature safety culture needs both signals: strong consequences for dangerous products and incentives for prevention.
|
Innovation readout: European safety policy combines enforcement with recognition. Participation from 18 countries demonstrates that product safety is also becoming a competitive design and innovation discipline. |
Country-Level Safety Gate Alert Activity
Where notification activity is concentrated
National Safety Gate activity is highly uneven. Italy notified 1,193 alerts in 2025, the largest total among the listed EU/EEA participants in the workbook. Germany followed with 465, France with 455, Sweden with 384 and Czechia with 359. Together, these five countries account for a substantial share of the listed-country subtotal, showing that the network's originating notification activity is concentrated rather than evenly distributed.
These totals must be interpreted carefully. A high alert count does not mean products manufactured in the notifying country are unusually unsafe. Authorities notify products found in their markets regardless of where those products were made. High totals can reflect inspection volume, testing priorities, customs work, specialist expertise, market size, online-surveillance activity or simply a more active reporting strategy.

Figure 6. Italy led 2025 Safety Gate notification activity by a wide margin, followed by Germany, France, Sweden and Czechia.
|
Country |
Alerts 2024 |
Alerts 2025 |
Change |
2025 share |
|
Italy |
1,089 |
1,193 |
+104 |
28.1% |
|
Germany |
471 |
465 |
-6 |
10.9% |
|
France |
315 |
455 |
+140 |
10.7% |
|
Sweden |
386 |
384 |
-2 |
9.0% |
|
Czechia |
332 |
359 |
+27 |
8.4% |
|
Hungary |
365 |
317 |
-48 |
7.5% |
|
Poland |
170 |
170 |
+0 |
4.0% |
|
Finland |
90 |
105 |
+15 |
2.5% |
|
Country-alert readout: National alert counts are indicators of surveillance activity rather than product-quality league tables. Strong notification numbers can reflect active detection and enforcement. |
Italy's Role in European Product-Safety Notifications
Italy's 1,193 alerts make it the clearest outlier in the 2025 country distribution. Based on the listed EU/EEA subtotal in the research workbook, Italy represents about 28% of alerts. Germany's 465 and France's 455 are both large totals, yet Italy notified more than twice as many alerts as either country individually.
The number should not be converted into a claim about Italian manufacturing quality. Notifier geography refers to the authority that entered the alert, not necessarily the country where the product was made. An Italian authority can notify a product manufactured outside Europe, sold through an international marketplace or distributed by a company based in another Member State.
For analysts, Italy is therefore a surveillance signal rather than a country-risk label. The useful questions are which categories and hazards appear in those notifications, whether the same suppliers or marketplaces recur and how often other countries take follow-up action.
|
Italy readout: Italy's exceptionally large notification volume gives it disproportionate influence over the EU alert total, but notifier country must not be confused with manufacturing origin. |
Germany, France, Sweden and Czechia
Germany and France form the second tier of 2025 notification activity, with 465 and 455 alerts respectively. Their totals differ by only 10 alerts, making them a useful large-market comparison. Sweden follows with 384 and Czechia with 359, showing that high reporting activity is not limited to the EU's largest consumer economies.
The Swedish and Czech figures are particularly useful for understanding why country rankings should not simply mirror population or GDP. Authority priorities, specialized testing and reporting practices can produce high notification volumes in smaller markets. This is a feature of a networked surveillance system: concentrated expertise in one country can produce information that benefits consumers elsewhere.
|
High-activity readout: European notification activity is concentrated but not limited to the largest consumer markets. Sweden and Czechia show that enforcement intensity can produce high reporting volumes outside the largest economies. |
Countries With the Largest Changes in Alert Activity
Year-over-year changes show how quickly national notification activity can move. Absolute change is the most useful first measure because a large percentage increase from a very small baseline can appear dramatic while adding only a handful of alerts. Percentage change remains useful as a second measure because it identifies countries whose surveillance output is changing rapidly relative to their own previous level.
These movements can result from changing inspection campaigns, staffing, product-category focus, major cases, reporting practices or the timing of coordinated surveillance projects. A one-year fall therefore should not automatically be interpreted as improved product safety, just as a one-year rise should not automatically be read as deteriorating product quality.
For businesses, change data identifies where monitoring priorities may be shifting. A country with rising alert activity may become a more important early-warning market, especially if the increases cluster in a relevant product category. A large decline can also warrant examination to determine whether the authority focus has shifted rather than disappeared.

Figure 7. The largest absolute increases and decreases show how national notification activity can shift materially from one year to the next.
|
Change readout: Year-over-year changes should be read alongside starting volume. A large percentage increase from a small baseline does not carry the same system impact as a moderate increase in a high-volume notifier. |
Follow-Up Actions by Country
Which national systems generate the strongest corrective response
Follow-up rankings reveal a different national map from alert rankings. Germany recorded 639 follow-up actions in 2025, while Denmark recorded 522 and Bulgaria 417. Several other countries also show substantial response activity even when their own alert totals are much lower.
Bulgaria is an instructive example. Its 2025 alert count is low compared with the leading notifiers, yet its follow-up count is high. Denmark shows another strong response profile. Such patterns should not be described as automatically better or worse. They indicate a national role with relatively more corrective action in response to circulating alerts.
Germany combines high performance on both dimensions, ranking near the top for alerts and leading follow-up actions. That makes it a broad surveillance-and-response node in the network. Italy, by contrast, leads alert origination by a wide margin but records far fewer follow-ups in the workbook.

Figure 8. National follow-up activity is led by a different group of countries than initial notifications, underscoring the network effect of Safety Gate enforcement.
|
Country |
Alerts 2025 |
Follow-ups 2025 |
Follow-ups per alert |
|
Germany |
465 |
639 |
1.37 |
|
Slovenia |
16 |
557 |
34.81 |
|
Denmark |
18 |
522 |
29.00 |
|
Poland |
170 |
419 |
2.46 |
|
Bulgaria |
11 |
417 |
37.91 |
|
Ireland |
94 |
369 |
3.93 |
|
Luxembourg |
31 |
338 |
10.90 |
|
Finland |
105 |
335 |
3.19 |
|
Country-enforcement readout: Notification leadership and follow-up leadership are not the same. Comparing both metrics reveals whether countries primarily detect risks, respond to alerts raised elsewhere, or perform both functions. |
Alerts Versus Follow-Ups: National Enforcement Profiles
Plotting alerts against follow-up actions separates national participation into several profiles. High-alert, high-follow-up countries are active in both originating serious-risk notifications and responding to alerts raised elsewhere. High-alert, lower-follow-up countries contribute strongly to detection. Lower-alert, high-follow-up countries appear more heavily in the corrective-response layer.
These profiles are descriptive rather than normative. Follow-up counts depend on whether products notified elsewhere are sold locally, while alert counts depend on national testing and reporting. Market size, category mix and authority strategy all influence where a country appears.
The framework is useful for company planning because it highlights where different operational risks arise. High-notification countries can provide early warning on product failures. High-follow-up countries can create rapid corrective obligations when a foreign alert affects local stock. Companies with pan-European distribution need processes capable of responding to both signals.

Figure 9. Alerts and follow-up actions reveal different national roles in the Safety Gate network; selected outliers show how strongly detection and response can diverge.
|
Enforcement-profile readout: Countries participate differently in the same network. Some primarily originate alerts, others generate large volumes of corrective responses, and a smaller group is highly active on both dimensions. |
Northern Ireland in the Safety Gate Network
Northern Ireland is useful for understanding why network totals and standard EU/EEA country tables must be handled carefully. The 2025 reporting includes 420 alerts associated with Northern Ireland and 9 follow-up actions. Because its institutional treatment differs from a standard EU Member State entry, the research workbook separates it when calculating the listed-country subtotal and national shares.
This distinction is more than a spreadsheet detail. If Northern Ireland is included in one total but excluded from another, country shares and rankings can become inconsistent. Analysts should therefore define the denominator used for each chart and table and avoid silently mixing network participation with EU-membership categories.
The same rule applies to any future expansion or special participation arrangement. Safety Gate is an operational network, while political and legal membership categories define a different classification. The report should preserve both concepts without forcing them into one table.
|
Network readout: Safety Gate analysis must distinguish institutional participation from EU-member-country rankings so network totals and national subtotals are not accidentally mixed. |
International Cooperation and Imported Product Safety
European product safety operates inside global supply chains. Products manufactured outside the EU can enter through traditional importers, direct-to-consumer sellers or online marketplaces, and the same product may be distributed to several Member States before a defect is identified. International information exchange is therefore essential when the source of supply lies beyond the jurisdiction that first detects the risk.
Imported-product safety depends on three types of visibility. Authorities need to identify the product and economic operator. Businesses need to identify the manufacturer, supplier and affected batch. Marketplaces need to identify the seller and every active listing. If one of these links is missing, the time between detection and effective removal increases.
|
International readout: EU safety standards increasingly depend on information exchange beyond EU borders because manufacturing, distribution and online sales routinely cross jurisdictions. |
Safety Standards by Product Lifecycle Stage
A lifecycle view converts regulatory requirements into operating controls. During design, hazards should be reduced before they become testing failures. During supplier selection, material and component evidence should be verified. Pre-market testing should confirm technical conformity, while documentation should preserve the product identifiers and economic-operator information needed later.
After sale, the safety system shifts toward monitoring. Complaints, returns, marketplace reviews and consumer reports can reveal emerging issues. Incident escalation determines whether the problem requires technical investigation or regulatory notification. Corrective action determines whether affected products are withdrawn, recalled or removed online.
The stages are connected. Poor supplier records make chemical investigations slower. Weak product identifiers make recalls less precise. Missing marketplace ownership can cause a removal deadline to be missed. Conversely, strong traceability improves every later stage because the company can determine where affected units were sold and which customers or sellers need to be contacted.
|
Stage |
Main control |
Failure signal |
Useful metric |
|
Design |
Risk reduction |
Hazard remains inherent |
Design-risk findings |
|
Supplier |
Material verification |
Restricted substance |
Supplier failure rate |
|
Testing |
Standards compliance |
Failed test |
Test pass rate |
|
Documentation |
Traceability |
Missing information |
Documentation completeness |
|
Market entry |
Listing control |
Unsafe offer |
Rejected listings |
|
Monitoring |
Complaint/incident review |
Recurring incident |
Complaint trend |
|
Corrective action |
Withdrawal/recall |
Slow response |
Days to action |
|
Verification |
Follow-up check |
Product remains available |
Reappearance rate |
|
Lifecycle readout: Safety is strongest when controls begin before market entry and continue after sale. Regulatory compliance is a lifecycle process rather than a one-time certification event. |
Building the EU Safety Standards Benchmark Index
Turning regulatory evidence into a weighted performance model
The EU Safety Standards Benchmark Index converts the report into eight weighted pillars. Risk assessment and hazard control receive 18%, the largest single weight, because safety begins with identifying and reducing foreseeable hazards. Chemical and material compliance receive 16%, reflecting the dominant 53% chemical-risk share in Safety Gate alerts and the importance of rapidly changing restricted-substance requirements.
Traceability and documentation receive 14%, matched by 14% for product testing and technical conformity. The equal weighting reflects two complementary needs: the organization must know that the product meets technical requirements and must also be able to identify the affected product, batch, supplier and economic operator when something goes wrong.
Corrective-action and recall capability receive 10%, digital marketplace compliance 9% and consumer reporting and transparency 7%. These lower individual weights should not be interpreted as optional. A product can score strongly in design and testing yet still create a weak overall safety outcome if the company misses a removal order or cannot contact customers during a recall.

Figure 10. The benchmark gives the greatest weight to hazard control and chemical compliance while preserving substantial weight for traceability, testing, surveillance and corrective action.
|
Pillar |
Weight |
Evidence to score |
|
Risk assessment |
18% |
Hazard identification and mitigation |
|
Chemical compliance |
16% |
Restricted-substance controls |
|
Traceability |
14% |
Batch/operator identification |
|
Testing |
14% |
Technical conformity results |
|
Surveillance readiness |
12% |
Monitoring and escalation |
|
Corrective action |
10% |
Withdrawal/recall speed |
|
Digital compliance |
9% |
Marketplace control |
|
Transparency |
7% |
Consumer information/reporting |
|
Index readout: A high safety score should not come from testing alone. Strong performance requires hazard control, documentation, traceability, corrective-action speed and digital-market readiness to work together. |
EU Safety Compliance Challenges
The first structural challenge is scale. Online commerce allows products from thousands of sellers to enter many national markets quickly. Traditional compliance models were built around identifiable importers and distributors, while modern marketplaces can contain independent sellers whose product data changes frequently.
The second challenge is traceability. Third-party sellers, contract manufacturing and multi-tier supply chains create gaps in economic-operator and batch information. These gaps become expensive during incidents because teams spend the critical first hours reconstructing who supplied the product, where it was sold and what inventory remains.
Changing chemical restrictions create a third challenge. A formulation that was historically accepted can become non-compliant, and old stock can remain in warehouses, distributors or online channels after the restriction changes. Supplier declarations must therefore be linked to a regulatory change process rather than stored as static documents.
|
Challenge readout: The main difficulty is no longer knowing that safety rules exist. It is maintaining traceability, current chemical controls and rapid corrective-action capability across complex physical and digital supply chains. |
90-Day EU Safety Standards Compliance Plan
Days 1 to 30 should establish the baseline. Catalogue every product category, applicable legislation, relevant standards, technical file, supplier declaration, responsible economic operator, batch identifier and online listing. Map restricted substances and connect existing complaints or incidents to the products involved.
Days 31 to 60 should test the controls. Select representative products for sample verification, audit documentation completeness, review chemical evidence and examine whether high-risk materials receive appropriate screening. Check marketplace listings against the approved product record and run a mock Safety Gate escalation.
Days 61 to 90 should simulate enforcement. Test whether a two-day marketplace removal order can be completed, whether a three-day safety notice can be investigated and whether four-day notification workflows are operational. Trace a selected batch from supplier through warehouses, listings, distributors and customers.
The final output should be a prioritized remediation plan, not simply an audit score. Critical gaps include missing responsible-operator information, inability to trace batches, outdated chemical declarations, unclear marketplace ownership and response times that exceed regulatory windows. These issues should be corrected before lower-impact formatting or documentation improvements.
|
Period |
Primary objective |
Key output |
|
Days 1–30 |
Know the product and supply chain |
Complete baseline |
|
Days 31–60 |
Verify controls |
Testing and audit findings |
|
Days 61–90 |
Prove response capability |
Simulated enforcement readiness |
|
90-day readout: The purpose is not merely to produce more compliance documents. It is to prove that the organization can identify an affected product and act within the same short timetable expected by EU regulators. |
Metrics Manufacturers and Retailers Should Track
Product-risk metrics should track test failures, chemical non-conformance, design-risk findings, incidents and injuries. Traceability metrics should measure documentation completeness and the time required to identify affected units. Marketplace metrics should cover unsafe listings, removal time and repeat appearances, while enforcement metrics should track corrective-action and recall timing.
Traceability metrics should measure the share of products with complete batch identification, supplier-document completeness, responsible-operator coverage and the time required to identify affected units. The most useful test is elapsed time. A record can be technically complete yet operationally weak if teams need several days to assemble it from separate systems.
Marketplace metrics should include unsafe listings detected, removal time, seller-response time and reappearance of products that were previously removed. Enforcement metrics should add corrective-action initiation, withdrawal completion, recall completion and response to authority contacts. Consumer metrics should track safety complaints, incident notifications, recall-contact rates and completion of refunds or replacements.
These measures create a leading indicator system. Sales and return rates can remain healthy while safety capability is degrading. Rising document gaps, slower identification time or repeated restricted-substance findings can reveal weakness before a major recall occurs. The best scorecard therefore combines product outcomes with system response measures.
|
Metric family |
Core KPI |
Strong signal |
Warning signal |
|
Testing |
Pass rate |
Stable high conformity |
Rising failures |
|
Chemicals |
Restricted-substance failure |
Near zero |
Repeated detection |
|
Traceability |
Identification time |
Hours |
Days/weeks |
|
Digital |
Listing-removal time |
Within required window |
Missed deadline |
|
Corrective action |
Recall initiation |
Rapid |
Delayed |
|
Consumer safety |
Incident trend |
Stable/declining |
Accelerating |
|
Scorecard readout: Revenue and sales volume describe commercial performance. Test failures, traceability speed, restricted-substance findings and corrective-action timing describe whether the product-safety system is functioning. |
How EU Safety Responsibilities Change by Business Model
Manufacturers have the greatest influence over design, materials, technical documentation, risk assessment and production control. Their evidence should show not only that a reference sample was compliant but that safety-critical components and formulations remain controlled across production. Change management is essential when suppliers, components or chemicals are substituted.
Importers and distributors occupy different positions but still need practical verification. Importers should ensure the necessary product and economic-operator information exists before goods enter the market. Distributors need enough visibility to recognize obvious compliance gaps and to react when a product is identified as unsafe.
Online marketplaces increasingly form part of the safety infrastructure. They need mechanisms for authority communication, seller traceability, notice processing, listing removal and prevention of reappearance. Brands and retailers add product selection, consumer communication and recall execution. Testing laboratories provide technical evidence, but a test result cannot replace the economic operator's wider responsibility for traceability and corrective action.
National authorities inspect, test, notify and coordinate, while consumers contribute incident signals through complaints and reporting channels. The value chain is therefore interdependent. Strong raw technical compliance can be undermined by weak marketplace controls, and excellent retail operations cannot compensate for an inherently unsafe design.
|
Business-model readout: Product safety is distributed across the value chain. A compliant manufacturer can still be undermined by weak marketplace controls, while strong retailers cannot compensate for unsafe product design. |
The EU Safety Standards Report FAQ
What is Safety Gate?
Safety Gate is the EU rapid alert system for dangerous non-food products. National authorities use it to circulate serious-risk information so other countries can check whether the same product is present in their markets and take appropriate action.
How many Safety Gate alerts were recorded in 2025?
The 2025 total was 4,671 alerts. That compares with 4,137 in 2024, 3,412 in 2023 and 2,117 in 2022, showing a substantial increase in annual surveillance and notification activity.
How many follow-up actions were recorded?
Authorities recorded 5,794 follow-up actions in 2025. Follow-ups can exceed alerts because one notification may trigger corrective activity in several countries where the product has been sold.
Which product category generated the largest share of alerts?
Cosmetics were the largest product group at 36% of alerts. Toys represented 16% and electrical appliances and equipment 11%.
What was the most common risk?
Chemical risk was the largest risk category at 53% of alerts. Injury represented 14% and choking 9%, making chemical control a central part of the current EU product-safety picture.
Why do some countries report more alerts than others?
National totals reflect testing activity, market size, inspection priorities, specialist expertise, customs work, online surveillance and reporting practices. They should not be treated as a direct ranking of the safety of products manufactured in each country.
Does a high alert count mean products made in that country are unsafe?
No. The notifying country is the authority that entered the alert. The product may have been manufactured elsewhere and simply detected in that national market.
How quickly must online marketplaces act?
Important GPSR processes operate within short working-day windows. Marketplace action on authority removal orders can be required within 2 working days, while product-safety notices may need processing within 3 working days.
What is the role of the Consumer Safety Gateway?
The gateway gives consumers a direct route to report products they believe are unsafe. More than 100 notifications were recorded in its first year, adding another source of post-market safety intelligence.
What should companies monitor internally?
Companies should track product testing, chemical non-conformance, batch traceability, supplier documentation, complaints, marketplace listings, incident trends and the time required to initiate corrective action or recall.
Final Takeaway
The EU product-safety system operates at unprecedented notification scale. Safety Gate recorded 4,671 alerts and 5,794 follow-up actions in 2025, with close to 100 alerts in an average week. The rise from 2,117 alerts in 2022 shows how rapidly surveillance output has expanded.
The risk mix is equally important. Cosmetics represent 36% of alerts, toys 16% and electrical equipment 11%, while chemical hazards account for 53%. BMHCA shows how one restricted substance can generate extensive enforcement when legacy, imported or non-compliant products remain in circulation.
Digital commerce adds a speed requirement to the technical framework. Marketplace action can be required within 2 working days, notice processing within 3 and several authority notifications within 4. These deadlines turn traceability, seller records and internal escalation into measurable safety controls. Together, these controls determine whether unsafe goods are prevented, traced and removed quickly across markets.