I'm going to tell you about a problem that took me six months to solve.

I'm a product developer for an outdoor brand. We source down jackets from a factory in Jiangsu. The factory is good. They've been making down products for 15 years. They know what they're doing.

In 2025, we started getting complaints. Some jackets were warm. Some were cold. Customers said: "Same jacket, same size, but one kept me warm and the other didn't."

We sent the jackets to a lab. The down was tested for fill power. All the jackets were within spec. The filling weight was consistent. The construction was identical.

Lab said: "They're the same."

Customers said: "They're not."

I had to find out why.

The First Clue

I started by weighing the jackets. Not the filling weight. The total jacket weight. There was a 40-gram difference between the warmest and the coldest jackets.

Forty grams is not a lot. But it was consistent. The warm jackets were heavier. The cold jackets were lighter.

I weighed the down filling separately. The filling weight was identical. The difference was somewhere else.

The Second Clue

I cut open a warm jacket. Then I cut open a cold jacket. I laid them side by side.

The down was different. Not the fill power. The down clusters were different sizes. The warm jacket had larger down clusters. The cold jacket had smaller ones.

The fill power was the same. But the cluster size was different. Larger clusters trap more air. More air means more warmth.

Why were the clusters different sizes?

The Third Clue

I asked the factory: "Where does the down come from?"

"From the same supplier," they said. "We buy by the batch. The batch is tested for fill power. It all passes."

I visited the down supplier. The down was stored in large bales. Each bale came from a different source—different farms, different regions, different seasons. The bales were mixed together to achieve a consistent fill power.

But the cluster size was not consistent. Some bales had large clusters. Some had small clusters. The mixing process created a batch with the right average fill power, but the cluster size varied.

The Fourth Clue

I watched the filling process at the factory. The down was blown into the jacket chambers. The operator weighed each jacket after filling. The weight was checked. If it was off, they adjusted.

But the operator was also shaking the jackets. Shaking distributes the down. It helps it settle. But it also breaks down the clusters. Aggressive shaking breaks large clusters into smaller ones.

The operator was shaking some jackets more than others. The jackets that were shaken more had smaller clusters. The jackets that were shaken less had larger clusters.

The Fifth Clue

I asked: "How much do you shake the jackets?"

The operator said: "Until they look full."

"Show me."

He grabbed a jacket. He shook it. He shook it again. He shook it a third time. He was shaking it until it looked right.

The shaking time was not controlled. It varied by operator, by shift, and by how the jacket looked. Some jackets were shaken for 5 minutes. Some were shaken for 15.

The Sixth Clue

I tested it. I took two jackets from the same batch. I filled them with the same down. I shook one for 3 minutes. I shook the other for 15 minutes.

The 3-minute jacket was 35 grams heavier. The down clusters were larger. The warmth was better.

The 15-minute jacket was lighter. The clusters were broken. The warmth was worse.

The Full Picture

Here's what happened.

The down supplier was mixing bales to achieve consistent fill power. But the cluster size varied. The cluster size affected the warmth.

The factory was shaking the jackets to distribute the down. But the shaking was breaking the clusters. The shaking time was not controlled. Some jackets were shaken too much.

The result was inconsistent warmth. The lab tests measured fill power. They didn't measure cluster size. They didn't measure shaking time.

The Solution

The solution was simple and immediate.

We added a specification: "Shaking time shall be standardized to 3 minutes per jacket. No operator shall exceed 4 minutes."

We also added a cluster size test to our quality check. Not just fill power. Actual cluster size distribution.

The factory complied. The shaking time was reduced. The cluster sizes became consistent. The warmth became consistent.

What I Learned

Here's what I learned from this.

First: The lab doesn't tell the full story. The lab measured fill power. The fill power was the same. But the jackets were different. The lab didn't measure cluster size. The lab didn't measure shaking time.

Second: The process matters. The filling was correct. The down was correct. But the shaking was wrong. The shaking broke the clusters. The warmth was lost.

Third: The same spec can produce different results. The fill power was the same. But the cluster size varied. The spec was not the whole story.

Fourth: The operator's eye is not a measurement. The operator shook the jackets until they "looked full." That's not a measurement. That's an opinion. You can't control quality with opinions.

What I'd Tell Every Buyer

If you're buying down products from China, ask these questions:

Question 1: "Do you test cluster size, or just fill power?"

If they only test fill power, they're not catching the variation. Fill power is an average. Cluster size is the distribution around the average.

Question 2: "How do you control the shaking process? Is the shaking time standardized?"

If they don't control shaking time, they're introducing variation. The same down can produce different warmth depending on how it's handled.

Question 3: "Where does the down come from? Is it a single source or a blend?"

If it's a blend, ask about the cluster size variation. Blends can achieve consistent fill power with inconsistent cluster sizes.

Question 4: "How do you train operators to handle down?"

If the training doesn't include cluster protection, operators will break the clusters without knowing it.

The Aftermath

We still buy from the same factory. They changed their process. Shaking time is now standardized. Cluster size is now tested.

The jackets are consistent. The complaints have stopped.

The fix was simple. Finding it took six months.

The Question I Still Have

I still wonder: how many other down jackets are being shaken too much? How many clusters are being broken because nobody thought to measure the shaking time?

I don't know the answer. But now I always ask.

We tested the down in the lab. Fill power was identical. The jackets should have been the same warmth. They weren't. Th

Same factory. Same down. Same fill power. Lab said they were identical. Real-world tests said they weren't. Six months to find out why.

Jangan direproduksi tanpa izin：Blog-Pulviso » The Down Jacket Mystery: Same Factory, Same Fill Power, Different Warmth