Center for Pet Safety Tested Pet Bags: Buyer Compliance Guide

Wholesale pet bag sourcing desk · Updated 2026-10-06 · 14 min read

Orders for bulk pet carriers are quoted on the same three levers as Material & Technology work - fabric weight, hardware grade and carton count - because freight, not unit labour, decides the landed figure. Buyers who buy pet carriers in mixed sizes can consolidate several size runs into one booking, which keeps the per-unit spread tight without forcing a single colourway to carry the whole minimum.

What a Third-Party Pet Safety Programme Actually Is

A third-party safety programme in this category is an independent organisation that publishes test criteria, tests products against them, and publishes the results. It is not a regulator, it does not issue a legal approval, and participation is normally voluntary and paid for by the manufacturer or brand. What it provides is comparability: two products evaluated against the same published criteria can be compared in a way two supplier-generated reports cannot.

The most recognised programme for pet containment is run by the Center for Pet Safety, which publishes its protocols and its results and is widely cited by retailers and by informed consumers. Buyers should read the protocol rather than the summary, because the value is in the criteria and the exclusions, both of which determine whether a particular product can pass and what a pass actually means.

Three features distinguish a credible programme from a vanity one. The criteria are published before testing rather than after. The results, including failures, are published rather than only the passes. And the tested product is identified specifically by model so a buyer can confirm that the listed product is the one being offered.

For a wholesale buyer the practical question is whether the programme's criteria match the product's intended use. A protocol written for in-cabin airline carriers addresses a different set of risks from one written for car travel, and a pass in one does not imply performance in the other. Matching the protocol to the claim is the first decision, and it should be made in writing before any design work starts, because everything downstream - geometry, materials, hardware and the instruction set - follows from it.

Buyers should also understand what a published programme does not do. It does not guarantee acceptance by an airline, a retailer or an insurer, and it does not transfer regulatory obligations to the organisation. It provides evidence, and the buyer still decides how that evidence is used and described.

Reading the Protocol: Criteria, Exclusions and Failure Modes

A protocol is a specification, and it should be read the way a buyer reads a tech pack: clause by clause, looking for the criteria that will be hardest to meet. Three groups of clauses matter most in pet containment protocols - containment and structural criteria, restraint and excursion criteria, and constructional requirements such as ventilation, closure type and labelling.

• Containment criteria: the product must retain the test device or mass, with no breach of the closure or structure.
• Restraint criteria: where vehicle use is in scope, the product must stay within defined limits when installed as instructed.
• Constructional criteria: ventilation area, closure type, escape-resistant features, and the presence and clarity of instructions.
• Exclusions: size, weight or product-type limits that determine whether a given SKU is eligible at all.

Exclusions are where programmes are lost. A protocol may limit eligibility by animal weight band, by product type or by construction, and a buyer who designs a range before reading the exclusions can find that most of it is ineligible. Reading them first and designing to them is far cheaper than redesigning later.

Instruction and labelling clauses are frequently the cheapest to satisfy and the most often missed. A protocol that requires a specific statement, a diagram or a warning on the product is asking for a printed element, and the cost of compliance is a label. Missing it fails the product for the want of a few square centimetres of print.

Device or test-mass clauses are the second group that deserves early attention. A protocol will specify what is placed inside the product and how, and a design that performs well with a compliant mass can behave differently with a live animal that shifts, braces and pushes. Reading the clause tells a buyer what the pass really certifies, and that understanding should shape both the design and the claim.

Certification vs Testing: What a Buyer May Say

Programme terminology is specific and buyers should use it precisely. Where a programme awards a certification or a listing, the product may be described using that programme's own terminology, normally subject to the programme's rules and often to a licence covering the mark. Where a buyer commissions an independent test that happens to use a published protocol, the accurate description is that the product was tested to that protocol - not that it is certified by the organisation, unless the organisation has actually certified it.

That distinction protects both sides. A buyer who claims certification without participating misuses the organisation's name and creates a claim that a retailer will check, because third-party marks are among the easiest claims to verify. A buyer who says tested to a named protocol, and can produce the report, is on solid ground.

Buyers should also check whether the programme permits use of its name and mark in marketing at all, and under what conditions. Some require specific wording, some restrict the mark to listed models, and some require the year of testing to be stated.

How Suppliers Usually Fail a Protocol Submission

Three failure patterns account for most unsuccessful submissions. Submitting a prototype rather than a production-representative unit, which makes the result non-transferable. Changing a listed component between test and bulk without assessment. And relying on a material upgrade to solve what is really a geometry problem, such as a closure that can be worked from inside. All three are avoidable with a clause-by-clause review before submission, which is why our production team runs one against the prototype rather than waiting for the laboratory to find the gaps.

Building a Product to the Criteria Rather Than to a Price

Designing to a published protocol changes the engineering conversation in a useful way, because the criteria are fixed and the trade-offs are explicit. Closure strength, anchorage reinforcement, frame stiffness and ventilation area stop being preferences and become thresholds. That usually raises the specification and the unit cost, and it also removes the argument about whether the extra cost was necessary.

Four design areas typically need attention for containment protocols. Closure integrity, because a protocol will test whether the animal can work the opening. Structural integrity under load, because panels and frames must not collapse inward. Ventilation area, because a minimum is often specified and it competes with structural reinforcement. And interior smoothness, because exposed edges and loose liners are both escape aids and failure points.

Ventilation is the one that reliably conflicts with everything else. Reinforcing a panel reduces open area, and meeting both usually means geometry rather than compromise: baffled vents, gusseted openings and framed mesh panels deliver airflow without creating a weak panel. Solving it geometrically is what separates a compliant product from a reinforced one that fails on airflow.

Interior smoothness, the fourth area, is worth expanding because it is where cheap decisions pay off disproportionately. Exposed seam allowances, loose liners, hanging labels and accessible zipper pulls are all escape aids, and each is removed by a construction decision costing cents rather than by a stronger material costing dollars. Protocols that test whether an opening can be worked from inside reward exactly this kind of detailing.

Cost lands where buyers expect once it is modelled honestly: reinforcement and rated hardware add a modest amount per unit, while the fixed programme and laboratory cost is what determines whether the volume supports the exercise at all.

Cost, Volume and Whether to Pursue the Programme

The decision to pursue a third-party programme is a channel decision before it is an engineering one. Specialty pet chains, veterinary-adjacent distribution and premium marketplaces recognise the benchmark and reward it with listing quality and reduced friction. Mass retail and low-margin marketplace channels generally do not, and a programme funded for those channels rarely recovers its cost.

The cost structure has the familiar two-part shape. The fixed part covers programme participation, laboratory time, reporting, the design iteration needed to pass and any licence fee attached to using the mark; the variable part covers the reinforced construction that recurs on every unit. Amortising the fixed part over a committed annual programme is what makes the numbers work, which is why these programmes are quoted against stable volumes rather than spot orders.

Iteration is the budget line buyers underestimate. A first submission frequently fails on a detail - a label, an instruction, a closure that can be worked from inside - and each cycle costs time and money. Reviewing the protocol clause by clause against the prototype before submission is the cheapest way to reduce cycles.

A comparable alternative exists for buyers who do not need the mark: commission an independent test to the published protocol and describe the result accurately as tested to it. That captures most of the engineering value and much of the marketing value at lower cost, and it is the right answer for programmes where the listing benefit is uncertain. Many buyers run the independent test first, prove the commercial return in one season, and then commit to the full programme with evidence that it pays.

The fixed part also includes internal time: protocol review, prototype iteration, retailer questionnaires and claim approval. On a first programme that internal cost often exceeds the laboratory invoice, and budgeting for it honestly is what prevents the mid-programme squeeze that leads teams to skip verification steps.

Retailer and Marketplace Recognition

Recognition is the commercial payoff, and it arrives in three forms. Line review is faster because the technical buyer recognises the benchmark and does not need to evaluate a bespoke criterion set. Listing content is stronger because a specific, checkable claim converts better than a general one. And complaint handling is faster because the file already answers the question.

Documentation should be prepared for upload, not just for filing. Retailers and marketplaces typically want the certificate or report, the model identification matching the listing, the date, and permission to use any mark. Having those four as separate named files shortens onboarding considerably. Buyers should also price the onboarding work itself - documentation uploads, wording rules and review cycles differ per channel - because it sits between the test and the first sale and is where a launch usually slips.

Model identification is the item that most often goes wrong. A listing describing a product as tested when the tested model is a different size, a different colourway or a pre-production revision is inaccurate, and it is easy for a reviewer to check. Buyers should confirm that the tested model identifier matches the SKU being listed, and should re-check after any revision. A simple control helps: keep one record mapping every listed SKU to the tested model identifier, and review it whenever the range is extended or revised.

General safety obligations apply regardless of programme participation. In the US the consumer product regime is administered by the Consumer Product Safety Commission, and chemical documentation for the EU market sits under REACH as administered by ECHA. A third-party listing does not replace either.

Recognition is rarely a single decision. Most retailers accept a third-party protocol listing as supporting evidence rather than as a substitute for their own safety questionnaire, and several require the test report to be re-issued within 24 months. Budget for re-testing at that interval when modelling programme cost, because the listing fee is not the only recurring expense.

Air Travel, Car Travel and Protocol Scope

Protocol scope is narrower than buyers often assume, and matching scope to claim is the discipline that prevents most problems. A protocol written for in-cabin air travel addresses ventilation, containment under handling and fit under an aircraft seat. A protocol addressing car travel addresses restraint, excursion and impact behaviour. A pass in one says nothing about the other.

Air carriage requirements themselves are set by airlines, shaped by the Live Animals Regulations maintained by IATA, with public guidance available from the FAA for travellers flying with pets. Those rules are airline-specific in detail and change periodically, so a third-party pass does not guarantee acceptance on a particular carrier. Buyers selling into travel should keep airline requirements in the same file as the protocol, because the two answer different questions and a customer will ask both.

Scope creep is the associated risk. A product tested for car travel is frequently marketed for air travel because the photography suits it, and that extension of the claim is not supported by the test. Keeping one approved-claims list per SKU, checked before every launch, is the control that prevents it.

Car travel sits closer to the vehicle-safety vocabulary discussed elsewhere in this series. Where a product is marketed for vehicle use, buyers should consider the dynamic criteria used for restraint products and should be explicit that pet containment is not covered by the vehicle standards themselves.

Welfare guidance supports the underlying case and belongs in instructions: the American Veterinary Medical Association publishes animal transport guidance that treats secure containment as a welfare measure, which improves customer compliance with whatever the product was tested to do. Instructions that explain why a feature exists, rather than only how to use it, measurably improve correct use, and correct use is the variable no protocol can control for.

Scope exclusions deserve as much attention as the criteria themselves. A protocol covering car travel does not extend to air travel, where airline rules and live-animal transport regulations govern instead, and presenting one as evidence of the other is a misrepresentation risk rather than a marketing opportunity. State the covered mode explicitly on the product page and in the listing copy.

Production Control: Keeping the Listed Product the Shipped Product

A listing describes a specific tested configuration, and production control is what keeps the shipped product identical to it. This is the part of the programme that continues after the certificate arrives, and it is where most programmes quietly fail.

1. Component lock: webbing, hardware, foam, mesh and closure components recorded by specification and part number for the tested build.
2. Revision control: any design change assessed against the protocol before it ships, with a written record.
3. Batch verification: anchorage pull testing and closure function testing per production batch to stated forces.
4. Instruction and label check: the shipped instruction and label set verified against the protocol's requirements.
5. Re-test triggers: annual re-testing plus immediate re-testing after any change to a listed component.

At goods-in, the defects that matter are the ones that remove a tested property: wrong closure hardware, missing reinforcement, mesh substituted for a different aperture, or a missing required label. Our production team inspects to AQL 2.5 with all of these classified as critical rather than cosmetic.

Records should be kept deliberately rather than incidentally. A single programme file containing the protocol version, the report, the tested model identifier, the component lock list, the instruction set and the re-test schedule lets anyone answer a question in minutes, and it is the difference between a programme that survives staff changes and one that quietly stops being maintained.

The most common drift occurs in consumable components. Webbing, buckles and foam are sourced lot by lot, and a substitution made to solve a two-week shortage is exactly the change that invalidates a listing. Our production team treats every component named in the test report as frozen and requires a written concession before any alternative enters bulk.

A Practical Sequence for Buyers

Read the protocol first, design to its criteria, prototype against the clause list, book the laboratory or programme slot, test production-representative units, then write the marketing from the result. Programmes that begin with marketing and work backwards are the ones that end up with claims the file cannot support.

The second season is where a programme either compounds or decays. Those that re-test on a schedule, hold the component lock and keep the model-to-SKU record current get cheaper and safer each year. Those that treat the result as a one-off certificate drift away from the tested configuration without anyone deciding to drift.

Begin with the protocol document rather than with a sample request. Reading the exclusion criteria first tells a buyer whether the product type is eligible at all, which saves both the sampling fee and the 6-10 working days of development time. Buyers who request samples before checking eligibility are the ones who report the programme as expensive.

Timing

Qualification sampling runs 6-10 working days before the laboratory or programme phase, which has its own scheduling and should be booked when the brief is finalised. Bulk production is 35-50 days from approved sample, with the approved sample being the tested configuration rather than a later revision.

Commercial Terms

Standard terms are MOQ 500 pieces per colourway, T/T 30/70 and FOB Xiamen. Because a listed configuration depends on stable component sourcing, and because a substitution is invisible on a finished sample, buyers should confirm that closure hardware, webbing and mesh specifications are locked for the programme year, and that substitutions trigger review before they reach the line. Buyers extending a range should also confirm whether a new size or colourway requires its own test, since eligibility and performance often differ by size rather than by colour.

Supplier Questions That Matter

Ask whether the tested unit came from production tooling, whether the protocol version is stated, whether the model identifier in the report matches the SKU being offered, and whether listed components are locked and change-controlled. Four clean answers indicate a supplier who has supported a documented programme before. Ask one further question - what happens if a listed component is discontinued - because the answer reveals whether change control exists at all.