Same Shoe, Different Factory

About one in five of the tips that come into Worse On Purpose’s tip line are about running shoes. I have seen some version of the following statement countless times: “I’ve run in [brand] shoes for ten years, and the last pair I bought fell apart in a couple months”

As an avid runner who has also noticed radical fluctuations in quality between pairs myself, I was excited to dig into this category.

The research revealed a mechanism unlike the ones we typically cover. There is no private equity firm to point a finger at in this story. No debt, no flip, no rollup. It starts with a foam innovation in the mid-1970s that moved the part of the shoe that wears out from the outside, where a cobbler could replace it, to the inside, where it cannot be repaired and often cannot even be seen.

Within a generation, that one change caused an entire industry to abandon control over how its own product is made.

What follows tracks how that happened, what the companies themselves disclose about who builds their shoes, and what it means for you when choosing your next pair.

There was a time, long ago, when athletic shoes were repairable. This is evidenced by a product a few of you may know about: Shoe Goo.

Shoe Goo was invented by Lyman Van Vliet, a 45 year old executive at Hughes Aircraft, working out of his garage. He was a tennis player who was “dissatisfied with the durability of the soles of his tennis shoes” and “sought a method to extend their life by repairing them”.

He invented Shoe Goo and founded Eclectic Products to sell it, advertising in tennis magazines and pro shops.

The product found its real market in the rapidly expanding jogging crowd. Joggers were hard on shoes, and Van Vliet had given them a way to fix the part that wore out.

In 1972, the part that wore out was the outsole. It sat on the outside of the shoe where you could watch it start to go and easily complete a DIY repair with a tube of glue. A cobbler could do the same job properly.

In 1974, Brooks introduced the Villanova. It was the first running shoe with an EVA midsole, developed with input from Marty Liquori, an Olympic middle-distance runner and Villanova alumnus. EVA came in as an alternative to heavier, less responsive rubber. It made shoes lighter and softer. Better for running.

The unintended consequence was that it changed how shoes wear out.

EVA does not abrade, it compresses. The failure is structural and it happens in the middle of the shoe, where no glue or cobbler can easily reach it. Worse, it happens invisibly. A dead midsole looks no different to a new one, at least not to a casual runner examining her shoes.

Incidentally, Shoe Goo did fine. It found other markets and is still sold today. I became aware of it during my misspent youth as a skateboarder. Skaters still repair their shoes to this day, long after runners stopped.

Running went the other way. Inside a generation, the category moved from a quality item to a disposable one. For those of us that care about quality and pride on the part of the producer in their work, that transition was stark.

We now replace running shoes on a schedule. We do not argue with the schedule, we do not try to fix them, and for most of us it has never occurred to us that we could.

Ask a runner when to replace your shoes and you will likely hear a number, rather than a description of condition. Three hundred miles. Five hundred. Somewhere in there. It is one of the most widely known rules in the sport, repeated by shops, coaches, training apps and the brands themselves.

This viewpoint appears to stem from a single study, published in 1985.

Cook, Kester and Brunet, writing in the American Journal of Sports Medicine, put running shoes on a machine and measured how much shock absorption they lost as miles accumulated. They tested at 0, 5, 10, 25, 50, 75, 100, 125, 150, 200, 250, 300 and 500 miles. The shoes held about 75% of their cushioning at 50 miles, about 67% at 100 to 150, and dropped below 60% somewhere between 250 and 500.

Note that five hundred miles is just where the testing stopped, not some magic threshold experimenters discovered. The rule every runner now takes to heart is bracketed, at its upper bound, by how far experimenters decided to run a machine in 1985.

There are two other important findings from that same study did not survive into folklore.

The first is that price and manufacturer had no bearing on how quickly shoes degraded. The cheap ones and the expensive ones died at the same rate.

The second is that when the researchers put the shoes on actual runners instead of the machine, the degradation was "not as great." Not as great is generous. Runners' shoes still held about 70% of their cushioning at 500 miles, a level the machine had them hitting by 150 miles. Real running wore shoes down around three times slower than the test that produced the rule.

Nike's own website is a good example of how brands now apply the number in contradictory and confounding ways. On a page titled "How Many Miles Are Running Shoes Good For?", Nike cites a 2011 review in Footwear Science stating that "high-quality running shoes can last for more than 600 miles."

Just a few beats down on the same page, Nike advises that "experts generally advise that you get new running shoes every 300 to 500 miles."

Then it converts that into a calendar. Casual runners, every 7.5 to 12.5 months. Training for a 5K or 10K, every 5 to 8 months. Half marathon, every 4 to 6 months. Marathon training, every 2 to 3 months.

The same page concedes that there is no set rule.

Clearly, these claims cannot all be true at once. Yet the brands will keep making them, because the replacement cycle benefits them massively.

My advice is to focus less on mileage, and more on your personal experience with the shoe. Examine it, be conscious of how it feels as you run. Base your decision to replace on your personal needs and experience, not a set mileage.

Trusting your own judgement is not easy in a market that sells you the way running brands do, though.

If you have bought running shoes at a specialty store, you know the routine. You run on a treadmill for thirty seconds. Someone films your feet. You watch a slow-motion replay of your ankles while a staff member explains that you overpronate, or underpronate, or are neutral. Then you are walked to the shelf that matches the verdict. Stability. Motion control. Neutral.

It is a good ritual. It feels like being measured for something.

In 2009, three researchers went looking for the evidence behind it. Richards, Magin and Callister, writing in the British Journal of Sports Medicine, searched for any study testing whether prescribing running shoes by pronation type actually reduced injuries.

They found none. Their words: "No original research that met the study criteria was identified." Their conclusion was that "the prescription of this shoe type to distance runners is not evidence-based."

Five years later somebody did the original research. Nielsen and colleagues followed 927 novice runners for a year, all of them in the same neutral shoe regardless of foot type. They published in the same journal, under a title that leaves little room for interpretation: "Foot pronation is not associated with increased injury risk in novice runners wearing a neutral shoe."

People inside the industry have been saying versions of this for years. Spencer White, then a vice president at Saucony, told Outside that "there is no correlation between how much the foot moves and who is getting hurt. It's not a motion thing, it is a stress on the body thing." In the same piece, the physical therapist Jay Dicharry asked whether anyone really believes "an 8 to 13 ounce piece of material is going to stop motion in a 180 to 200 pound skeleton."

The industry agrees, and has been quietly walking away from pronation control for years.

The same Outside article cited above traces Brooks’ removal of the medial post, a firm wedge of foam on the inside of the midsole, which was the entire physical basis of a stability shoe. Brooks started removing it some years before 2020, when the Beast launched without one. By 2021 even the Addiction, its maximum support model, lost its post. The rest of the market followed, replacing posts with guiderails, sidewalls and rocker geometry.

They were right to. The science was never there.

But there was no big announcement. Nobody wrote to the runners who had spent twenty years being told they needed a particular kind of shoe to mention that the reasoning behind it had been quietly dropped.

And the treadmill, along with the bogus gait measurement, is still in the store.

The apparatus of expertise around running shoes is elaborate. The mileage rule, the gait analysis, the categories, the vocabulary of pronation and support and stack height. The brands control their sales and marketing strategy closely.

What they don’t control are the factories producing their shoes. The brand on the box does not own them. It does not employ the people inside it. In several cases it will not commit to buying from it past the current order.

I wish I could say that last sentence was hyperbolic. It is not. Here’s what the major shoemakers disclose in their filings:

Two brands own factories. The rest rent.

Once a shoe could no longer be fixed and the customer no longer expected longevity from it, owning a factory, and with it control over production quality, stopped being worth the cost. In fact it became a handicap. A brand that owned factories was competing against brands that only owned designs. Within a couple decades, the entire market let go of production control.

And they rent from a small pool. Nike's own filing states that "four footwear contract manufacturers each accounted for greater than 10% of footwear production and in the aggregate accounted for approximately 59% of NIKE Brand footwear production."

Four companies make well over half of all Nike shoes.

One of the largest of those contractors is Yue Yuen, which shipped 252.2 million pairs in 2025 at an average factory price of $21.00 a pair (versus $35-$45 in the US). Its named customers include Asics, Adidas and Nike. Three rivals, all manufacturing in one building.

Of Nike's roughly thirty suppliers, at least 22 are Taiwanese firms, accounting for over 60% of the 550 million pairs Nike sells each year. One industry executive describes Feng Tay, another massive contractor, as "the TSMC of Taiwan's shoe industry."

When a brand does not own the factory, quality is not something it controls. The best it can do is to specify some level of quality, then hope it is achieved. The brand writes the standard. Somebody else, in another country, decides how closely to hold it.

That would be manageable if the arrangements were stable. Often they are not. Deckers, which owns Hoka, tells its investors it buys "on the basis of individual purchase orders, rather than maintaining long-term purchase commitments." In the same 10-K, Deckers also says it "continue[s] to diversify our independent manufacturers and the regions in which they operate."

Think about what that means in practice. The same shoe, sold under the same name at the same price, may be built next year in a different factory, in a different country, by people who have never made it before.

That is the mechanism behind the reader frustration this piece started with. You can’t move production lines searching for fatter margins and expect quality to be maintained.

It is important to note that just because a brand uses contract manufacturers does not necessarily mean that the product is bad. Apple owns no factories. Yue Yuen and Feng Tay are highly capable firms building complex products at enormous scale.

Given that virtually every player in this industry outsources their manufacturing, the simple fact that a brand uses contract manufacturers is not a good indicator of quality. In order to assess likely future quality, we must examine how deep these contractor relationships are, and how they are maintained over time.

As we’ve established, only two major running brands own the factories that make their shoes. If buying from a brand that owns the means of production is a strict criterion for you, I’m afraid your shortlist is very limited:

New Balance is the more available of the two. It runs five company-owned US factories in Maine and Massachusetts employing around 1,300 people, and recently put $65 million into doubling capacity at its Skowhegan plant. It is worth knowing that its Made in USA line carries a disclaimer of "70% or greater" domestic value, and that a class action filed in December 2021 alleged roughly 30% of components in those shoes are made elsewhere. The case was voluntarily dismissed in 2024.

Mizuno is the other, stating in its own sustainability reporting that it "remain[s] committed to manufacturing in Japan" alongside its contracted production.

That is not a silver bullet. It is two brands, and purchasing from neither guarantees a shoe entirely manufactured in-house.

For everyone else with more lenient criteria, the useful question is not whether a brand contracts out, but how it maintains its contractor relationships.

Adidas discloses that it has worked with 65% of its independent manufacturing partners for at least ten years, and 37% for over twenty. That is a company with some skin in the game. If one of their older partners leaves, they lose something truly valuable.

Brooks does not own anything, but it publishes its final-assembly factories by name and location and discloses all four tiers of its supply chain.

Deckers, by contrast, tells investors in its annual filing that it buys "on the basis of individual purchase orders, rather than maintaining long-term purchase commitments," and that it "continue[s] to diversify our independent manufacturers and the regions in which they operate."

None of the companies are hiding this. Their supplier practices are right there in the annual reports.

So, practically:

Judge the shoe, not the odometer. Nike's own page cites research saying more than 600 miles.

Ignore the fitting verdict and buy what feels right on your foot. Two studies, in 2009 and in 2014, found no evidence the prescription prevents injury.

And if you find a model that genuinely works for you, buy a second pair while it exists. Next year's version may be built in a different country by people who have never made it before.

A very brief honorable mention for three smaller brands that seem to be pulling in the opposite direction:

NNormal, founded by the mountain runner Kilian Jornet with the Spanish family-owned shoemaker Camper, sells a shoe built around three interchangeable midsoles that you swap yourself. Given that the midsole is the part that dies, this is an interesting approach.

Vivobarefoot runs an actual resoling service in the US, UK and EU through a restoration company called NuShoe. They are upfront that it only covers a handful of models so far. You also have to be ok with running in a barefoot footbed. Not for everyone.

Carson Footwear makes what it calls "American-made running shoes built by hand."

I have not tested any of them. Everything above comes from the companies themselves, and after spending this piece demanding evidence from everyone else, I am not going to recommend shoes on the strength of their own marketing, however cool I think their business models are.

Your mileage may vary.

As always, you can track the status of the brands mentioned in this essay in the ‘Footwear’ section of The Brand Ledger.