Waterproof Testing Standards for Custom Bags: the ISO, ASTM and AATCC System, and How to Submit Properly

A standard is a method, not a promise

The most expensive misunderstanding in this field is treating a standard reference as a quality claim. It is not one. Two laboratories working independently produce comparable numbers from the same material because waterproof testing standards tell both of them to do the same thing. They standardise the procedure, not the outcome. When a supplier writes "tested to ISO 811" on a quotation, they have told you which recipe was used; they have not told you the result, the criterion, or whether the number means anything for your product. In practice the useful shortlist of ISO ASTM AATCC methods is short, and almost all of them work this way.

This has a direct drafting consequence. A specification has two halves that must both be present: the method, which says how to measure, and the criterion, which says what counts as acceptable. "Hydrostatic resistance tested to ISO 811" is a method without a criterion. "Hydrostatic resistance not less than 3,000 mm when tested to ISO 811, current edition, on production material, at three points across the roll width" is a specification, and it is enforceable.

It also explains a pattern that confuses buyers. Two suppliers can both hold reports to the same standard, one showing a large number and one showing a modest one, and both can describe themselves as tested to that standard. The standard did its job. What differed was the product. This is why asking for the result and the criterion, rather than for the standard name, is the question that produces usable information.

Who publishes them and how the numbering works

Half a dozen bodies produce nearly everything relevant to waterproof bags, and knowing which is which saves time when a report arrives from an unfamiliar laboratory. The names appear in a fixed short form on every report, and each body has a different default philosophy about what its documents cover.

• ISO, with the ISO and ISO/IEC prefixes, publishes international methods across textiles, plastics and enclosures. It is the most widely recognised across markets, and its editions are revised periodically, so a citation needs a year.
• ASTM International, prefix ASTM, produces US-origin methods with very detailed procedure and apparatus requirements, frequently specifying conditioning and test speed more tightly than the ISO equivalents.
• AATCC covers US textile methods and is heavily used for water resistance and colour, with particularly well-established spray, rain and impact penetration procedures.
• CEN and CENELEC, prefix EN, publish European standards adopted nationally across the EU, and matter most for protective-clothing style claims.
• IEC covers electrotechnical subjects including the enclosure protection code that underpins the rating system used across this industry.
• JIS and GB/T are the Japanese and Chinese national methods, frequently requested by domestic retailers in those markets and similar in principle to the ISO methods but different in detail.

The numbering is not random and it is worth reading. In the ISO system the number identifies the document and the suffix year identifies the edition, so a citation without a year is incomplete. Parts appear as a dash and a second number. ASTM uses a letter prefix for the category — D for materials, F for specific applications — followed by a number, with a two-digit year suffix and often an M suffix indicating metric units.

One practical trap: methods that sound equivalent are not. An ISO hydrostatic method and an AATCC hydrostatic method can differ in the rate at which pressure is raised, the area under test and the endpoint definition, producing numbers that are similar in magnitude but not interchangeable. Never compare a result obtained under one body’s method with a criterion written for another’s.

Methods versus specifications

The distinction sounds pedantic and it changes what a purchase order can enforce. A test method describes a procedure and produces data. A specification sets requirements, usually including minimum values, and very often references methods to define how those requirements are demonstrated. Some documents do both. Knowing which you are citing determines whether your clause is a request for information or a requirement.

• A test method gives a repeatable procedure and a number — how to measure resistance to water penetration, for example.
• A material specification sets minimum requirements for a material, and in a waterproof programme it sets the performance floor for a coating or film at purchasing.
• A product specification sets requirements for the finished item, which is what an assembled bag actually needs.
• A classification standard defines bands or classes with limits, and assigns a protection class to a measured result.
• A management system standard sets requirements for the organisation rather than the product, governing records, competence and traceability.

The most common specification error is citing a management-system standard as though it were a product requirement. Certification to a quality-management standard says the organisation keeps records and follows its own procedures; it says nothing measurable about whether a particular bag keeps water out. Both matter, and they are frequently confused in procurement documents because both appear on the same certificate page. Our note on quality management systems in waterproof bag manufacturing sets out what that certification does and does not cover.

The second common error is citing a material specification where a product requirement is needed. A fabric that meets a demanding material specification can be assembled into a bag that leaks at the closure, because the specification measured a flat swatch and the product has geometry, seams and an opening. Write product-level requirements for product-level claims.

The short list that actually applies

There are hundreds of textile and plastics methods and a custom bag programme needs perhaps a dozen. The list below is the set that comes up repeatedly, grouped by the question each one answers. Confirm the current edition and the exact scope before citing, because methods are revised and occasionally withdrawn.

The full texts of these methods are published by the issuing bodies, and the catalogues of ISO, ASTM International and AATCC are the places to confirm the current edition and exact scope before citing any of them.

Two comments on using this list. First, the hydrostatic methods answer a question about flat material and should never be used to support a claim about an assembled product; for that you need a product-level method or an immersion method appropriate to the claim. Second, the light-exposure methods produce a change in properties, not a service life, and any durability statement built on them needs a correlation exercise rather than a conversion factor.

Where a programme sells into more than one market, national variants matter. Chinese retail buyers frequently ask for the GB/T hydrostatic and spray methods, and Japanese buyers for the JIS equivalents. These are generally similar in principle to the ISO methods but differ in detail, so a report obtained under one cannot be presented as evidence of compliance with another without re-testing.

How to read and write a citation

A correct citation removes ambiguity, and an incomplete one creates it. The minimum is the issuing body, the number, and the edition year. Where the document has parts, the part is required. Where the method offers options — different rates, different endpoints, different atmospheres — the option chosen must be stated, because the result depends on it and the report will otherwise be silent on something that matters.

• Write the edition year every time. "ISO 811" is underspecified; "ISO 811:2018" is a document.
• State the part where one exists; the part changes the apparatus and the result.
• State the option or procedure letter where the method offers choices, especially pressure ramp rates and endpoints.
• State the conditioning atmosphere, because moisture content changes mechanical and barrier behaviour.
• State the specimen type: flat material, seam, or assembled product. It is the largest single source of misleading reports.
• State the criterion separately from the method, with a number and a unit.

Conditioning is the item most often omitted and it is not cosmetic. Textile and coated materials are hygroscopic to varying degrees, and a specimen tested straight off a humid production floor can behave differently from one conditioned in a standard atmosphere. Standard conditioning atmospheres exist precisely to remove this, and the reference atmosphere should be named in the instruction rather than assumed.

When reading a report, check these six items before reading the result. A result obtained from the wrong specimen type, in the wrong atmosphere, under the wrong edition, is not a wrong result — it is a result answering a different question, and it is worth exactly nothing for your purpose.

Choosing which standard to cite for a given claim

The right method is determined by the claim you intend to make, not by the method your supplier happens to have a report for. Working backwards from the claim prevents the most common mismatch: a material-level report used to support a product-level statement.

The third and fourth rows are where programmes get into difficulty with marketplace reviews, because the claim language implies a level of protection that the cited method does not address. A bag described as waterproof with a fabric hydrostatic certificate behind it has made a claim about an assembly using evidence about a swatch.

A useful discipline before any submission: write the claim sentence first, then ask what would have to be tested to make that sentence true, then choose the methods. Working in that order takes an extra ten minutes and eliminates most mismatches. The wider commercial picture of what different markets expect is covered in our review of certification requirements across global markets.

Accreditation: what the certificate on the wall covers

Laboratory accreditation to a competence standard — the widely used international one for testing and calibration laboratories — certifies that the laboratory operates a competent, impartial and consistent system for the specific tests listed in its scope. It does not certify every test the laboratory offers, and this distinction causes real problems when a buyer assumes a report is accredited because the laboratory is.

• Ask for the scope document, not the certificate. The scope lists the specific methods and materials covered.
• Check that the method on your report appears in the scope, with the same edition.
• Check whether the report is endorsed as accredited or issued as a non-accredited service; laboratories issue both.
• Understand that accreditation covers competence in performing the method, not the interpretation of the result.
• Note that subcontracted tests appear on reports from accredited laboratories but may be performed elsewhere.
• Remember that flexibility is legitimate: many accredited laboratories develop or modify methods for non-standard products, and those are not accredited by definition.

The last two points matter for waterproof bags specifically. A finished bag is not a standard specimen, and a laboratory testing it will often adapt a method or use an in-house procedure. That is entirely proper, but the resulting report is not accredited evidence, and it should not be presented as though it were. Ask which parts of the work were performed under accredited methods and which were not.

International recognition of accreditation operates through mutual recognition arrangements between accreditation bodies, which is why a report from an accredited laboratory in one country is generally accepted in another. The arrangements are maintained by ILAC, and confirming that the accrediting body is a signatory is a fast way to check whether a report will travel.

Writing the instruction a laboratory can act on

Laboratories do not guess. If the instruction is vague they will either ask — costing you a week — or choose a default, which may not be what you needed. A good instruction takes about eight lines and answers every question the laboratory would otherwise have to raise. This is the single highest-leverage document in the whole submission, and it is usually written in three hurried lines.

• What is the sample: flat material, component, seam specimen, or finished product? Say explicitly.
• What is the question, in one sentence? Not "test waterproofness" but "does water penetrate the assembled bag at the closure under these conditions".
• Which method, with edition, part and any options selected?
• What is the criterion, if the laboratory is being asked to judge pass or fail?
• How many specimens, and how should they be selected or prepared?
• What conditioning is required before testing?
• What must the report contain, and by when?
• Who is authorised to approve deviations, and how should the laboratory proceed if the method cannot be applied as written?

The last line is the one everyone omits and it is the one that saves the most time. When a method turns out not to fit the product — which happens constantly with finished bags — the laboratory needs to know whether to stop, call, or proceed with a documented deviation. Deciding that in advance avoids either a stalled job or a report with an undocumented modification buried in it.

It is also worth stating what you will do with the report. If it will be shown to a retailer, a marketplace or a customs authority, the laboratory can advise on what needs to appear on it, and sometimes on whether an accredited method is required. That conversation is much cheaper before the work starts than after the report is issued.

The submission packet, and why submissions stall

Most delays in third-party testing are caused by the submission rather than by the test. Samples arrive unlabelled, in the wrong quantity, without the paperwork, or from a different production lot than the one documented. The laboratory then has to stop and ask, and a two-day test becomes a two-week correspondence.

Quantity deserves a specific note. Testing is usually destructive, and a programme that sends exactly the minimum has no margin for a failed specimen, a retest, or a request to test an additional condition. Send more than the method requires and give the laboratory permission to use spares as needed.

Photographs of the sample as submitted are worth more than they appear to. Disputes about whether the tested item was representative are almost impossible to resolve without them, and they cost nothing at the point of packing.

Laboratories rarely reject work outright; they stall it. Understanding the usual causes turns a three-week cycle into a one-week one, and most of the causes are entirely within the buyer’s control.

• Insufficient sample quantity for the number of conditions requested.
• Samples unlabelled or labelled inconsistently with the paperwork.
• The requested method does not fit the product, and no deviation authority was given in advance.
• The edition cited is withdrawn or superseded, so the laboratory must confirm which to use.
• Conditioning was requested but the samples arrived outside the stated storage history.
• The criterion was not provided, so the laboratory cannot issue a pass or fail statement.
• Invoicing or purchase-order details were missing, so the job was not released to the bench.
• The named contact is unreachable when the laboratory has a question.

The withdrawn-edition one deserves emphasis because it is invisible to most buyers. Methods are periodically revised, and a citation that was correct three years ago may now refer to a superseded edition. Laboratories will usually apply the current edition and note it, which is proper, but if your specification cites the old one, your acceptance criterion may no longer match the method that produced the result. Checking editions annually against the publisher is a fifteen-minute task.

The fix for almost everything on the list is a template. Build one submission instruction per product family, keep it current, and reuse it. Programmes that do this see their turnaround drop sharply, and the laboratory starts treating their work as routine, which has its own benefits when a rush job appears.

Auditing the report

A report should be read as an audit trail rather than as a conclusion. Six items, checked in order, determine whether it is usable. None of them require technical expertise; they require knowing what to look for and refusing to skip to the result.

Individual values matter more than summaries. A report showing only a pass or fail discards the information that lets you judge how close to the limit the product sits, and whether the result is representative. Ask for the underlying values as a matter of routine; laboratories keep them and will usually provide them on request.

Deviations should be read carefully rather than skipped. A documented deviation is a sign of an honest laboratory, not a poor one — but it changes what the result means, and if a deviation affects the aspect of the test you care about, the result may no longer answer your question. This is particularly relevant to connecting laboratory results to real-world performance, where a small change in conditions can change the mechanism being tested.

Cost, turnaround and scheduling

Testing cost is driven by machine time and by reporting effort, which means it scales with duration and with the number of conditions far more than with the number of samples. A five-day exposure costs noticeably more than a one-hour one regardless of how many units are in the chamber, and a report requiring interpretation costs more than one producing a number. Understanding that shape helps a programme spend where it gets information.

Turnaround is dominated by queue position and conditioning time rather than by the test itself. Conditioning alone can take a day before anything starts. Realistic planning figures for a straightforward waterproof method are a few working days once the job is released, rising to several weeks where exposure, ageing or sequential conditioning is involved. Rush slots exist and are expensive; the more reliable lever is submitting early in the development cycle rather than at the end of it.

The scheduling mistake that costs the most is testing at the end. A programme that waits until the pre-shipment stage has no time to act on a failure, so the test becomes a gamble rather than a control. Testing at material qualification and again at pilot-batch stage gives two opportunities to change something, and the second is still cheap compared with a bulk run discovered to be wrong.

Budget for the recurring side as well. A certificate from the development stage supports one claim about one batch; a programme that intends to keep making the claim needs periodic re-verification, and the cost of that is a permanent line rather than a one-off. Our companion guide on working with third-party testing laboratories covers the commercial side of that relationship in more detail.

Keeping standards current

Standards are living documents. New editions appear, parts are added, methods are confirmed, withdrawn or replaced, and national adoptions of international standards do not all move at the same speed. A specification written three years ago can quietly point at a document that no longer exists, and the laboratory will not always notice the mismatch in your paperwork.

• List every standard cited in your specifications, with the edition, in one register.
• Check the register annually against the publisher’s catalogue; it is a short task and prevents drift.
• When an edition changes, re-read the scope and any changed procedure before simply updating the number.
• Where a method is withdrawn with no replacement, decide deliberately whether to keep an internal method or move to another body’s equivalent.
• Update internal work instructions and inspection documents at the same time, not only the specification.
• Keep superseded reports, clearly marked, so historical comparisons remain possible.

The check that matters most is not whether the number changed but whether the procedure changed. A new edition can alter the apparatus, the rate or the endpoint, which means results obtained before and after the change are not directly comparable. Where a programme has historical data, note the edition boundary on the chart so nobody reads the step as a process shift.

There is also a commercial angle. A retailer or marketplace may specify a minimum edition, and a report against an older one can be rejected even though the result is sound. Knowing which edition your customers expect, and citing it explicitly, removes a class of avoidable rejection.

Turning the system into a working specification

Everything above converges on one short document: the testing clause in a purchase order or technical file. Written well, it takes half a page and removes most of the arguments that otherwise occur after something goes wrong. Written badly — usually as a list of standard names with no criteria — it creates the appearance of control without any of the substance.

• For each claim, one line: what is tested, by which method and edition, on what specimen, to what criterion.
• State who performs the test, where, and whether an accredited method is required.
• State the frequency: at development, per lot, per season, or on a sampling basis during the run.
• State what evidence is retained, and for how long.
• State what happens when a result fails, including the disposition of product already made.
• State who pays for initial testing, for re-testing after a failure, and for periodic re-verification.

The final point about payment is worth deciding explicitly, because it is the one most often left open and most often argued about later. Testing at development is usually a development cost; testing triggered by a failure is usually the supplier’s; periodic re-verification depends on who benefits from the claim. Allocating these in advance costs a sentence.

The other half of a working specification is the internal one: the methods the factory runs during production, which are usually simpler, faster versions of the same ideas. Those need to be tied to the external methods by at least one correlation exercise, or the two sets of numbers will drift apart and nobody will know which to believe. Our guide to seam integrity testing methods describes the in-house side in practical terms.

If you want help mapping your intended claims to the right methods before you commission any testing, send us the claim you intend to print. We will tell you which methods support it, which do not, and which specifications we would write into the order. Minimum per style is 500 pieces, samples in 6–10 working days, bulk in 35–50 days, quoted FOB Xiamen.

Frequently Asked Questions

Q1. Does testing to a standard mean the product passed?

No. A standard specifies a procedure, not an acceptable outcome. It produces a number; whether that number is good enough is a separate decision that has to be written as a criterion alongside the method.

Q2. What is the difference between a test method and a specification?

A method describes how to measure and produces data. A specification sets minimum requirements and often cites methods to define how those requirements are demonstrated. Only the second is enforceable in a purchase order.

Q3. Why do two suppliers with reports to the same standard perform differently?

Because the standard produced comparable numbers, and the numbers differed. The standard did its job; the products are not equivalent. Always ask for the result and the criterion, not just the standard name.

Q4. How should I write a standard citation?

Body, number, edition year, part where one exists, and any options selected, plus the conditioning atmosphere and the specimen type. A citation without a year or a specimen type is incomplete.

Q5. Can I compare results from an ISO method with an AATCC method?

Not reliably. Similar-sounding methods can differ in rate, area, endpoint and conditioning, producing numbers that look comparable but are not. Compare like with like, or convert only with a documented correlation.

Q6. Which method do I need for a hydrostatic resistance claim?

A hydrostatic method on flat material, such as the ISO or AATCC versions, with the edition and the criterion stated. Use it for material selection, never as evidence about an assembled product.

Q7. What do I cite for a claim about the finished bag?

A product-level method that tests the assembly, typically a rain or immersion method appropriate to the claim, plus verification of the closure. Material results do not support product claims.

Q8. What does laboratory accreditation actually certify?

That the laboratory operates competently for the specific tests in its accredited scope. It does not cover every test the laboratory offers, so check the scope document and whether the report is endorsed as accredited.

Q9. Why was my report not accredited even though the laboratory is?

Possibly because the method is outside the accredited scope, because the work was subcontracted, or because the method had to be adapted for a non-standard product. All three are normal and all three change what the report proves.

Q10. How many samples should I send for testing?

Enough for the planned specimens plus spares, typically five per condition. Testing is destructive and re-tests are common, so sending exactly the minimum leaves no margin.

Q11. What is the most common cause of laboratory delays?

The submission rather than the test: insufficient quantity, unlabelled samples, missing paperwork, or an unreachable contact when the laboratory has a question. A reusable submission template prevents most of it.

Q12. What should I check first when reading a test report?

Identity of the samples, the method with its edition, the conditions and any deviations, whether individual values with units are shown, and whether the method was within the accredited scope. Read the result last.

Q13. How long does waterproof testing take?

A few working days once the job is released for a straightforward method, rising to several weeks where conditioning, exposure or sequential ageing is involved. Queue position and conditioning dominate, not the test itself.

Q14. Should I test at the end of development or during it?

During it. Testing at material qualification and again at pilot-batch stage gives two chances to change something; testing only at pre-shipment turns the test into a gamble with no time to act.

Q15. How often should I re-check which edition of a standard applies?

Annually. Methods are revised, withdrawn or replaced, and a specification written three years ago can point at a document that no longer exists or whose procedure has changed.

Q16. What happens if a standard I cite is withdrawn?

Decide deliberately: move to the replacement, keep an internal documented method, or switch to another body’s equivalent. Note the edition boundary on any historical data so a change in procedure is not read as a process shift.

Q17. Who should pay for testing?

Decide in advance and write it down. Development testing is usually a development cost, testing triggered by a failure usually falls to the supplier, and periodic re-verification depends on who benefits from the claim.

People Also Ask

What is a waterproof testing standard?

A published procedure that defines how to condition, expose and measure a specimen so different laboratories produce comparable results. It does not define whether a result is acceptable.

Which standards apply to waterproof bags?

Hydrostatic and spray methods for fabric, rain and immersion methods for assemblies, enclosure codes for ratings, and peel or tensile methods for seams and coated materials.

Is ISO or AATCC better for waterproof testing?

Neither is better; they answer the same questions with different details. Choose based on the market and the claim, then never compare results across the two without a correlation.

Does an accredited lab report guarantee the product is waterproof?

No. Accreditation covers competence in performing the listed method within its scope, not the interpretation of the result or the suitability of the claim.

How do I submit samples to a testing laboratory?

Send enough units plus spares, label each one, provide lot traceability and a written instruction naming the method, edition, criterion and conditioning, and name a contact who answers quickly.

Why do test reports get rejected by retailers?

Usually because the method does not match the claim, the edition is outdated, the sample is not identified, or individual values with units are missing from the report.