Thread Count, Fill Power and Other Misleading Numbers

Macy’s was sued in 2017 over bedsheets it sold as 900 thread count. The lawsuit alleged that the true count was only 249. Macy's ultimately settled the case for $10.5M. The court estimated the class at “around 2.5 million individuals,” everyone who bought one supplier’s cotton-poly sheets from Macy’s over nearly 10 years.

If the lawsuit was right, then Macy's number was simply false. But a true number can mislead just as well. In almost every industry I've looked at, manufacturers print numbers that sound precise but tell you very little about the product on their own.

Thread count, down fill power, horsepower and hardness ratings are a few examples of common measurements that manufacturers love to quote. Manufacturers print them not to inform, but to drive sales.

For today’s article I picked five numbers I know well. For each of them, I'll explain what the number actually measures, what it leaves out, and what to check instead.

Many of you will know the numbers in your own areas of expertise better than I do. Please do share that knowledge in the comments, or in the community, if you’re a member.

Thread count is the number of threads in a square inch of fabric, counted in both directions. What is often left out is ply. Ply counts how many individual yarns are twisted together to create one piece of thread. By counting each ply as a separate thread, manufacturers can arbitrarily double or triple their thread count number.

In 2005, the FTC issued guidance to the National Textile Association on how sheets should be labeled. James Kohm, the FTC's associate director for enforcement, wrote that sellers should "state count followed by ply, as in: 300 thread count, two-ply yarn." Calling that same sheet a 600 "would likely mislead consumers about the quality of the product being purchased."

Bear in mind, this is just staff guidance, so it isn’t legally binding.

Even an honest count that follows the FTC guidance doesn’t tell you much. After washing sheets repeatedly in its tests, Consumer Reports "could not find any correlation between manufacturer claims of materials or thread count and sheet performance." One of its top-rated sets is a 350. A 550-count Macy's set sat near the bottom after it shrank too much to fit a 10-inch mattress.

What to check instead: look for "single-ply" on the package. Then read the material content, because the fiber itself tells you more than any thread count number about how the sheet will behave. In Consumer Reports' tests, cotton sheets "shrank up to 6 percent." The polyester sheet "hardly shrank at all," but it "doesn't absorb water," so some people find it sweaty. Bamboo viscose shrank by more than 15 percent.

Fill power measures how much volume down takes up. REI defines it as "how many cubic inches of loft one ounce of that down fill produces." The testing lab IDFL calls 700 and above "extremely high quality." The fill power on a label is allowed to be off by 5% in either direction.

What fill power leaves out is fill weight: the amount of down actually in a jacket or sleeping bag. The point of down is warmth, and to compare the warmth of two jackets you need both numbers. Fill power alone tells you how light and packable the down is. It can't tell you how warm the jacket is.

REI sells a lot of down jackets, and its own guide admits "Very few jacket makers report fill weights." Patagonia's Down Sweater page, for example, lists "800-fill-power" down and a total weight of 13 ounces, with no figure for the down itself.

Here's how the two numbers interact, using a couple hypothetical jackets. Jacket A holds 3 ounces of 800-fill down, giving it 3 × 800 = 2,400 cubic inches of loft. Jacket B holds 4 ounces of 650-fill down, for 4 × 650 = 2,600 cubic inches. Jacket B is the warmer of the two, even with the smaller fill power number.

Higher fill power buys you less weight and bulk for the same warmth. In REI's words, "In theory, 50 g of 900-fill-power down would be the same warmth as 100 g of 450-fill-power down; it would also compress to about half the space."

As a backpacker and general outdoor tomfoolerer, I think high fill power is worth paying for. My frustration is how rarely brands quote both numbers, which makes clean comparisons between options almost impossible.

The FTC's guidance covers what's in the fill. A product labeled "down" should be "more than 70%" down, and a manufacturer "may not 'shoot for the tolerance.'" It says nothing about fill weight.

What to check instead: ask for the fill weight in ounces or grams. If the brand won't give it, be skeptical, since the fill power alone doesn’t tell you much.

Leather grades describe which layer of the hide was used and how its surface was treated. Each grade suits different things. Not everything needs to be full grain, and full grain does not mean high quality.

Full grain keeps the outer layer intact. It's the most durable grade and develops a patina, so it's the one for boots, belts and bags you plan to keep for years.

Top grain is the outer layer sanded and refinished for an even look. It gives up some durability and patina for suppleness, and works well for jackets, handbags and furniture.

Corrected grain is treated more heavily to hide flaws. It's cheaper and holds up to wear, so it shows up in budget furniture, car seats and shoes.

Split leather and suede come from the lower layers of the hide. They're softer and weaker than grain leather, but split cowhide is common in work gloves because it resists abrasion and costs less.

Bonded leather is shredded leather fiber mixed with a binder. It's the cheapest and least durable grade, and prone to cracking. Avoid.

None of these categories have a legal definition. In 2008, the leather industry's trade group asked the FTC to define the grade words, starting with "top grain or full grain leather." The FTC declined:

The FTC’s Leather Guides have not changed since 1996. They say the bare word “leather” should only describe a product “composed in all substantial parts of leather,” and that bonded material must state its share of leather fiber.

So the “genuine leather” stamp you see everywhere tells you only that the material is leather. Every grade outside of bonded leather qualifies. It tells you nothing of the quality of the product.

What to check instead: the grade on the tag is only a starting point. There's good full grain and bad full grain. Telling them apart takes a trained eye. If you don't have one, borrow an expert's. Bedo's Leatherworks, a third-generation cobbler, and Rose Anvil, who cuts boots in half to show how they're built, are two of the best places to start.

Wet/dry shop vacuums are sold on peak horsepower. Often the number is printed right there in the product name. Ridgid sells a "6 Gal. 3.5 Peak HP" model at Home Depot. Shop-Vac sells a "5 Gallon 6.0 Peak HP" model. Its product page explains what the number means:

That disclosure came out of a 2016 settlement. The lawsuit claimed an independent lab found the vacuums generated "only a small fraction of the advertised 'Peak Horsepower.'"

You can check the number yourself directly from the spec sheet. The 6.0 peak horsepower Shop-Vac is rated at 11.8 amps on a 120-volt circuit, and 120 × 11.8 = 1,416 watts. One horsepower equals 746 watts, so the most this motor can draw from the wall is about 1.9 horsepower. It delivers less than it draws. Producing 6.0 horsepower would take 4,476 watts, or about 37 amps.

The numbers that matter are airflow, measured in cubic feet per minute (CFM), and sealed suction, measured in inches of water lift. Air watts combine the two. According to Pro Tool Reviews, airflow keeps debris moving, and water lift is what gets it moving in the first place.

The Ridgid, rated at 3.5 peak horsepower, lists 64 CFM, 67 inches of water lift and 186 air watts. The Shop-Vac, rated at 6.0, lists 90 CFM, 50 inches of lift and 180 "peak air watts." The Shop-Vac moves more air, the Ridgid pulls harder, and their air power is about the same. Peak horsepower alone tells you none of that.

What to check instead: look for air watts, CFM and water lift on the spec sheet. The amp rating tells you how much power the motor really draws.

A knife's performance depends on three things: the steel, its hardness and the angle of its edge.

Hardness is measured on the Rockwell C scale, written as HRC. Wüsthof prints 58 for its Classic line. A higher number means the steel resists wear for longer, but that comes at a cost. Larrin Thomas, who runs Knife Steel Nerds, explains it: "Higher hardness/strength reduces the toughness of steel. Toughness is the resistance to fracture or chipping."

The type of steel used lets you look up its mix of toughness, edge retention and rust resistance. Those properties trade off against each other, so the right steel depends on the job. A hard-use outdoor knife needs toughness. A kitchen knife needs a thin edge that won't roll or chip.

The thing that matters most, though, has nothing to do with materials or manufacturing. A knife’s edge angle is set by the owner every time they sharpen their blade. How well a knife’s edge has been maintained is more important than its steel type or hardness.

Thomas puts it plainly: "sharpening in terms of grit finish and edge angle have a greater effect on performance than steel choice." He tested 48 steels and built a model from the results. In it, four extra points of hardness added about 63 mm of cardstock cut. Sharpening the edge 5 degrees thinner per side added about 89.

What to check instead: a named alloy, and hardness as a tradeoff. Then put your effort into keeping the edge sharp.

That's why I'm comfortable using Victorinox knives exclusively, even though Victorinox publishes no steel grade or hardness. America's Test Kitchen, which tests sharpness with an industrial machine, calls the Fibrox Pro 8" chef's knife its "top choice for decades."

Three of the five numbers discussed here have no outside standard behind them. Sellers choose how to count threads, and the FTC's guidance on ply isn't binding. Leather grades have no legal definition. Peak horsepower is a figure from the maker's own lab tests. The numbers that hold up better are measured with a standardized test that someone other than the seller wrote. Ridgid, for instance, says its air watts are "tested to ASTM standard."

One great example of standardized testing is the grade stamp on a piece of lumber. A stamped 2x4 shows the grading agency, the mill number, the grade, the species and the moisture level when it was graded.

The grading rules sit under PS 20, a voluntary federal product standard. The American Lumber Standard Committee's Board of Review "accredits agencies to grade and inspect under those rules and monitors the agencies' performance." The agencies then send inspectors to the mills.

A standardized test does two things. It makes products comparable, because every brand is measured the same way. It also holds companies’ marketing claims to account. The Macy's lawsuit could set 900 against 249 because a thread can be counted. There is no equivalent number to dispute when a wallet says "genuine leather."

In Between Craft and Quality, I argued that products have grown more complex while each of us has become less able to detect quality at the point of sale. I wrote then that "a quality nobody can detect is a quality nobody pays for, and a quality nobody pays for stops being made."

A shared, independent test is one way to make quality detectable again. When buyers can compare real numbers, the better product is more likely to win, and a company that cuts corners shows up in the results.

Perhaps it's time to bring standards like these to more consumer industries. Exactly how is an open question. My instinct is that independent private bodies, like the lumber committee, would do it best. The government could do it too.

I have shared only a few examples of misleading numbers from the categories I am most familiar with. I am certain that there are many more. If you work with fabric, wood, steel, engines or electronics, you'll know which numbers matter in your trade, and which ones a seller leaves off.

Tell me yours in the comments. If you're a member, post it in the community, where readers compare notes on exactly this. I’ll follow up with a more complete list once we’ve built it.

Thanks as always for reading.