StartMade Blog Episode 155 ·

The "90% Goose Down" Long Puffers Leaked Feathers and Went Flat Within a Month
Down Fill Content & Fill Weight Verification Guide - Why Down Jackets and Duvets That Passed Inspection Go Flat and Leak Feathers, Down, Feather, Duck, Recycled Down and Synthetic Fill Compared, Hand Feel, Opening, Fill Weight, Content Analysis, Fill Power and Cleanliness Tests Compared, Approximate Criteria for Down Content, Fill Weight, Fill Power and Cleanliness, Gaps in Supplier Reports and RDS Certificates, Supplier Substitution Such as Good Material for Samples Only, Blended Duck Down, Reduced Fill Weight and Synthetic Fill in Some Panels, Spec Sheet Items, Stage-by-Stage Checks and a Checklist

Hello, this is StartMade.

A Korean brand selling winter apparel online had a down garment factory in Xiaoshan (萧山), Zhejiang make goose down long puffer coats. The label read "Goose down 90/10 (90% down cluster, 10% feather), fill weight 280g." The first season's 2,000 coats sold well, so the following year the brand reordered 4,000 from the same factory. The pre-shipment inspection passed on dimensions, stitching, zippers, labels and packaging, and the factory sent the down supplier's content test report along with a copy of its RDS certificate.

When the second batch went on sale in December, the reviews changed. "Feathers poke through the fabric every time I wear it," "it went flat after a month," "it smells fishy when it gets wet in the rain." The brand opened a few coats from remaining stock and sent them to a lab: down content was 61%, goose and duck down were mixed, and the fill weight was 231g, almost 50g short of the label. To save on material, the factory had blended in feather-heavy duck down and cut the fill weight as well.

Down is tricky because you cannot tell the difference by looking at a finished garment. A 90% down coat and a 60% down coat both feel lofty and warm the first time you wear them. The difference only shows after a few weeks of compression, wear and washing, as poor recovery, feather leakage and odor. This article covers how to check down content, fill weight and fill power, approximate criteria, the substitutions factories commonly make, and how to build down checks into your spec sheet and inspection checklist, from a practical sourcing perspective.

The "90/10" on the label and the down supplier's test report are not enough. Put down content, goose ratio, fill weight by panel, fill power and cleanliness criteria into the spec sheet as numbers, and test garments you pick yourself from production by opening them up. That is how you prevent under-content labeling and feather leakage claims.

Note This article is practical information as of November 2026, based on common practice at Chinese down suppliers and down garment and bedding factories, commonly used industry criteria, Korea's household textile safety and labeling standards under the Electrical Appliances and Consumer Products Safety Control Act, China's GB/T 14272 (down garments), GB/T 10288 (down test methods) and GB/T 17685 (down and feather), the International Down and Feather Bureau (IDFB) testing regulations, and the European EN 12934 labeling standard. Required criteria and test conditions vary by product type, destination country and sales channel, so set actual criteria based on your product and input from a test lab. Figures and amounts in this article are examples for illustration.

1. Why Down Jackets Go Flat and Leak Feathers So Quickly

Down fill is a mix of down clusters (绒子, 朵绒) from the underside of waterfowl and feathers (羽毛, 毛片) that have a quill. A down cluster has no quill, just fine filaments spreading out like a ball, so it traps a lot of air and springs back after being compressed. A feather's stiff quill supports weight but traps less air, and the quill tip easily pokes through fabric. So the lower the down ratio, the less warmth you get from the same weight, the slower the recovery after compression, and the more leakage.

The first thing a factory touches to cut cost is the down ratio, because 90% down and 60% down material differ by nearly double in price per kilogram. Blending in cheaper duck down instead of goose, or mixing in recycled down (回收绒, 再生绒) that has been taken out of old products and rewashed, is also common. Cutting fill weight makes the shell look less lofty, so factories compensate by switching to thinner fabric or narrowing the baffles to keep the volume.

Odor usually comes from insufficient washing. Down suppliers wash and dry the material to remove oil, dust and protein residue. When this process is shortened, the leftover organic matter produces a fishy smell when it meets moisture, and it gets worse if the goods absorb moisture in the container. Moisture problems are covered in Episode 118.

There are five main fill materials you will deal with on the ground.

FillChinese TermCharacteristicsWatch For
Goose down鹅绒, 白鹅绒·灰鹅绒Large clusters, good recovery, expensiveDuck down blended in
Duck down鸭绒, 白鸭绒·灰鸭绒Smaller clusters than goose, cheaperOdor if poorly washed
Feather-heavy material毛片, 含绒量低Heavy, quills poke through fabricGap between labeled and actual content
Recycled down回收绒, 再生绒Down taken from old bedding or garments and rewashed, broken clustersUndisclosed blending, odor and dust
Synthetic fill丝棉, 仿丝绵, 羽丝绒Polyester fiberfill and similar synthetic insulationSecretly used in some panels of a down product

Using duck down does not make a product bad. The problem is putting in material different from the label, or raising the feather ratio and mixing in recycled down while still selling it as "goose down 90/10." Apparel imports in general are covered in Episode 67 and children's products in Episode 65. This article focuses on confirming that the down you received has the content and weight you were promised.


2. Verification Methods — From Hand Feel to Lab Tests

Ask "Is the down content right?" and the factory will almost always answer "绝对是90绒" (it's definitely 90 down). Decide on verification methods before you rely on words.

MethodHowWhat It Tells YouLimits
Feeling from outsideRub a baffle between your fingers to feel for quillsRough sense of whether feather ratio is very highWeak basis for a decision
Visual check after openingOpen one baffle, put the fill in a clear bag and lookFeathers, quill fragments, synthetic fill, suspected recycled downCannot give content as a number
Fill weight measurementRemove all down from the garment or by panel and weigh itActual weight vs labeled fill weightDestroys one garment
Content analysisLab hand-sorts down, feather and residue and calculates weight ratios (GB/T 10288, IDFB)Down content, goose vs duck ratio2-5 days, samples must be chosen correctly
Fill power testPut a set weight of down in a cylinder, compress and measure recovery heightWarmth and material gradeUnits differ by test method
Cleanliness testsTurbidity (透明度), oxygen number (耗氧量), odor ratingInsufficient washing, risk of odor claimsRequires a lab

Fill weight can be measured on site

The simplest and most reliable on-site method is to open one garment, take out all the down and weigh it. A scale, a large plastic bag and about an hour is all it takes. Weighing by panel, such as body, sleeves and hood, also shows which panel was underfilled and whether any panel has synthetic fill instead of down. Losing one garment may feel wasteful, but as a way to prevent labeling violations across thousands of units it is cheap.

Agree on fill power units first

US and European brands state fill power as cubic inches per ounce (cu.in/oz), such as "650FP" or "800FP," while Chinese factories often state it as height in centimeters per 30g (蓬松度, cm). Test methods and conditioning also differ, so simple conversions do not line up. Write in the spec sheet which test method the value is based on, so there is no argument later.


3. What Criteria to Set

The standards a down product must meet differ by destination country. In Korea, household textile safety and labeling standards require at least 75% down content before a product can be labeled "down," and the gap between labeled and actual content must stay within tolerance. China's GB/T 14272 treats garments with at least 50% down content as down garments, and Europe sets labeling rules through EN 12934. Approximate values commonly used for quality control are below.

ItemApproximate CriteriaNotes
Down contentWithin -3 percentage points of the label (e.g. 87% or more for a 90% label)Check destination country tolerance; in Korea, below 75% cannot be labeled "down"
Goose ratio"Goose down" labels often require 90% or more gooseIf the rest is duck, state the ratio separately
Fill weightWithin -5% of label, each panel within ±5% of specOpen one garment and weigh by panel
Fill powerAt or above spec (e.g. 650FP or more)State the test method; drops when material grade changes
Cleanliness and odorTurbidity 500mm or more, oxygen number 10mg/100g or less, odor rating passReveals insufficient washing
Down leakage (钻绒)Number of fibers escaping after tumble or rub test at or below specDepends on downproof fabric and stitch density

The values to watch most closely are down content and fill weight. Both tie directly to cost, so they are what factories adjust first when squeezing price, and they are also behind most problems flagged in Korean consumer agency tests and safety surveys. Fabric shrinkage and colorfastness are covered in Episode 132 and seam strength in Episode 142.


4. Why You Should Not Simply Trust Supplier Reports and RDS Certificates

When you ask for proof of down content, the factory usually sends a test report from its down supplier or an RDS (Responsible Down Standard) certificate. These documents have gaps.

GapWhat It MeansWhat to Check
Supplier reports describe the raw materialResults from a sample the down supplier pulled from a bale; says nothing about which bales the factory actually usedTest results from opened finished garments
Old reportsA report from a different lot months ago reused again and againTest date, lot number, bale label photos
RDS is an animal welfare certificationCertifies no live plucking; does not guarantee content or fill weightTransaction certificate (TC) and content test, separately
Special test samplesGarments made with good material just for testingRandomly select from finished goods and seal before testing
Claiming RDS requires a transaction certificate To use the RDS logo on your product page, every stage from the down supplier to the sewing factory must be certified, and a Transaction Certificate (TC) must be issued for that specific shipment. If the factory only sends a copy of its supplier's certificate and cannot provide a TC, there is no proof RDS material went into that shipment. Ask for both the certificate and the TC, and verify validity on the certification body's website yourself.

5. Common Supplier Substitutions

Fill makes up a large share of the cost of a down product, and a 10-point difference in down content shifts the cost of one garment by a few dollars. When price pressure comes, factories start here. These are the types we have actually encountered in consultations.

TypeHow It WorksHow to Prevent It
Good material for samples onlyApproved sample and first order are 90/10, reorders drop to 70/30 or 60/40Content test on approved sample, compare every lot
Blending in duck downDuck down mixed into products labeled goose downCheck goose vs duck ratio in content analysis
Cutting fill weightLabeled 280g, actual 230g, baffles narrowed to keep volumeOpen and weigh by panel
Blending in recycled downDown from old bedding and garments rewashed and mixed inCheck for broken clusters and dust, cleanliness tests
Synthetic fill in some panelsBody is down, sleeves, hood and sides are polyester fiberfillBOM by panel, open and check each panel
Cutting the washingCheaper material washed and dried fewer timesOdor rating, turbidity and oxygen number tests

The most common pattern is switching material on reorders. Factories are careful with the first order because the customer is watching the response, but from the reorder they lower the content while presenting the same report again. A process for requiring prior approval of material or process changes is covered in Episode 119, and the same principle of secret material substitution is explained for leather in Episode 153 and silicone in Episode 154.


6. The Cost of Down Substitution

Down problems do not show at shipment and only appear after customers have worn the product for a few weeks, but the damage spreads well beyond the price of the garment.

Recalls and product liability are covered in Episode 53, and claims and disputes in Episode 36.


7. What to Put in the Spec Sheet

A single line of "goose down 90/10" is interpreted differently by every factory. Organize the items below into a table and attach it to your spec sheet.

ItemExample
Material type and contentWhite goose down, 90% down content or more (GB/T 10288 or IDFB method), 90% goose or more
Down supplierFixed down supplier and material code, prior written approval for changes
ProhibitionsNo recycled down or synthetic fill, no undisclosed duck down
Fill weight by panelFront and back body, sleeves, hood and collar listed by baffle in grams, total 280g ±5%
Fill power and cleanliness650FP or more (state test method), turbidity 500mm or more, oxygen number 10mg/100g or less, odor rating pass
Fabric and stitchingDownproof fabric spec, stitch density, down leakage test criteria
DocumentsIncoming material lot number, bale label photos, down supplier report, TC if claiming RDS, filling records

Writing fill weight for each baffle is tedious, but it is that much more reliable. If you only write "fill weight 280g," you have no basis to object when the factory shorts the sleeves and packs the body, but with per-baffle figures a single scale settles the question. Spec sheet and BOM writing is covered in Episode 59.


8. How to Add Down Checks to Your QC Checklist

Sample stage

When you receive the approved sample, open one garment and weigh the fill by panel, then send the removed down for content and fill power tests to set your baseline. This record becomes the reference point for comparing production goods. Sample management is covered in Episode 34.

Production stage

When down arrives before filling, get bale label photos, lot numbers and the down supplier's report. On the filling line, get the filling machine settings and per-baffle weighing records, and if possible take a sample directly from an incoming bale for a content test. If photos of the filling room show bales with different labels mixed together, ask immediately. Production tracking is covered in Episode 102.

Pre-shipment inspection stage

The inspector picks packed goods personally, weighs them, and opens one garment to check fill weight by panel and any synthetic fill. If results differ significantly from the approved sample or there are noticeably many quill fragments, seal randomly selected samples and send them to a third-party lab for content analysis and cleanliness tests. Down jackets are often needle-detected as well, so see Episode 140. Pre-shipment inspection is covered in Episode 100 and third-party inspection in Episode 47.

Sample down control conditions White goose down with 90% or more down content, 90% or more goose, fill weight 280g ±5% (each panel within ±5% of spec), fill power 650FP or more, no recycled down or synthetic fill. For each lot, weigh 10 randomly selected finished garments, open one to check fill weight by panel, and seal the removed down for content analysis and cleanliness tests. Compare with approved sample measurements. If anything falls outside criteria, increase the sample and recheck; if it still fails, hold the lot and refill or remake. Figures are examples; actual conditions must match the product type and destination country standards.

9. Sample Request Wording for the Factory

Putting the wording below in your purchase order or spec sheet takes material choice out of the factory's discretion and makes it easier to assign responsibility if a claim arises.

"本产品须使用指定原料商、指定型号的白鹅绒,含绒量不低于90%(按GB/T 10288或IDFB方法检测),鹅绒比例不低于90%,蓬松度不低于650FP。各部位充绒量须按规格书执行,总充绒量280g,允许偏差±5%。严禁掺入回收绒、再生绒、丝棉或任何未标示的鸭绒,任何部位不得以化纤填充物替代。未经我方书面同意,不得更换原料商、原料型号或生产工厂。每批须提交原料入库批号、原料包装标签照片、原料商检测报告及充绒记录;如标示RDS,须提供对应批次的交易证书(TC)。检测样品由我方验货员从成品中随机抽取并封样。如有争议,以第三方实验室的检测结果为准。如有不合格,须加抽复检;复检仍不合格的批次暂停出货,并使用合格原料重新充绒或重新生产。"
(This product must use white goose down of the designated supplier and material code, with down content of at least 90% (tested by GB/T 10288 or IDFB method), at least 90% goose, and fill power of at least 650FP. Fill weight for each panel follows the spec sheet, with a total of 280g and a tolerance of ±5%. Recycled down, synthetic fill or any undisclosed duck down must not be mixed in, and no panel may be replaced with synthetic fill. The down supplier, material code and production factory may not be changed without our written consent. For each lot, submit the incoming material lot number, bale label photos, down supplier test report and filling records; if RDS is claimed, provide the transaction certificate (TC) for that lot. Test samples will be randomly drawn from finished goods and sealed by our inspector. In case of dispute, third-party lab results will prevail. If a sample fails, increase the sample and recheck; lots that still fail will be held and refilled or remade with compliant material.)


10. Three Real Cases

These are anonymized versions of patterns we often see in consultations. Amounts and figures are examples.

Case 1 — Goose down long puffer coats

This is the case from the opening. When the brand confronted the factory with the test results, it held out, saying "the supplier report says 90%." But bale label photos showed a 70/30 duck down code from the same supplier mixed in, and there were no filling records at all. The spec sheet only said "goose down 90/10, 280g" without a material code, fill weight by panel or test method, and that became the weak point. Returns, label corrections and seasonal discounting added up to a loss of about 65 million won, and the factory covered only part of the material price difference. From the next season, the brand writes the material code and per-baffle fill weight into the spec sheet and opens one garment at every pre-shipment inspection to weigh it.

Case 2 — Goose down duvets

A bedding brand that imported 1,500 goose down duvets kept getting messages that "the room smells fishy after sleeping under it." The only document the factory had provided was the down supplier's content report; no cleanliness test had ever been done. A lab test found content close to the label, but the oxygen number was nearly three times the limit. It was cheap, under-washed material. The brand now gets turbidity, oxygen number and odor rating along with content for every lot.

Case 3 — Children's lightweight down jackets

A children's apparel brand sold 3,000 lightweight jackets labeled "down 80/20," and some reviews said "only the sleeves are oddly thin." Opening one showed down in the body but polyester fiberfill in the sleeves and sides, panels you could not tell apart by feel. The brand now lists fill material and weight by panel in the BOM and opens each panel to check on first orders and every reorder.


11. Common Mistakes


12. Down Fill Control Checklist

Spec stage

Sample and production stage

Pre-shipment inspection stage


Closing - Check With a Scale and Scissors, Not the Label

A puffer leaving the factory looks equally lofty and warm whether it holds 90% down or 60%. Only after customers have worn, compressed and washed it for a few weeks do the leaking feathers and lost loft become obvious. Writing the material code, fill weight by panel and test method into the spec sheet, and opening one garment per lot to put on a scale, does not cost much. Set one garment and a few hundred dollars in lab fees next to tens of thousands of dollars in seasonal stock, and it is clear which is cheaper.

There is also something you can do right now. Ask the factory currently in production for the down supplier and material code, incoming lot numbers, bale label photos and per-panel filling records. If the reply is a single line, "都是按标准做的" (everything was done to standard), put together a down control conditions sheet and send it before your next order.

StartMade supports you on the ground, from writing spec sheets with material codes and per-panel fill weights to recording approved sample content, checking incoming material and filling records during production, and pre-shipment inspections where we open garments ourselves. If you have already received leakage or content claims, we start by testing remaining stock; if you are preparing a new season's product, we start with material grade and factory selection.

Is the Down in Your Jackets the Content on the Label?

From setting material codes and per-panel fill weights to incoming checks, opening and weighing, and pre-shipment inspection
We check on the ground before leakage claims arrive

Phone   010-9980-9959
Email   deanpark@greenfrogseoul.com
KakaoTalk   pf.kakao.com/_XkfuX
Website   www.startmade.co.kr

Frequently Asked Questions

How do you test down content?
Accurate down content comes from a lab that removes the fill and hand-sorts down clusters, feathers and residue to calculate weight ratios (GB/T 10288 or IDFB method). On site, the most you can do is open one garment, weigh the fill by panel and look for quill fragments or synthetic fill. Testing the approved sample the same way gives you a baseline to compare production against.
What is the difference between goose down and duck down?
Goose down has larger clusters and better recovery, so it is warmer for the same weight, while duck down has smaller clusters and costs less. Duck down with high content and proper washing is still a good fill. The real problem is products labeled "goose down" with duck down mixed in, so check the goose to duck ratio when you test content.
Is it a defect if feathers come out of a down jacket?
A few stray fibers are normal for down products. But if quills poke through the fabric every time you wear it, the feather ratio is likely high or the downproof fabric and stitching are not good enough. Open one baffle to see how many quill fragments are mixed in, and if it looks far from the labeled content, send it to a lab for content analysis.