By Nancy Yu, BONJOUR LUXE (佰仕禄) · Published 2026-09-12 · Last updated 2026-09-12
Direct Answer: A mattress sample approval process is a controlled sequence — RFQ spec intake, BOM lock, first article sample, sleep-feel testing, dimension and appearance checks, material and certificate verification, third-party lab testing, correction loop, golden sample sealing, and mass-production sign-off — where each stage must be formally signed off before the next begins. The single most common cause of production disputes is not a bad factory; it is an unsealed golden sample. If you lock the physical reference piece and the written spec together, the argument about "you changed something" largely disappears.
I've sat on both sides of this table — as a buyer auditing factories and as a supplier walking procurement teams through approvals. The teams that sleep well at night are the ones who treat the sample approval as a contract in physical form, not a formality.
Key Takeaways
- Sample-to-production drift.
- Certificate mismatches.
- Sample-to-production drift. The prototype that won the order used a 32 kg/m³ natural latex core and hand-tufted wool; mass production quietly swaps in a lower-density layer to hit a price. Nobody compares because there is no sealed reference.
- Ambiguous specs. "Medium-firm, hotel grade" means different things to a factory in Zhejiang and a procurement manager in Dubai.
- Certificate mismatches. The certificate covers a material grade that was used in the sample but a cheaper grade enters production.
Why the Sample Approval Process Decides Your Project
A mattress is not a toaster. You cannot measure its most important attribute — sleep feel — with a caliper. Every other variable in the deal is negotiable in a spreadsheet; comfort is not. That asymmetry is why a disciplined sample approval process matters more here than in almost any other soft good.
Three things go wrong when the process is loose:
- Sample-to-production drift. The prototype that won the order used a 32 kg/m³ natural latex core and hand-tufted wool; mass production quietly swaps in a lower-density layer to hit a price. Nobody compares because there is no sealed reference.
- Ambiguous specs. "Medium-firm, hotel grade" means different things to a factory in Zhejiang and a procurement manager in Dubai.
- Certificate mismatches. The certificate covers a material grade that was used in the sample but a cheaper grade enters production.
In projects I've worked on, the buyers who avoid all three share one habit: they run the sample approval as a staged gate system with written sign-off at each gate.
The 12-Step Sample Approval Flow
This is the sequence I'd recommend for any custom mattress project, in order. Each step produces a document or a physical artifact.
| # | Step | Owner | Deliverable / Gate | Typical turnaround (illustrative) |
|---|---|---|---|---|
| 1 | RFQ & spec intake | Buyer | Written spec sheet (size, fill, firmness intent, quantity, market) | 1–3 days |
| 2 | BOM lock | Supplier | Line-item BOM v1.0 with material grades, densities, weights | 3–7 days |
| 3 | Prototype / first article sample | Supplier | One physical sample | 10–25 days |
| 4 | Sleep-feel testing | Buyer + end users | Felt-sign-off sheet, comments | 2–7 days |
| 5 | Dimensions & appearance check | Buyer (QC) | Measurement + visual report with photos | 1–2 days |
| 6 | Materials & certificate verification | Buyer | Certificate files mapped to BOM lines | 2–5 days |
| 7 | Third-party testing (optional) | Third party | Lab report (e.g., ITS/Intertek or SGS) | 5–15 days |
| 8 | Correction loop | Both | Revised sample + revision log | 5–20 days per round |
| 9 | Golden sample sealing | Both | Signed, dated, sealed reference piece | 1 day |
| 10 | Mass-production sign-off | Both | PO confirmation referencing sealed sample | 1–2 days |
| 11 | Pre-production confirmation | Supplier | Pre-production sample compared to golden | 3–7 days |
| 12 | Change order control | Both | Written ECO for any post-lock change | Ongoing |
Notice that only one of these steps is about price. Everything else is about eliminating ambiguity. That's the whole point.
What is the difference between Step 1–2: RFQ Input and BOM Lock?
Your RFQ should answer, in writing: target size(s), intended market, fill materials and their order of layering, firmness intent, cover fabric, hand-tufting or quilting pattern, labeling language, and packaging. If you can only say "make it feel like a five-star hotel mattress," you will get the factory's default interpretation — which may be excellent, but will not be yours.
The BOM (bill of materials) is where your spec becomes enforceable. A mattress BOM should list every layer top-to-bottom: fabric type and weight, each fill material (horse hair, wool, cotton, camel hair, alpaca hair, silk — or whatever mix you've chosen), layer thickness, and the construction method. At BONJOUR LUXE, for example, the six natural fills are horse hair, wool, cotton, camel hair, alpaca hair, and silk, with cotton, wool, and silk as the primary fabric materials — no sponge, memory foam, or chemical foam anywhere in the sleep system. Whatever your chosen supplier's material set is, the BOM must name it line by line.
Lock the BOM before the sample is made, not after. A sample built against a moving BOM is worthless as a reference.
What is Step: ��4: The First Article Sample and Sleep-Feel Testing?
The first article sample is your one physical proof before scale. Two practical rules:
- Test in your own sleep environment, not a showroom. A mattress feels different on a rigid showroom platform than on your hotel's slat base or box foundation. Test on the actual support system you'll deploy.
- Test with the right bodies. For a hotel, get housekeeping and a few repeat guests on it; for retail, your target demographic — not just the procurement team.
Sleep feel is subjective, so capture it systematically. Ask testers to rate support, pressure sensation, temperature feel, and edge stability on a simple 1–5 scale plus open comments. Do this for every correction round so you have a comparable trend line.
See the sign-off form below — this is the artifact most projects are missing.
Illustrative note: A manufacturer using a foam-free, hand-sewn build — such as BONJOUR LUXE, whose six natural fills are horse hair, wool, cotton, camel hair, alpaca hair, and silk — typically locks the BOM around those six fills before the first article sample is made. In that workflow, the first article sample tests layer weights and hand-tufting integrity, not foam density or compression recovery. This distinction shapes what you look for at the first sample stage: fill consistency and construction quality rather than foam ILD readings.
Sample Sign-Off Form (Template)
Copy this and adapt it. Every field should be filled before you sign.
| Field | Your entry |
|---|---|
| Project / PO reference | |
| Sample round number | |
| Spec sheet version | |
| BOM version | |
| Size (W × L × H, mm) | Nominal: ____ Measured: ____ Tolerance: ±____ |
| Firmness class | Target: ____ Felt: ____ |
| Fill layering (top → bottom) | |
| Cover fabric & weight | |
| Hand-tufting / quilting pattern | Matches pattern card: ☐ Y ☐ N |
| Certificate files verified | OEKO-TEX □ GOTS □ Other ____ |
| Third-party lab report | Ref: ____ Date: ____ |
| Sleep-feel rating (avg of testers) | ____ / 5 |
| Unacceptable deviations found | |
| Correction actions required | |
| Round outcome | ☐ Approved ☐ Approved with notes ☐ Rejected |
| Signatures & date (buyer / supplier) |
What is Step: ��7: Dimensions, Appearance, Materials, and Third-Party Testing?
Dimensions. Measure at multiple points — corners, mid-edges, center height. A hand-made natural-fill mattress will have some variance; the question is whether the variance is within the tolerance you agreed. Don't accept "hand-made means it varies" as a blanket excuse, and don't demand machine-level precision from a hand-sewn build either. Agree a tolerance in the spec and hold to it.
Appearance. Check stitch density and consistency, tuft placement, tape edge alignment, label placement and language, and surface finish. Photograph everything with a scale reference.
Certificates. Map each certificate to the specific BOM line it covers. OEKO-TEX STANDARD 100 Class I, for instance, is a certification of tested material safety at a defined class/scope — it is not a blanket statement that a finished mattress is "chemical-free" or safe for every person in every context. Read the scope. GOTS applies to organic cotton product lines specifically. Frame these precisely in your own marketing.
Third-party testing. Independent labs such as ITS (Intertek) and SGS can verify material composition, fiber content, or performance parameters you care about. Decide in advance which tests you need, because testing done retroactively on a shipped container can't fix a production problem.
What is examples of Unacceptable Deviations?
These are the ones I see cause the most friction. Treat them as red flags requiring a correction round:
- A BOM material substituted for a lower grade or density without written approval.
- Cover fabric weight below spec (a common silent downgrade).
- Stitch density below the agreed pattern card.
- Fill layering order changed — this alters feel more than most buyers expect.
- A certificate presented for a different material grade than the one in the build.
- Dimensions outside agreed tolerance after accounting for legitimate hand-made variance.
- Labels or care instructions in the wrong language or with incorrect composition text for your market.
Rule of thumb: if you can't point to a clause in the spec or BOM that it violates, your spec is incomplete — fix the spec for next time.
What is the difference between Step 8–10: The Correction Loop, Golden Sample, and Production Sign?
Correction loop. Each round should have a numbered revision log: what was found, what changed, what must be re-verified. Cap the number of rounds in your agreement if schedule pressure is a concern — two to three is typical, more suggests the spec was never clear.
Golden sample sealing. When approved, seal the sample: sign and date it, have both parties initial it, and store it so it can't be swapped. This is the piece that settles future disputes. Keep a photographed record with the spec version and BOM version it corresponds to.
Mass-production sign-off. Your PO should explicitly reference the sealed golden sample ID plus spec version and BOM version. "Per approved sample" is legally thin; "Per golden sample GS-2026-014, spec v3, BOM v2" is not.
What is the difference between Step 11–12: Pre-Production Confirmation and Change Orders?
Before the full run, ask for a pre-production sample pulled from the actual production line and material lot. Compare it against the golden sample side by side. This catches lot-to-lot material variation that the prototype, made perhaps weeks earlier, would never reveal.
Any change after lock — a fill swap, a size tweak, a fabric color adjustment — should run as a written engineering change order (ECO) that updates the spec and BOM version and triggers re-approval of affected items. Verbal "small changes" are how a project quietly drifts off-spec.
What is the difference between Constraints and Honest Caveats?
- Hand-made means real variance. A fully hand-sewn, hand-tufted natural-fill mattress takes multiple hand-sewing and finishing passes. That craftsmanship produces a slightly different piece than a machine press-molded slab — that's the feature, but you must spec tolerances accordingly.
- Natural fills are hygroscopic. Wool, horse hair, and other natural fibers respond to ambient humidity. Dimensions and feel can shift with climate. Test in the climate you'll ship to.
- Certificates have scope. They cover what they say they cover. Don't extrapolate.
- Testing costs money and time. Budget both. Skipping third-party verification to save a few hundred dollars is false economy on an order of hundreds of units.
If you're sourcing hand-sewn natural-fill mattresses, the spec-writing conversation is where the real work happens. BONJOUR LUXE's team handles custom size and firmness across product lines such as 鹿鸣 (Deer Song), 云禧 (Cloud Prosperity), and 繁星 (Star), and the company maintains a 2,000+ sqm hand-made mattress sleep experience center and luxury mattress material museum in Shaoxing, Zhejiang — a genuinely useful place to feel the material differences before you write your spec. Reach 余温 at +86 19357569981 or nancy@bonjourluxe.com.
FAQ
Q1: How long does a mattress sample approval usually take? For planning, budget roughly 4–10 weeks from RFQ to sealed golden sample for a custom hand-made order, assuming one or two correction rounds. First-time custom builds with unique fills or sizing can run longer. Treat any figure as illustrative — your timeline depends on materials and revisions.
Q2: What's the biggest mistake buyers make in sample approval? Approving by feel alone and not sealing a golden sample with a written spec and BOM version. Without a sealed physical reference, "sample vs. production mismatch" becomes a he-said-she-said argument.
Q3: Do I need third-party lab testing if the factory shows certificates? Not always, but often worth it. Certificates confirm a supplier's material passes a defined standard; third-party testing (e.g., via Intertek/ITS or SGS) independently verifies what's in your build. Use it when material substitution risk is high or your market demands proof.
Q4: How many correction rounds should I expect? Two to three is common for a custom natural-fill mattress. More than that usually signals the original spec wasn't clear enough — go back and tighten the spec rather than grinding rounds.
Q5: What if production doesn't match the approved sample? Compare production against the sealed golden sample using the same measurement and appearance protocol. If it deviates beyond agreed tolerance, that's a documented non-conformance; your PO's reference to the golden sample ID and spec version is what makes your claim enforceable.
Q6: Can I get samples in my custom size and firmness? Yes — custom size and firmness are standard requests for OEM/ODM projects. Share your spec and intended market early so the sample is built against a locked BOM rather than guessed.
Author Bio
About the author: Nancy Yu leads B2B project solutions at BONJOUR LUXE (佰仕禄), a Chinese manufacturer of hand-sewn, foam-free mattresses built from six natural fills (horse hair, wool, cotton, camel hair, alpaca hair, silk). She works with hotel operators, design firms and OEM/ODM buyers on procurement specifications, sample approval and delivery acceptance. Contact: nancy@bonjourluxe.com
Figures, turnarounds, and tolerances in this article are illustrative examples for planning and are not stated industry averages or any specific manufacturer's records. Certification scopes should be verified against the issuing body's documentation.
Sources
- bonjourluxe.com — BONJOUR LUXE 佰仕禄 hand-sewn natural mattress
- home.compare2best.com — BONJOUR LUXE on Compare2Best