Pet Bag Manufacturing Agreement: Clauses Every Buyer Needs

Wholesale pet bag sourcing desk · Updated 2026-10-06 · 15 min read

How to source pet bags is mostly a question of sequence - spec, sample, test, then price - and Market & Business Strategy sits in the second step. Pet bag sourcing that begins with a photograph rather than a technical pack tends to add two rounds before anyone can quote.

Why the Agreement Matters More Than the Payment Method

Payment structures get most of the attention in sourcing because they are where money visibly moves, but the agreement is where outcomes are decided. A letter of credit protects against non-shipment; it does nothing about a colour that is two shades off, a pattern that was reused for another customer, or a delivery that slipped three weeks inside a window the buyer never pinned down. Those are contract questions.

The asymmetry is worth stating plainly. Payment instruments allocate who holds money during the cycle. The contract allocates who bears the cost when something goes wrong, and because something eventually goes wrong in every program, the contract is the instrument that gets used.

Three failures recur in this category and all three are contractual rather than operational. The first is a specification that exists only as a sample, so nobody can say whether production conformed. The second is a delivery date expressed as an aspiration rather than as a computable period from a defined trigger. The third is silence on intellectual property, which becomes expensive the moment a successful SKU appears elsewhere.

Buyers should also understand that the agreement governs the relationship in both directions. A buyer who cancels late, changes specification mid-production, or delays approval is in breach of it just as much as a supplier who ships late. A balanced agreement is easier to enforce than a one-sided one, and suppliers negotiate harder against documents that only impose obligations on them.

The practical conclusion is that the agreement should be written once, carefully, as a template, and then applied per order with a short order schedule. Building it properly takes a few hours and it is reused for years.

Payment instruments decide who holds money during the cycle; the manufacturing agreement decides who bears the cost when something goes wrong, which is why it is the document that eventually gets used.

The Specification Clause: Making the Product Definable

This is the clause that prevents disputes, and it is the one most often reduced to a sample and a photograph. A sample is useful evidence and a poor definition, because a sample ages, fades, and cannot state a tolerance.

A workable specification clause has two parts. The first is a written schedule listing material, dimensions, colour reference, hardware, closure, weight capacity, labelling and packing. The second is an approved reference sample, identified by a number and a date, which governs appearance and feel where words cannot.

Tolerances deserve particular care. No production process is exact, and a clause that demands perfection is unenforceable in practice. Stating a dimensional tolerance, conventionally plus or minus one to two centimetres on soft goods, and a shade tolerance against the approved sample, converts an argument about acceptability into a measurement.

The equivalence trap should also be closed. Where a material is described as a named supplier's product, the clause should say whether an equivalent is acceptable and, if so, who approves it. Without that sentence, a supplier will substitute freely and reasonably claim they were entitled to.

Finally, the clause should state that the specification can only be changed in writing and that the approved sample reference is updated when it is. Version control is what prevents the common failure where production is made to a superseded specification because one party's file was older.

Price, Quantity and Change Control

The price clause should state the unit price, the currency, the quantity band it applies to, and what it includes. Whether the price includes packing, labelling, testing, tooling amortisation and inland haulage to port is the most common source of invoice disagreement, and it is entirely preventable with one sentence.

Quantity clauses in this category need a tolerance. Production overruns and underruns are normal in cut-and-sew, and the conventional tolerance is plus or minus five percent at the supplier's option, settled at the unit price. Buyers who need an exact quantity for a retail allocation should say so and should expect to pay for it, because exact quantity means the supplier absorbs the risk of cutting losses.

Change control is the clause buyers undervalue most. It should state three things: that specification changes must be in writing; that the supplier must advise the cost and schedule effect before implementing; and that the buyer must approve before the change takes effect. Without it, a buyer's casual comment becomes a binding variation and a supplier's unilateral substitution becomes irreversible.

Price adjustment mechanisms suit long programs. Where a program runs for a year or more, a clause tying price to a named input index, with a review date and a notice period, is more workable than a fixed price that the supplier will inevitably seek to renegotiate.

Tooling and sampling costs should be allocated explicitly. Who owns the cutting dies, the embroidery files, the printing plates and the moulds, and what happens to them if the program ends. This is a small clause that prevents a large argument, particularly where tooling was partly funded by the buyer.

Buyers should also state that the price applies to the specification as written. A quotation given against a loose description is not a price for a defined product, and saying so protects both parties from the argument that follows when the specification is later tightened.

Delivery, Delay and Remedies

A delivery clause fails when the date is not computable. Phrases such as within eight weeks or as soon as possible cannot be enforced, because both parties will compute them differently. The clause should state a period and a trigger: bulk production completed within 35-50 days after written sample approval, with the approval date recorded.

The trigger matters as much as the period. Sample approval, deposit confirmation and specification freeze are all candidate triggers, and the agreement should name one. Where a program has several gates, the clause should run from the last of them, because that is what the supplier actually needs before starting.

The remedy is what makes the date real, and it should be proportionate rather than punitive. A clause that is too harsh will be resisted in negotiation and ignored in practice; one that is specific will usually be honoured. Common workable structures include the supplier bearing the cost of expedited freight where a delay is theirs, a small percentage credit against the next order, and a right for the buyer to cancel with deposit return where delay exceeds a stated number of days.

Force majeure should be defined narrowly. A clause that excuses any delay for any reason outside the supplier's control swallows the delivery obligation, because material availability and labour supply are arguably outside it. A workable version lists genuinely exceptional events, requires prompt notice, requires mitigation, and gives the buyer a cancellation right if the event persists beyond a stated period.

Buyers should also address their own obligations. Approval within a stated number of days, payment within a stated number of days, and a consequence if the buyer delays, because a supplier whose production slot is held open by buyer inaction has a legitimate claim.

Finally, the clause should state what happens on late approval of samples. Where the buyer takes three weeks to approve, the 35-50 day period cannot reasonably run from the original order date, and the agreement should say so.

A delivery date is only enforceable if it runs from a defined trigger and carries a proportionate remedy, which is why the clause should name the approval event and the consequence together.

Quality, Inspection and Acceptance

The quality clause should name the standard, the timing and the consequence. Naming AQL 2.5 as the acceptance level is the conventional starting point for this category, and it should be stated together with the lot definition and the inspection basis.

Two inspection points are worth providing for. An inline inspection when bulk reaches roughly eighty percent completion, which allows correction while there is still time, and a final pre-shipment inspection against the approved sample and the written specification. The agreement should state that the buyer or their nominated inspector may attend or perform both.

Independent providers matter here. Inspection by the supplier's own team is useful but self-certifying; providers such as SGS issue reports that both parties and any forum will accept. Naming an inspector in the agreement means their report is the agreed evidence rather than one party's opinion.

The consequence of failure should be stated, with three options: rework at the supplier's cost and re-inspection, acceptance at an agreed price adjustment, or rejection. Rejection is the strongest and the most disruptive, and buyers should be clear that a rejected lot does not relieve the supplier of the delivery obligation.

Latent defects deserve a separate sentence. A defect that could not reasonably be found by pre-shipment inspection, such as a coating failure that appears after some weeks, should be covered for a stated period after receipt, with a defined remedy. Without it, the buyer has no claim on the failure mode that matters most to a retailer.

Quality systems evidence supports the clause rather than replacing it. Certification to a recognised management standard such as ISO 9001 indicates that processes exist, but the agreement still needs to state what happens when output falls below the standard.

Buyers should also require retention samples. Two sealed units per lot, held by each party for a stated period, resolve most later arguments about what was actually shipped.

Intellectual Property and the Confidentiality Interface

Three categories of intellectual property arise in a pet bag program, and the agreement should allocate each. The first is the buyer's brand and any design they supplied, which should be acknowledged as the buyer's property and used only for the buyer's orders. The second is tooling and patterns created for the program, where ownership should be stated along with who may use them. The third is any design contribution made by the supplier, which is the one most often left unallocated and most often disputed.

The clause should state that the supplier will not manufacture the buyer's design for any other party, will not sell it through any channel, and will not use the buyer's branding, photography or listing content. It should also state what happens to physical tooling, digital pattern files and samples at the end of the relationship, including whether they are returned, destroyed or retained.

A practical point worth stating explicitly: restrictions should survive termination for a defined period, commonly two to three years. An IP clause that ends with the agreement is close to worthless, because the exposure begins after the relationship ends.

Confidentiality interfaces with this and is frequently assumed rather than drafted. A short mutual clause covering specifications, drawings, forecasts, pricing and customer information is enough for most programs, and where a buyer is sharing genuinely novel designs a separate agreement is appropriate. Our guide to non-disclosure agreements covers when that separate document is warranted.

Where the supplier contributed design work, the cleanest arrangement is that the buyer owns the design for the product and the supplier retains the right to use generic construction knowledge. That distinction is commercially sensible and it is the one most suppliers will accept.

Buyers should also consider registering their rights in the relevant markets before sharing new designs widely, because a contract binds the counterparty and nothing else. Contractual protection and registered protection do different jobs, and a program that matters commercially usually needs both.

Compliance, Documentation and Product Safety

Compliance is a contractual obligation before it is a technical one. The agreement should state which market requirements the goods must satisfy, who provides the evidence, and what happens if they do not. Naming the destination market is essential, because the same product faces different requirements in the European Union, the United Kingdom and the United States.

The documentation clause should list what the supplier must provide: test reports from recognised laboratories, material declarations, a certificate of origin, packing lists, and any declaration required by the destination. Naming them means a missing document is a breach rather than an administrative delay, which matters when a shipment is held.

Chemical compliance deserves specific wording. Restricted substance requirements differ by market and are updated regularly, and a clause that requires conformance to the current version of the applicable regulation, at the time of production, is more workable than one that lists substances that will be out of date. Regulators such as the Consumer Product Safety Commission publish the requirements that apply in the United States market.

Recall and withdrawal provisions belong in the agreement as well. Who notifies whom, who bears the cost where the cause is attributable to one party, and what records must be retained for traceability. Buyers supplying large retailers frequently have recall obligations in their own supply contracts, and those need to flow down.

Traceability is the practical backbone of all of this. A clause requiring lot identification on cartons and retention of production records for a stated period converts a recall from a company-wide event into a targeted one, and it is inexpensive to comply with.

Finally, the clause should state that compliance obligations survive shipment and do not end with acceptance. A product that passes inspection and later fails a market surveillance test is still the supplier's responsibility where the cause is material non-conformance, and the agreement should say so.

Term, Termination and Dispute Resolution

Two structures suit this category. A framework agreement covering the relationship, with short order schedules for each purchase, works well for repeat programs because the negotiated clauses are settled once. A single order agreement works for occasional buying but reopens the same negotiation every time.

Termination should be available for cause and, in framework arrangements, for convenience with notice. Termination for cause should name the triggers: persistent late delivery, repeated inspection failure, breach of the IP clause, insolvency, or a compliance failure. Naming them prevents the argument about whether a particular failure was serious enough.

The consequences of termination need stating. What happens to work in progress, to deposits, to tooling, and to any confirmed but unshipped orders. The most common dispute at the end of a relationship is not about who was right but about who owns the half-finished goods.

Dispute resolution should be practical. A clause requiring good-faith negotiation for a stated period, followed by a defined forum, is workable. Where the counterparty is in a different jurisdiction, the choice of forum and governing law is one of the most consequential clauses in the document, and it should be considered deliberately rather than accepted as boilerplate.

Notice provisions sound trivial and decide cases. A clause stating the address and method for valid notice, including whether email suffices, prevents a party from claiming they never received something that matters.

Buyers should also include a short clause on assignment and subcontracting. Whether the supplier may subcontract production, and whether they remain liable for a subcontractor's output, is a real question in this category, and silence effectively permits it.

Finally, review the agreement annually. Requirements change, volumes change, and clauses written for the first order rarely fit the tenth. A short annual revision, applied to the next order schedule, keeps the document useful rather than archival.

Applying the Agreement to a Single Order

An agreement is only useful if it can be applied quickly, and the test is whether an order can be raised without reopening negotiation. A workable order schedule is short: SKU, specification version number, approved sample reference, quantity with tolerance, unit price with currency and delivery term, the delivery period running from the stated trigger, the inspection standard, and the payment milestones.

Each of those fields should point back to the framework agreement rather than restate it. The order schedule says what is being bought this time; the agreement says how it will be judged. Where an order schedule starts restating quality or remedy language, the framework is not doing its job.

Buyers should resist the tendency to renegotiate per order. A supplier who accepts a framework and then seeks a different position on every order schedule is signalling that they did not intend to be bound, and the correct response is to return to the framework rather than to accommodate.

Where a genuine exception arises, such as an unusual material, a different destination market or a compressed schedule, it should be handled as a written variation referencing the clause it varies. Variations are normal; undocumented ones are what create disputes.

The final test is simple. Hand the order schedule and the agreement to someone who was not involved, and ask them to decide whether a particular delivery was late and whether a particular lot passed. If they can, the documents work. If they cannot, the missing field is the one that will be argued about later.