You Don't Have a Safety Gear Problem. You Have a Specification Problem.

I review 350 to 450 safety product submissions a year before they're allowed into our catalog. Four years of doing it now. In 2024 I rejected 22% of first submissions, and almost none of those rejections were about the product being bad. They were about the product not being defined. (Mental note: build the rejection-reason dashboard I promised my boss in January. It's February. Fine.)

But the submissions aren't the interesting part. The questions are. Here's a sample from the last eight months:

• "Can I use nitrile gloves for food prep?"
• "Do I need dedicated shooting safety glasses, or are my regular ones fine?"
• "Was Charlie Kirk wearing a ballistic vest?"
• "Is the Radians RadWear safety vest I ordered actually Class 2, or is it just orange?"
• "Why is a high tensile fence listing sitting next to Radians coveralls in the same search results?"

Five questions. Four industries. No obvious relationship between them. Same shape though — all five assume the hard part is picking the right item off a list.

The problem isn't the picking

By the time somebody types "can I use nitrile gloves for food prep" into a search bar, the real decision was already made somewhere upstream — the category, the price band, the type of seller. What's left is confirmation.

Which would be fine, except the upstream decision was usually made by someone who never saw the hazard. OSHA 1910.132(d)(1) puts the hazard assessment on the employer, and in a lot of companies that assessment gets done once, filed, and then never consulted again when somebody needs to reorder gloves. The person clicking "buy" is three steps removed from the person who knew what the risk actually was.

I've been on both sides of that gap. When I first started doing vendor qualification, I assumed the problem was bad data — that if we just got cleaner spec sheets, we'd get better decisions. Three years in, I don't believe that anymore. We have plenty of data. What we don't have is anyone asking what the product is supposed to do before asking what it costs.

Three things are making it worse

1. "Ballistic vest" isn't a specification

Every time a public figure gets shot, search volume for ballistic vests spikes, and the question comes in: was he wearing one? I'm not going to speculate about that case — I don't know, and it isn't the part I care about. What I care about is that the question sounds answerable and isn't.

"Ballistic vest" describes a category with at least two completely different products inside it. Soft armor — what most people picture, the thing you wear under a shirt — is tested against handgun threats. NIJ Level II and IIIA under the 0101.06 scheme. Rifle threats require hard armor plates: different weight, different fit, different procurement, different lead time.

Soft armor rated for handgun threats and hard plates rated for rifle threats are not upgrades of each other. They're responses to different problems. Buying the wrong one doesn't give you less protection than you expected — it gives you protection against a threat you weren't facing.

So the honest answer to "was he wearing a ballistic vest" is that the question can't resolve into yes or no until you specify the threat. And a buyer who doesn't know that distinction is going to end up making a deployment decision based on a feeling instead of a rating.

2. The datasheet is the product

In PPE, the paper and the product are effectively the same thing. A vest without a legible ANSI/ISEA 107 class marking isn't a Class 2 vest. It's orange fabric. An eye protector without the Z87+ marking isn't high-impact rated, no matter what the product photo suggests.

I've had vendors argue with me about this. "It's the same material, we just didn't print the label." Doesn't matter. The marking is how a safety officer, an inspector, or an auditor knows what they're looking at. If it can't be verified at the point of use, it doesn't count — because the person enforcing it can't count it.

The nitrile glove question is the same problem in a different costume. ASTM D6319 tells you a glove was made to a medical exam glove standard. It says nothing about food contact. Food contact is FDA territory, and what your inspector will ask for is a declaration or documentation citing the applicable food contact regulation. A glove can be a perfectly good exam glove and still show up at your inspection with an empty file folder behind it.

Point being: the material isn't the question. Latex versus nitrile isn't the question either, despite what every forum thread will tell you. The documentation is the question.

3. The catalog got flat

Third thing, and probably the biggest: the marketplace flattened everything.

It used to be that you bought gloves from a glove distributor who knew something about gloves. Now the same search results page holds a case of Radians coveralls, a high tensile fence supplier, a no-name hi-vis vest, and three sellers who turn out to be the same company. Everyone looks the same. The seller name stopped being a signal years ago.

I have vendors who list 40 categories and stock none of them. When I ask for lot-level documentation, I get a link to a general ISO certificate from 2019. That's not a compliance problem in the abstract. It's a specific problem the day something fails and I need to know which production run it came from.

What the wrong answer costs

Not "could cost." Does cost. Three examples from the last two years.

Food service: the whole case gets pulled

A food prep customer in 2023 was using gloves that passed every visual check — good fit, no tears, arrived on time. Then a county inspector asked for the food-contact documentation and nobody had it. They pulled 60 cases and switched suppliers mid-quarter.

The gloves were never the problem. The file folder was. Sixty cases, a rush reorder at a premium, and about three weeks of a kitchen manager's time. Landed north of $4,000 — caused by a missing PDF.

Eyewear: the rating you assumed

Shooting safety glasses are where assumptions do the most damage, because the category name sounds like it covers everything. ANSI Z87.1 eyewear is tested against a high-mass impact at a defined velocity. That's real protection, and it's the floor. It was not designed around the fragmentation threat profiles that military-oriented eyewear specs test for — different projectile, substantially higher velocity, much higher energy.

I'm not saying Z87.1 eyewear is inadequate. I'm saying a buyer who reads "safety glasses," assumes every pair clears the hardest standard available, and hands them out at a range is making a bet they didn't know they were making. The marking on the lens is the only way to know which pair you actually have.

High-visibility: the stop-work order

A customer running night operations near traffic ordered hi-vis vests at a price that made sense for their volume. They arrived, and they were hi-vis in the sense that they were bright. The reflective material didn't meet the class requirement for their work zone speed, and a state inspector shut the shift down for two hours while they swapped out 14 vests.

That's the version of this problem I care about most, because nothing about those vests looked wrong. They looked right. The buyer had no reason to doubt them until someone who knew the standard walked up.

What we changed

Two things. Neither is clever.

We buy specs, not products. Every purchase order now names a standard instead of an item. "ANSI/ISEA 107 Class 2, Type R" rather than "hi-vis vest." "Z87+ marked" rather than "safety glasses." "Food-contact declaration on file" rather than "nitrile gloves." It reads like overkill on a small order. It also means I never have to explain to a customer that the thing they bought is fine but isn't the thing they needed.

And we pay for the delivery date. This is the part nobody likes. When we need stock in four days and the cheap supplier says "probably four to six business days," we don't take the bet anymore. We pay more and get a committed date with a stocking warehouse behind it.

I used to think rush premiums were a markup on convenience. They aren't. They're the price of knowing. In March 2024 a customer paid $400 in expedite fees on a small order so their crew wouldn't be standing around waiting for vests to clear a regional warehouse. Missing that work window would have cost them somewhere in the low five figures. Four hundred dollars to remove a maybe is one of the better trades I've watched.

Looking back

If I could redo my first two years here, I'd stop treating spec questions as a customer service problem and start treating them as a procurement problem. We were answering the questions well. We just weren't changing anything that generated them.

None of the questions above are dumb. Nitrile gloves for food prep, ballistic vest ratings, shooting glasses, RadWear class markings, fence listings appearing where they shouldn't — those are all reasonable things to type into a search bar. They're just being asked at the wrong end of the process. Somebody upstream picked a category and a price band, and now everyone downstream is trying to make it fit.

Fix the upstream part. The rest gets a lot quieter.