Watch Quality Control & Inspection: How to Verify Quality from a China Watch Factory

*By JOEFOX Watch Factory, Guangzhou — Est. 1990*

---

Last November, a European buyer placed an order for 500 stainless steel automatic watches with a Guangzhou factory. He'd received three perfect samples. The dials were clean, hands aligned, cases polished without a blemish. He wired the deposit. Production wrapped in 28 days.

Then the pre-shipment inspector pulled 80 watches at random.

Thirty of them had visible scratches across the dial hands. Not micro-scratches you need a loupe to see — scratches visible from arm's length under normal light. The root cause took two hours to trace: a new assembly line worker had been stacking finished hands into trays without foam separators between layers. One tray of hands slid against another with every workbench vibration.

Thirty-eight percent defect rate on hands alone. Every single watch had to be opened, the hands pulled, fresh hands installed, case re-sealed, and water resistance re-tested. The shipment was delayed by two months. The buyer missed Black Friday. His brand's holiday-season launch window — gone.

A $500 third-party inspection before shipment would have caught this within 48 hours of the defect appearing. Instead, it cost him a season.

That story isn't unusual. In 35 years on a Guangzhou factory floor, I've watched it play out across thousands of orders: buyers who treat QC as a checkbox and buyers who treat QC as a system. The checkbox buyers learn expensive lessons. The system buyers build brands that last.

This guide is the system.

Key Takeaways

- A four-stage QC system (IQC → IPQC → FQC → PSI) catches defects at every production phase — fixing a dial flaw during assembly costs pennies; fixing it after shipment costs your brand reputation

- For a 500-unit watch order at AQL 2.5, you inspect 80 pieces — if 6 or more show major defects, the entire batch fails

- The #1 watch defect worldwide is dial dust contamination — and it's entirely preventable with a proper cleanroom assembly line

- Third-party inspection costs $160–$300 per visit; a bad batch that reaches customers can cost you your Amazon account

- A factory that refuses a third-party inspection is a factory that's hiding something — full stop

---

The Four-Stage QC System Every Watch Factory Should Run

Most buyers think "QC" means someone glancing at finished watches before they go in a box. That's not QC. That's a prayer.

Real quality control in watch manufacturing runs on four distinct stages. Each stage has different tools, different checkpoints, and different decision logic. Skip any one of them, and you're gambling with your order.

Stage 1: IQC — Incoming Quality Control

Before a single watch gets assembled, every component that arrives from our suppliers goes through IQC. This is the gate that stops bad material from ever touching a production line.

Movement pre-screening. Every batch of movements — whether Japanese Miyota quartz or Seiko NH automatic calibers — gets sampled on arrival. For quartz, we check current draw, battery voltage, and initial accuracy against a Weishi timing machine. A standard Miyota 2035 quartz should pull 0.55–0.90 μA with a fresh battery. Anything outside that range suggests contamination during movement assembly — and we reject the batch.

For automatic movements, the checks are more involved. On the timing machine, we look at three numbers: daily rate (ideally within ±15 seconds/day for a Seiko NH35), amplitude (270°–310° in the horizontal position), and beat error (no more than 0.3ms). A movement showing 0.8ms beat error means the impulse jewel isn't centered on the balance staff — it'll run, but it'll never keep consistent time across positions. Those go back to the supplier.

Case inspection. We measure every critical dimension on a sample of cases from each batch. Lug width (±0.05mm is our internal tolerance), case diameter, thickness at the movement cavity, crown tube threads. One undersized crown tube thread means the crown won't screw down properly. That means no water resistance. That means returns.

We also check the steel grade. 316L stainless steel contains 2.0–3.0% molybdenum — that's what gives it corrosion resistance beyond standard 304. A fast molybdenum spot test tells us whether the case is genuine 316L or 304 that's been labeled as 316L. Yes, it happens.

Dial and hands. Print clarity under 5x magnification. Lume fill — no bubbles, no overflow past the indices, even brightness under UV. Applied indices — no tilt, no glue residue visible at the edges. Hands — perfectly flat under a profile gauge, no bends, no scratches, second hand tube not crushed.

A few years back, a dial supplier shipped us 200 dials where the minute track printing was shifted 0.3mm off-center. Not visible at a glance — but every single watch would have looked slightly "off" to an attentive customer. IQC flagged it in 20 minutes. The dials went back. The production schedule didn't slip.

Straps and bracelets. Leather straps get checked for color fastness (rub a white cloth across the surface — any color transfer is a fail), stitch density (minimum 4 stitches/cm), and even thickness across the length. Metal bracelets get the bend-and-flex test: 50 cycles of folding each link to 90 degrees. A link that snaps at cycle 12 tells you the steel is too brittle. Rubber and silicone straps get stretched to 150% of their resting length — no tearing, no cracking, no permanent deformation.

IQC at JOEFOX samples to AQL 2.5 for major defects and AQL 1.0 for critical defects. A batch that fails IQC doesn't enter the production floor. Period.

Stage 2: IPQC — In-Process Quality Control

IPQC lives on the assembly line. It's not a separate department that shows up at the end — it's embedded in the workflow. The goal isn't to catch every defect. The goal is to catch the *process* that's creating defects before it produces a hundred bad units.

On our floor, IPQC works like this:

First article inspection (FAI). At the start of every shift, the first three fully assembled watches come off the line and go straight to the QC bench. Every dimension, every function, every cosmetic surface gets checked. If watch #1 has a second hand that's sitting 0.2mm too high (and therefore close to rubbing the crystal), we adjust the hand-pressing fixture before watch #4 gets assembled. FAI takes 15 minutes. Skipping it can cost an entire shift's output.
In-line checks at critical stations. Dial installation: is the dial seated flat against the movement ring? Is the date wheel centered in the window? Hand installation: are the hour and minute hands parallel? Does the second hand clear the minute hand and the crystal? Crown installation: does the stem engage smoothly through all positions? Does the crown screw down completely with the correct number of turns?
Cleanroom discipline. This sounds obvious, but it's the single biggest predictor of dial dust defects. Our assembly room runs at Class 10,000 (ISO 7) — meaning fewer than 10,000 particles per cubic meter at 0.5μm or larger. Workers wear full anti-static suits, hair nets, and finger cots. The room runs positive pressure so when someone opens the door, air flows out, not in. Every workstation has ionizing fans to neutralize static charge that attracts dust to dial surfaces.

I've toured factories where the "cleanroom" was a regular workshop with an air conditioner and a sign on the door. Their dial dust rate ran 8–12%. Ours runs under 0.5%. The difference isn't fancy equipment. It's discipline.

CNC process checks. For custom cases, IPQC measures wall thickness at the lug holes, crown tube bore diameter, and bezel seat flatness during the machining run — not after. A tool that's worn 0.01mm past tolerance will produce cases that look fine but won't seal properly.

Stage 3: FQC — Final Quality Control

FQC happens after assembly but before packaging. This is the 100% inspection stage — every single watch gets checked, not a sample.

Movement accuracy — 100% on the timing machine. Every quartz watch gets tested for current draw and initial accuracy. Every automatic gets the full three-position timing check: dial up, crown down, and vertical. The goal isn't COSC certification. It's consistency. A batch of automatics where 80% run +5 to +15 seconds/day and 20% run -25 seconds/day — that's a problem, even if the "average" looks acceptable.
Water resistance — 100% dry test, sample wet test. Every watch rated for water resistance goes through an air pressure leak test. The watch sits in a chamber; the machine pressurizes the air around it and measures whether the case deforms slightly (indicating air entering or escaping). A pass means the seals are intact. A fail means the watch comes apart for root cause analysis. Additionally, we wet-test a random sample in a water pressure chamber at the rated depth for 10 minutes — because air testing can miss seal failures that only appear under actual water pressure.

For the full deep dive on waterproof testing, read our guide on IP68 waterproof watch manufacturing in China.

Cosmetic inspection — 100% under 1000 Lux. Every watch goes under bright, diffused light at arm's length and then under a 5x loupe at critical areas: dial center, date window, hands, crystal edges, case flanks, crown, case back. Inspectors rotate the watch through three angles. What they're hunting: dust under the crystal, hand scratches, dial printing flaws, case polishing inconsistencies, bezel alignment, crown tube gap.

A typical FQC inspector at our factory examines roughly 200 watches per day. More than that and fatigue sets in — the defect catch rate drops measurably after hour six. So we staff accordingly.

Functional checks. Crown operation through all positions (winding, date set, time set — should feel smooth, not gritty). Date change at midnight (instant flip for Miyota quartz, slow engagement over 30–60 minutes for most automatics). Chronograph start/stop/reset (all hands must return exactly to zero). GMT hand independence. Bezel rotation (120 clicks for a dive bezel, each click crisp and distinct, no back-play).

One functional defect that buyers rarely think to check: the crown at the date-setting position. A crown stem cut 0.3mm too long will engage the date mechanism when the crown is pushed to the winding position — meaning the date changes every time the wearer manually winds the watch. We caught this on a 300-unit order last year during FQC. The stem supplier's batch had a slight dimensional drift. All 300 movements got fresh stems. The buyer never knew there was a problem.

Stage 4: PSI — Pre-Shipment Inspection

PSI is the final gate. After FQC, after packaging, after the watches are boxed and cartoned and stacked on a pallet — one last random-sample inspection before they leave the factory floor.

We pull samples according to the AQL table (more on this in the next section), open the packaging, and re-inspect every watch in the sample. This catches packaging damage, labeling errors, accessory omissions (missing manual? wrong warranty card?), and the rare defect that slipped through FQC.

At JOEFOX, PSI happens 48 hours before the scheduled ship date. That gives us a full business day to rework any issue without delaying the shipment. The PSI report — with photos, sample data, and pass/fail on every checkpoint — goes to the buyer before the pallet leaves our dock.

---

AQL Sampling for Watches: How Many Units Should You Inspect?

AQL — Acceptable Quality Limit — is the statistical backbone of every professional QC system. It answers one question: *out of an order of X units, how many do I need to inspect to be reasonably confident the entire batch is acceptable?*

The standard is ISO 2859-1. The logic: for a given batch size and AQL level, the standard tells you how many units to pull and how many defects you can accept before the batch fails.

A critical point most buyers miss: AQL is not an average. It's a limit. AQL 2.5 doesn't mean "the batch averages 2.5% defects and that's fine." It means "if the sample shows more than the allowable number of defects, the *entire batch* is suspect." The math assumes that if you find too many defects in the sample, the rest of the batch is statistically likely to be worse.

Watch-Specific AQL Reference Table

Different defect types demand different AQL levels. You don't tolerate a non-functioning movement the same way you tolerate a faint polishing mark on the case back.

| Defect Category | AQL Level | Sample Example (500 units) | What It Covers |

|---|---|---|---|

| Critical — safety hazard, movement dead on arrival, glass cracked | AQL 0.65 | Sample 80; reject if ≥2 found | Non-running movement, shattered crystal, crown stem snapped, leaking battery |

| Major — affects function or obvious at arm's length | AQL 2.5 | Sample 80; reject if ≥6 found | Visible hand scratches, dial misprint, bezel misalignment, water resistance failure, date doesn't change |

| Minor — visible on close inspection, doesn't affect function | AQL 4.0 | Sample 80; reject if ≥8 found | Faint polishing swirl, slight lume unevenness, strap keeper slightly loose, dust speck visible under 5x loupe only |

For a 500-unit order using AQL 2.5 for major defects, the math breaks down like this:

- Batch size: 500 → Sample size code letter: J → Sample size: 80 units

- AQL 2.5 → Acceptance number: 5, Rejection number: 6

- Translation: pull 80 watches. If you find 5 or fewer with major defects, the batch passes. If you find 6 or more, the entire batch fails.

For critical defects at AQL 0.65 on the same 500-unit order: acceptance number is 1, rejection is 2. Find two dead movements in your sample of 80? The batch fails. No exceptions.

At our factory, we run internal standards tighter than what the buyer specifies. If a client accepts AQL 2.5 for major defects, we sample internally to AQL 1.5. The buffer gives us room to catch and fix issues before they ever reach the client's inspector. It's not about being generous — it's about never having a buyer's inspection fail on our watch.

---

Component-Level Inspection: What to Check, Piece by Piece

You can't inspect a watch the way you'd inspect a T-shirt. A watch has 100+ individual components, each with its own failure modes. This is the checklist we use on the factory floor, broken down by component group.

Case and Case Back

The case is the first thing a customer touches. It sets the expectation for everything else.

Polishing and brushing. On a case that mixes polished and brushed surfaces — polished bezel, brushed lugs — the boundary between the two should be razor-sharp. No "bleeding" of the brushed texture into the polished zone. We check this under 1000 Lux light, tilting the case against the reflection. A good polish looks like a mirror. A rushed polish looks like a cloudy mirror.
Plating and coating adhesion. For PVD-coated cases (black, gold, rose gold), we do the tape test: apply 3M 600 tape to the coated surface, press firm, peel at a 90-degree angle. Any coating transfer to the tape is a fail. We also check for color consistency between cases within a batch — PVD process parameters drift over a production run, and case #50 can look subtly different from case #1 if the chamber temperature isn't controlled tightly.
Dimensional accuracy. Digital caliper on lug width, case diameter, thickness, lug-to-lug. Crown tube bore diameter gets checked with a go/no-go gauge. Thread engagement on the case back — it should thread smoothly for at least 3 full rotations before seating, with consistent torque throughout. A case back that binds after one turn has a cross-threading problem waiting to happen.
Edge finishing. Run a fingertip along every edge of the case. Any sharp spots — particularly at the lug tips and crown guards — fail. A watch that cuts the wearer or snags clothing is a return, guaranteed.

Dial and Hands

This is where most visible defects live. The dial is what the customer stares at every time they check the time.

Print quality. Under 5x magnification: logo crisp, no ink bleeding at the edges, text legible, minute track continuous with no breaks, date window cut clean with no rough edges. The tolerance for print registration error is effectively zero at this magnification — if the naked eye can see it's "off," it fails.
Lume application. Checked under UV light after a 30-second charge: even glow across all indices and hands, no dark spots (indicating thin application), no overflow onto the dial surface, no bubbles in the compound. Lume on hands should match the color and brightness of lume on indices — mismatched lume is a surprisingly common defect in OEM watches.
Index alignment. Applied indices checked for tilt, centering, and glue residue. Each index marker should sit perfectly perpendicular to the dial surface. A 1-degree tilt is invisible to the naked eye but catches light differently — the watch looks "cheap" without the buyer knowing why.
Hand alignment. Hour and minute hands must be parallel to the dial. Second hand must clear the minute hand, the crystal, and any applied indices. The three hands must not touch each other at any point in their rotation. We check hand clearance at 12:00, 3:00, 6:00, and 9:00 positions — these are the most likely interference points.
Dust. The #1 defect in watch manufacturing, period. A single visible dust particle under the crystal is a major defect. The cause is almost always one of three things: worker discipline (touching a non-cleanroom surface then handling a dial), static attraction (no ionizing fans at the workstation), or inadequate final cleaning (no compressed air blow-off before case closing). Fix the cleanroom, and you fix the dust problem.

Movement

The movement is what makes a watch a watch. The inspection here depends on whether you're running quartz or automatic.

Quartz movement checks:

- Current draw: Miyota 2035 should pull 0.55–0.90 μA with battery installed. Higher draw suggests contamination or a short.

- Accuracy: ±20 seconds/month is the standard for a basic Japanese quartz. Anything worse suggests a damaged quartz crystal or circuit issue.

- Battery voltage: fresh battery should read 1.55V minimum. Under-voltage causes erratic timekeeping.

- Hand clearance: the second hand on a quartz should advance in clean 1-second increments with no stutter or hesitation.

Automatic movement checks (three-position timing machine):

| Position | Target Rate | Target Amplitude | Beat Error |

|---|---|---|---|

| Dial Up | ±15 s/day | 270°–310° | ≤0.3 ms |

| Crown Down | ±20 s/day | 250°–290° | ≤0.4 ms |

| Vertical (12 up) | ±25 s/day | 240°–280° | ≤0.5 ms |

Amplitude below 240° in the horizontal position means the movement isn't fully wound — or the mainspring is tired. Beat error above 0.5ms means the balance isn't centered — the watch will run differently in different positions. Neither is acceptable on a new movement.

Date mechanism. Check that the date clicks over cleanly at midnight (instant or slow-creep, depending on the movement). Check that the quick-set date works from the crown's middle position. Check that the date can't be set between 9 PM and 3 AM on movements that warn against it (this is a movement-specific spec — most Seiko NH-series movements don't have a date-change danger zone, but many ETA-based movements do).
Chronograph (if applicable). Start, stop, reset. All chronograph hands must return exactly to zero. The central seconds hand must sweep smoothly, not stutter. The minute totalizer must advance cleanly at each 60-second mark. Test three full cycles to check for cumulative error.

Strap, Bracelet, and Clasp

The strap is the most abused part of any watch. It gets pulled, bent, soaked, and rubbed against desks. It needs to hold up.

Leather straps: Color fastness (white cloth rub test), stitch density and consistency (no skipped stitches, no loose thread ends), even edge painting (no cracks, no peeling), thickness uniformity. The quick-release spring bar must engage smoothly and hold firmly — a quick-release pin that pops out spontaneously is a warranty claim waiting to happen.
Metal bracelets: Link flexibility (every link should pivot freely, no binding), clasp snap tension (should close with an audible click, should not open with moderate pulling), end link fit to case (no gaps, no rattling), brushing direction consistency across all links.
Rubber/silicone straps: No flashing (mold seam lines), no air bubbles in the material, no sticky or tacky surface feel. Stretch test to 150% of resting length — should return to original shape with no deformation.
Tensile test: Each strap type gets a pull test. Leather should withstand 30N without tearing at the lug or buckle holes. Metal bracelets should hold to 50N at the clasp. A spring bar should not deform or release at under 20N of lateral force.

Want to see how this QC checklist runs on a real production line? Schedule a factory visit at JOEFOX — walk our Guangzhou floor, watch our QC team in action, and hold the inspection tools yourself. No sales pitch, just the production reality.

---

Third-Party Inspection vs. Factory Internal QC: A Watch Brand's Decision Framework

Every buyer eventually faces this question: *should I trust the factory's own QC, or should I pay an independent inspector?*

I'm the owner of a watch factory, and I'll tell you straight: I recommend third-party inspection for every new buyer relationship. Let me explain why — and when you can skip it.

*If you're launching a private label brand, see our private label watch manufacturing guide for QC considerations specific to branded production.*

The Honest Comparison

| Dimension | Third-Party Inspection | Factory Internal QC | What 35 Years Has Taught Me |

|---|---|---|---|

| Objectivity | No incentive to ship bad product | Under pressure to meet shipping deadlines | A good factory's internal QC is actually *stricter* than most third-party inspectors — we bear the cost of returns directly |

| Cost | $160–$300 per visit + travel | Already built into your unit price | On a 500-unit order at $25/unit, a $300 inspection is 2.4% of order value. A 5% return rate costs you far more |

| Watch-specific expertise | Generalist inspectors; may know electronics better than watches | Specialists who inspect watches 200+ days/year | A third-party inspector might not know what correct chronograph hand reset looks like. Ask about their watch-specific experience before booking |

| Speed | Requires scheduling; 2–3 business days lead time typical | QC is on the floor right now | For time-sensitive restocks, internal QC is faster. For first-production runs, the scheduling wait is worth it |

| Leverage | Neutral third party; report carries weight in disputes | Internal QC findings can be dismissed by the factory owner (if they're dishonest) | A factory that pushes back against third-party findings is telling you something about their quality culture |

When to Use Which

Start with third-party inspection when:

- You're working with a supplier for the first time

- The order value exceeds $5,000

- The watches include complications (chronograph, GMT, automatic)

- Your brand reputation depends on this batch (holiday launch, retail placement)

- The factory doesn't openly share their internal QC data

Factory internal QC is sufficient when:

- You've completed 3+ orders with the same factory with zero quality disputes

- The factory provides detailed QC reports (not just "passed inspection" — actual data)

- You've visited the factory and seen their QC process firsthand

- The order is a straightforward restock of a proven design

The hybrid approach that works best: third-party inspection on the first order. If the factory's internal QC matches the third-party findings (both catch the same issues at similar rates), reduce to random spot-checks via third party for subsequent orders — maybe one in every three or four shipments. If the factory's internal QC consistently misses defects that the third party catches, you have your answer about whether to continue the relationship.

At JOEFOX, we actively encourage new buyers to send a third-party inspector. It's not that we don't trust our own QC — our QC team catches 99%+ of defects before packaging. It's that the third-party report gives the buyer peace of mind that no factory sales pitch can provide. Transparency builds trust faster than any marketing claim.

---

The 7 Most Common Watch Defects — and How to Stop Them Before They Start

Over 35 years on a factory floor, certain defects show up again and again — across different factories, different brands, different price points. These seven cause the most returns, rework, and buyer frustration.

| # | Defect | Frequency | Root Cause | Prevention |

|---|---|---|---|---|

| 1 | Dial dust or fingerprints | Very High | Cleanroom discipline failure; static attraction; worker touching face/hair then dial | Class 10,000 cleanroom; ionizing fans at every station; full anti-static PPE; compressed air blow-off before case closing; 100% visual under 5x loupe |

| 2 | Hand scratches or misalignment | High | Hands stacked without protective separation; hand-pressing fixture out of calibration; rushed assembly | Foam-separated hand trays; daily FAI with hand clearance check at 4 positions; line speed limits that can't be overridden for rush orders |

| 3 | Water resistance failure | Medium-High | O-ring compression ratio wrong (target 25–35%); case back not torqued to spec; crown tube thread damage; crystal gasket pinched during installation | 100% air pressure leak test; wet-test sample per batch; torque-controlled case back closing tools; O-ring durometer incoming inspection |

| 4 | Movement accuracy out of spec | Medium | Movement not broken in or regulated before installation; position error ignored; quartz circuit damage from static discharge during handling | 24-hour run-in before timing check; three-position testing on automatics; ESD-safe workstations for quartz movement handling |

| 5 | PVD plating color inconsistency | Medium | Batch-to-batch PVD chamber parameter drift; different base material surface roughness between case batches | Batch color reference samples kept for comparison; spectrometer spot-check on coating thickness; same surface finish spec enforced at IQC |

| 6 | Strap or clasp failure | Medium-Low | Material not tested before production; spring bar temper wrong; clasp spring fatigue after limited cycles | Tensile testing on incoming materials; buckle cycle test (100 open/close cycles minimum); spring bar lateral force test at IQC |

| 7 | Crown or pusher binding | Medium-Low | Crown tube thread tolerance too tight or too loose; stem cut to wrong length; debris in pusher mechanism | Go/no-go thread gauge at IQC; crown torque test during FQC (smooth engagement through all positions); 100% functional check on pushers |

Let me tell you about #5. Three years ago, we received a batch of 300 rose gold PVD-coated cases from a plating subcontractor. The samples were perfect — rich, warm rose gold, exactly matching the Pantone reference. The production batch arrived in three cartons of 100 each. Carton 1 matched the sample. Carton 2 was slightly warmer — still rose gold, but leaning more toward copper. Carton 3 was noticeably lighter, almost a champagne tone.

The plating house had run the three cartons on three different days without resetting the PVD chamber parameters between runs. The titanium nitride layer — which creates the gold color — varied in thickness by roughly 0.3 microns across the three batches. Invisible to a thickness gauge at a glance. Visible to the human eye when two cases from different cartons were placed side by side.

All 300 cases went back. The plating house stripped and re-coated them on a single production day with locked-in parameters. We now require a reference sample from every PVD batch — not just the first one — and we compare batch #2 and #3 against batch #1 before accepting delivery.

That's the difference between a factory that does real QC and a factory that checks a box.

---

Pre-Shipment Inspection: The Final Gate Before Your Watches Leave China

PSI is your last chance to catch a problem before it lands on your customer's doorstep. After this point, the watches are in a shipping container. Any defect found after PSI is a defect your customer will find.

A proper PSI on the factory floor runs through six steps:

Step 1: Verify the order. Check that the carton count, unit count, model numbers, and packaging match the purchase order. Sounds basic. Gets missed more often than you'd think — especially on orders with multiple SKU variants (different dial colors, strap options).
Step 2: Pull the AQL sample. Use the ISO 2859-1 table to determine sample size based on batch quantity and your agreed AQL levels. Pull cartons from different pallets, different positions in the stack — not just the top layer. A factory that knows which cartons you'll sample can game the results.
Step 3: Piece-by-piece re-inspection. Open each sampled watch's packaging. Re-check every item on the FQC checklist: cosmetic, functional, accuracy, water resistance. This isn't redundant — packaging can introduce new defects (strap creasing from incorrect box placement, crystal smudges from final handling).
Step 4: Packaging and labeling check. Barcode scans correctly? FNSKU matches the product (critical for Amazon FBA)? Hang tag attached? Manual and warranty card included in the correct language? Box exterior clean and undamaged?
Step 5: Document everything. Photos of the sample layout, close-ups of any defects found, a tally sheet showing pass/fail counts by defect category, and a clear pass/fail verdict against the agreed AQL standard.
Step 6: Send the report before the shipment leaves. At JOEFOX, our PSI report goes to the buyer 48 hours before the scheduled pickup. That window gives you time to review, ask questions, and request rework if needed — without delaying the shipment.

A PSI that happens *after* the shipment has left the factory is not a PSI. It's a post-mortem.

---

10 Questions to Ask Your Watch Supplier Before Placing an Order

The best QC system in your supplier's factory is worthless if you don't know how to verify it. These questions cut through the sales talk and tell you whether a factory actually runs QC or just talks about it.

*Before you get to inspection, make sure you're sourcing from the right factory in the first place — our guide to sourcing watches from China covers supplier vetting, factory visits, and red flags to watch for.*

1. "Walk me through your QC stages — from IQC to PSI — for a typical watch order."

A real factory answers this without hesitation, in sequence, with specific checkpoints and tools mentioned. A trading company gives you vague generalities about "strict quality control."

2. "What AQL levels do you apply to critical, major, and minor defects?"

If they can't tell you the difference between AQL 0.65, 2.5, and 4.0 — and when each applies — they're not running statistical QC. They're eyeballing it.

3. "Do you have a QC team that's separate from your production team?"

QC that reports to the production manager has a conflict of interest. The production manager's bonus depends on hitting output targets. The QC manager should report to a different chain of command. Ask about the reporting structure.

4. "Can I see your water resistance testing equipment — and its calibration records?"

Ask for a video call showing the equipment in operation. Check the calibration sticker date. An uncalibrated pressure tester is worse than no tester — it gives false confidence.

5. "What's your 100% inspection policy versus sampling? Which checks does every watch go through?"

Every watch should get: timing machine check, air pressure leak test, cosmetic visual inspection, and crown/pusher function test. If they sample these instead of doing 100%, that's a red flag.

6. "When your QC finds a batch defect — what happens next?"

They should describe a process: isolate the batch → root cause analysis → corrective action → re-inspection → preventive measure update. "We fix it" is not a sufficient answer.

7. "Can you send me QC reports from three recent production runs — including ones where defects were found?"

Reports with zero defects ever found are suspicious. Every production line produces occasional defects. A factory that shares honest reports is a factory you can trust.

8. "Do you welcome third-party inspections? Which companies have your buyers used before?"

The speed and tone of their answer tells you everything. "Absolutely, we work with SGS, QIMA, Bureau Veritas regularly" — green flag. "Why would you need that? We have our own QC" — red flag.

9. "What traceability system do you use? If a customer reports a defect six months after delivery, can you trace it to a specific production batch?"

Batch numbers on movements, case lot tracking, and production date codes should be standard. If they can't trace a defect to its source, they can't prevent it from happening again.

10. "What's your typical defect rate — and your typical return rate from buyers?"

There's no single "right" number here — it depends on watch complexity and price point. But any factory that claims zero defects or zero returns is lying. A factory that gives you a specific number and explains what drives it is being honest.

---

The Bottom Line

QC isn't a cost. Bad QC is a cost. Good QC is the cheapest insurance your brand will ever buy.

A $300 inspection that catches a dial-hand scratching problem saves you from 500 angry customers, 500 return shipping labels, and an Amazon listing buried under one-star reviews. A factory that runs disciplined IQC saves you from discovering — three months after launch — that your cases were 304 steel sold as 316L.

The best factories don't fear inspection. They invite it. Because their QC system works, and they have the data to prove it.

At JOEFOX, we've been building watches in Guangzhou since 1990 — 35 years of learning what breaks, why it breaks, and how to stop it from breaking in the first place. Our QC team catches defects before they become your problem. And for new buyers, we actively recommend sending a third-party inspector on your first order. Walk our floor. Watch our process. Read our reports.

Transparency is the only quality guarantee that actually means something.

Start your watch project with JOEFOX → — or schedule a factory visit and see our QC floor for yourself. Factory-direct from Guangzhou since 1990.

---

*Looking for more? Read our complete guide to IP68 waterproof watch manufacturing — or compare OEM vs ODM watch manufacturing to understand which model fits your brand.*

---