It’s not the cap. Not really.

I got a call last spring from a customer who’d just received 5,000 16-oz amber Boston rounds. About 300 had leaked in transit—soaked through the corrugate, labels peeling, the whole mess. Their first instinct: blame the cap. And I get it. The closure is the most visible point of failure. But when we ran our post-mortem, the root cause wasn’t the cap’s thread pitch or the liner material. It was the finish—the glass neck’s dimensions.

It’s tempting to think a tight seal is purely a function of the closure. But identical caps from the same batch performed perfectly on jars from a different production run. The variance was in the glass, not the plastic. That’s the kind of headache that costs you a $22,000 redo and delays your launch by six weeks.

Let me walk you through what I’ve learned reviewing over 200 unique packaging SKUs annually—and why the spec you’re probably ignoring is the one that matters most.

The surface problem: Leaks look like a closure issue

When a jar leaks, the obvious culprit is the lid. It’s the part that moves, the part you torque, the part you see. So you check the cap: is the liner seated properly? Is the thread count right? Did the capper over-torque and warp the seal?

But here’s the thing: a cap can be perfectly molded, with the correct liner material and proper torque application, and still fail if the glass finish isn’t within tolerance. The finish is the rim of the jar—the flat sealing surface and the threads molded into the glass. If that surface has even a 0.5mm deviation, your cap’s liner can’t compensate.

I still kick myself for not catching this earlier in my career. Our first big quality audit in Q1 2022 revealed that 4% of the jar inventory had finish dimensions outside our spec. The vendor claimed it was “within industry standard.” They weren’t wrong—but industry standard is a range, not a target. We tightened our spec by 0.3mm on the sealing surface flatness. Rejection rate dropped to under 0.5%.

“It’s tempting to think a tight seal is purely a function of the closure. But identical caps from the same batch performed perfectly on jars from a different production run.”

The deeper problem: Glass consistency is harder than you think

Glass is a tricky material. It’s not like injection-molded plastic where the tooling produces near-identical parts. Glass is made in a hot-end process—molten gobs dropped into molds, blown, annealed. The molds wear. The temperature fluctuates. The annealing lehr (the cooling oven) can vary by 10°C from one side to the other.

What does that mean for you? Variations in:

• Finish dimensions – The sealing surface isn’t perfectly flat, or the thread depth is inconsistent.
• Wall thickness – Thin spots create stress concentration, which can lead to cracking under vacuum sealing or thermal shock.
• Weight – If the gob weight drifts, you get jars that are heavier or lighter than spec. Heavier jars cost you more in freight. Lighter jars might not have enough glass in critical areas.

On a 50,000-unit run, even a 1% variation in weight adds 500 lbs of unexpected shipping cost. Or, worse, 500 units that fail.

Most buyers focus on price per unit and cap style. They don’t ask about the glass manufacturer’s process capability index (Cpk). I’ve had vendors quote me a beautiful price only to discover their Cpk on finish flatness was 0.8—meaning nearly 3% of jars would be out of spec. We aim for 1.33 or higher.

The specific spec you should be checking

If you’re sourcing glass containers, here’s the single most important dimension to put in your spec: sealing surface flatness. Per industry standards (ASTM C147-86), the acceptable deviation is typically 0.030 inches (0.76mm) for continuous thread finishes. But that’s a wide band. We specify 0.020 inches (0.5mm) maximum. It adds about $0.01–0.02 per jar, depending on volume. On 10,000 units, that’s $200 for a measurable reduction in leakers. In my experience, that trade-off pays for itself in fewer returns and happier customers.

The cost of ignoring this: Not just leaks

Leaks are the dramatic failure. They’re the ones that get photographed and sent to your quality team in an angry email. But the less visible cost of inconsistent glass is chronic underperformance.

• Vacuum seal failures – For food products requiring a hermetic seal (jams, sauces, pickles), finish variations mean some jars don’t hold vacuum. You lose shelf life. You lose product.
• Label application issues – If the jar’s wall thickness varies, so does its diameter. Label applicators that expect a tight tolerance can misplace labels or create wrinkles.
• Cap torque variability – An out-of-round finish means the cap applies uneven pressure. The capper applies the same torque to every jar, but the effective sealing pressure varies. The result: some jars are over-tightened (liner extrusion) and some are under-tightened (leaks).

I ran a blind test with our packaging team: same jar, same product, same cap. We sorted jars by finish flatness (within 0.5mm vs. 0.8mm deviation). The “tight spec” jars had zero leakers in a 500-unit simulated transit test. The “industry standard” batch had 3 leakers—0.6%. On a 50,000-unit order, that’s 300 leakers. At $3.50 per jar (product + container), that’s $1,050 in direct spoilage. Add shipping, handling, and customer goodwill damage, and it’s easily $2,000–3,000.

“Most buyers focus on price per unit and cap style. They don’t ask about the glass manufacturer’s process capability index (Cpk). We aim for 1.33 or higher.”

So what do you do about it?

This is the part where I could give you a shopping list of specs and testing protocols. And if that’s what you need, I’m happy to share our template. But the real takeaway is simpler: stop assuming the cap is the problem.

When you’re qualifying a new glass supplier—or even checking a batch you’ve already received—ask for finish dimension data. Specifically:

1. Finish diameter (T-dimension) – the diameter across the threads
2. Sealing surface flatness – the flatness of the rim the cap liner presses against
3. Finish height (H-dimension) – the distance from the sealing surface to the top of the thread

These three measurements, taken with a simple go/no-go gauge or coordinate measuring machine (CMM), will tell you more about leak risk than any cap inspection alone.

An informed customer asks better questions and gets better product. I’d rather spend 10 minutes on the phone explaining finish tolerances than deal with a $22,000 redo. And honestly, most vendors appreciate when a buyer knows what to look for. It separates serious orders from price shoppers.

At Fillmore Container (fillmorecontainer.com), we maintain finish tolerance data for every glass SKU we stock. It’s part of how we keep our rejection rate below 0.5% across 200+ SKUs annually. And while I can’t promise every jar will be perfect—nothing made of glass ever is—I can promise we measure what matters.

Pricing for reference only; verify current rates with your supplier. Finish specs based on ASTM C147-86 and in-house testing protocols effective Q1 2025.