Net-content inspection needs a suitable measurement method, known tare behavior, product density at the relevant temperature and the correct market-specific decision rule. This page separates internal process controls from legal lot assessment and provides a fictional calculation that can be checked; it is not a verified factory report.
Why Accurate Fill Matters for Perfume Brands
Beyond the obvious customer satisfaction angle, fill accuracy touches several business-critical areas:
- Legal criteria: Identify the actual destination’s quantity rules and sampling/decision method before assessing a lot.
- Process control: Track actual fill variation against approved internal criteria; do not use an invented ±5 mL market pattern.
- Material reconciliation: Measure actual overfill, loss and cost rather than assuming a fixed giveaway erases profit.
- Traceability: Link fill data to bulk input and losses with documented units, density assumptions and measurement limitations.
Common Inspection Methods
Choose a validated method for the actual product and package. Mass-based conversion can support a volume-labeled product, but visual level, gross weight and net volume are different quantities. Confirm applicable legal methods and sampling separately.
1. Gravimetric (Weight) Inspection
A gravimetric approach derives net mass from the applicable gross/tare method, then divides by the relevant measured density to estimate volume. Tare variability, density variation, temperature and measurement uncertainty affect the result. It is not inherently the most reliable method for every product or bottle.
- Equipment: Confirm the available scale/checkweigher and its validated range, resolution, calibration and suitability for the selected method.
- Measurement capability: Resolution alone is not accuracy. Verify repeatability and uncertainty against the actual decision rule; do not assume a standard perfume density or universally sufficient 0.01 g display.
- Temperature: Measure density and assess volume at the applicable reference conditions, using validated correction where appropriate.
2. Level / Fill Height Inspection
A fill-height observation can be a process indicator when validated against actual volume for the selected bottle. Geometry, meniscus and bubbles can affect it; do not equate a common visual line with verified net contents.
- Manual level screen: A drawing/fixture reference can flag differences, but needs validation before being used as quantity evidence.
- Automated level screen: Verify actual sensor suitability for the bottle; no installed camera/laser system is claimed.
3. Automated Checkweighing
An inline checkweigher, if actually available and validated, may monitor gross mass at production speed. Its resolution, uncertainty, rejection function and tare strategy need verification. Subtracting one stored average tare can misrepresent individual contents when glass/closure weights vary; no 0.1 g capability or full-line inspection is verified here.
4. Legal Lot Assessment and Internal Sampling
Separate internal quality sampling from the legally applicable net-content test. A generic AQL 2.5 plan does not establish legal fill compliance. Identify the lot, selection method, sample size, individual and average criteria and decision rule required for the actual jurisdiction and package.
For a traceable U.S. method example, NIST Handbook 133 (2026), Table 2-1 uses Category A samples of 12 for inspection lots of 12–250 packages, 24 for 251–3,200 and 48 for larger lots. Random selection, the permitted number of errors exceeding the MAV, the average test and the applicable tare procedure are all part of the method. Use Chapter 1, Chapter 2 and Chapter 3 for the full procedure and its application to the quantity and package. An internal AQL inspection or a different receiving sample cannot be described as passing this legal net-content test merely because its sample size looks similar.
| Method | Useful role | Important limitation |
|---|---|---|
| Gravimetric | Net mass and justified conversion to volume | Requires appropriate tare, density/temperature and method capability |
| Fill-height screen | Process observation for validated bottle geometry | Not direct volume proof; opacity/meniscus/bubbles may interfere |
| Inline checkweigher if available | Controlled production monitoring | Gross mass and average tare can hide individual content errors |
| Legal lot test | Destination-specific compliance assessment | Use applicable sampling and individual/average criteria, not generic AQL |
What Tolerances Should You Expect?
Legal quantity allowances are not a symmetric filling specification. Keep the declared nominal quantity, the legal rules for the destination, and the manufacturer’s internal fill target and warning limits separate. Set process controls from actual fill variation, tare, density, temperature and measurement capability; a generic Cpk threshold or internal AQL plan does not prove legal net-content compliance.
EU, Great Britain and Northern Ireland: define the applicable system
The EU’s voluntary ℮-mark system covers qualifying prepackages from 5 mL to 10 L (or 5 g to 10 kg). Directive 76/211/EEC, as amended, provides the harmonised framework; national requirements also matter, particularly outside that scheme. Great Britain uses the Packaged Goods Regulations 2006, while Northern Ireland uses the Packaged Goods Regulations (Northern Ireland) 2011. Their general quantity range is 5 mL–25 L (or 5 g–25 kg), subject to the regulations’ scope and exceptions. Do not treat a 2 mL sample or a different sales system as covered by the examples below.
Under these average-quantity rules, the batch average must be at least the nominal quantity; the number of packages short by more than the tolerable negative error (TNE) must meet the applicable reference-test criteria; and a package short by more than twice the TNE cannot qualify. For the EU scheme, such a package must not bear the ℮-mark. The NI three packers’ rules and GB guidance distinguish these simultaneous requirements. TNE is not an amount that may be deducted from every bottle.
| Nominal quantity, Qn | TNE | Qn − TNE | Qn − 2 × TNE |
|---|---|---|---|
| 10.0 | 0.9 | 9.1 | 8.2 |
| 30.0 | 2.7 | 27.3 | 24.6 |
| 50.0 | 4.5 | 45.5 | 41.0 |
| 100.0 | 4.5 | 95.5 | 91.0 |
These values follow the 9% TNE band from 5–50 mL and the 4.5 mL band from 50–100 mL. Where a percentage produces a fraction of 0.1 mL, round the TNE upward to the next 0.1 mL. See GB Schedule 3, NI Schedule 3 and the Commission-listed WELMEC 6.4 (2015), section 4.1 and Table 1, which explains the EU directive. WELMEC guidance is not itself legislation. A result below Qn − TNE counts toward the reference test’s short-package limit; the lower boundary is not a recommended fill target. These are not U.S. MAV values.
Reference sampling is separate from production control
The EU Annex II reference method checks individual contents and the sample mean; its non-destructive and destructive routes have different conditions. The Commission-listed WELMEC 6.5, Annex H explains reference testing and random selection. Do not replace the applicable method with a generic AQL 2.5 inspection.
For a specific GB/NI example, the single-sample, non-destructive route in GB Schedule 2 and NI Schedule 2 uses the following separate individual-content and mean checks. Here, a defective package contains less than Qn − TNE; s is the sample estimate of the standard deviation in the same units as Qn.
| Packages in batch | Individual-content sample | Accept / reject on defective count | Mean-check sample and acceptance criterion |
|---|---|---|---|
| 100–500 | 50 | Accept ≤3; reject ≥4 | 30; mean ≥ Qn − 0.503s |
| 501–3,200 | 80 | Accept ≤5; reject ≥6 | 50; mean ≥ Qn − 0.379s |
| 3,201 and above | 125 | Accept ≤7; reject ≥8 | 50; mean ≥ Qn − 0.379s |
Both checks must pass, and the separate twice-TNE prohibition still applies. These sample-mean thresholds account for statistical sampling; they do not authorise a production average below Qn. Follow the schedules’ batch definition and random-selection rules: the batch is capped at 10,000 outside end-of-line checks; end-of-line checks use maximum hourly output. For a batch below 100, this non-destructive route checks every package; apply the separate small-batch acceptance rule in Schedule 2 paragraph 3.6 rather than treating Qn as a per-bottle minimum. Double-sampling and destructive routes have their own criteria and must not be mixed with this table.
For the GB/NI reference test, volume is measured at, or corrected to, 20°C; the measurement error must not exceed one fifth of the relevant TNE. These are method requirements, not verified instrument performance at LAEYO. Record the actual destination, legal system, quantity declaration, method, batch definition and results before making a lot decision. No actual shipment is assessed by these examples.
Challenges You’re Likely to Encounter
Foam and Bubble Traps
Record any bubbles or foam and their effect on the chosen measurement. Have engineering investigate the actual product/process and approve any filling or conditioning change. Do not assume surfactants or prescribe a nozzle change as the cause/remedy without evidence.
Neck Fill and Headspace
Assess headspace against actual bottle capacity, thermal behavior, closure and validated use/transport conditions. No universal 10–15% headspace rule is established. The label’s quantity basis remains relevant even if mass is used for internal filling control.
Temperature Influence During Inspection
Record the product temperature and density basis. A volume-labeled product cannot avoid reference-temperature requirements simply by adopting a mass limit. Use an appropriate validated conversion/correction and account for actual measurement uncertainty.
Opaque and Decorated Bottles
For opaque bottles, select a suitable validated method such as gravimetric testing with the appropriate tare/density approach, or a controlled destructive volumetric method performed by trained personnel. Avoid relying on an unverified average glass weight; opacity does not mandate one specific factory machine.
Fictional tare-distribution example
Assume only for this arithmetic example: nominal 50.00 mL and density 0.8500 g/mL at 20.0°C, giving 42.50 g nominal liquid mass. Three complete empty-package tares are 98.00, 100.00 and 102.00 g, averaging 100.00 g. These are invented values, not measured LAEYO data or a recommended sampling plan.
| Unit | Actual tare, g | Gross, g | Actual net / volume | Using 100.00 g average tare |
|---|---|---|---|---|
| A | 98.00 | 140.50 | 42.50 g / 50.00 mL | 40.50 g / 47.65 mL |
| B | 100.00 | 142.50 | 42.50 g / 50.00 mL | 42.50 g / 50.00 mL |
| C | 102.00 | 144.50 | 42.50 g / 50.00 mL | 44.50 g / 52.35 mL |
| D, separate underfill illustration | 102.00 | 142.80 | 40.80 g / 48.00 mL | 42.80 g / 50.35 mL |
Volume = (gross mass − applicable tare mass) ÷ density. Unit D shows how heavier glass can conceal a lower actual fill when an unsuitable average tare is used. This does not condemn every approved average-tare method: legal procedures specify their own tare and variability controls. No row is a legal pass/fail determination.
Actual measurement and decision record
| Record | Project field |
|---|---|
| Identity/sampling | Order, lot, unit IDs, quantity declaration, legal method/edition, sample selection ___ |
| Tare distribution | Component lots; individual/approved tare method; sample size, range and variability ___ |
| Measured density | Actual result ___ g/mL; method ___; product temperature ___ °C; density variation/uncertainty ___ |
| Calibration | Scale/density/temperature instrument IDs; calibration dates and references; resolution/uncertainty ___ |
| Calculation | Gross ___ g; tare ___ g; net ___ g; volume ___ mL; correction basis ___ |
| Separate decisions | Internal process limits ___; legal criteria ___; result/disposition ___; qualified reviewer/date ___ |
EU/GB/NI sources checked on 14 October 2026: European Commission ℮-mark guidance and its listed WELMEC guides; GB SI 2006/659, Schedules 2–3; NI SR 2011/331, regulation 4 and Schedules 2–3; and current NI cosmetics guidance. The earlier U.S. source review of NIST Handbook 133, 2026 edition is retained. These are source checks for a general guide, not a legal approval or an assessment of any actual batch. The applicable country, package, quantity declaration and measurement records still need to be identified for a shipment decision.
FAQ
What’s the best way to verify fill accuracy when I receive my shipment?
Agree a representative receiving plan with a qualified reviewer. Use suitable calibrated equipment and an approved tare/density or volumetric method. Do not adopt an arbitrary 20-bottle sample, assume identical empty bottles or claim 0.01 g resolution is always sufficient.
Should I use volume or weight as my primary specification?
Keep the declared quantity and legal assessment basis distinct from a mass-based process target. For a volume-controlled example, target net mass equals volume multiplied by density at the relevant reference temperature. Actual density and uncertainty must be measured; a mass target alone does not establish legal compliance.
How do cosmetic regulations affect perfume fill inspection?
Quantity labeling and legal metrology overlap with cosmetic requirements. The 2026 NIST Handbook 133 provides U.S. package-test procedures within federal, state and local requirements. For the EU ℮-mark scheme, use the relevant harmonised rules and national application; GB and NI have separate packaged-goods regulations. The scoped TNE and GB/NI reference-test examples above are not universal perfume tolerances and do not replace cosmetic labeling requirements or actual batch measurements.
Can a manufacturer fudge fill data?
Request traceable raw data, sample selection, instrument/calibration records and approved decision criteria. Assess unexplained gaps or inconsistent records rather than accusing a supplier of falsification without evidence. An agreed independent inspection can support verification where appropriate.
What documentation should I request regarding fill inspection?
Request a traceable inspection report with gross/tare/net measurements, density and temperature, unit/lot IDs, sampling basis, instrument/calibration/uncertainty records, calculations and signed disposition. Where a COA is used, it may reference the underlying report rather than contain every field itself.
To discuss fill inspection, agree the actual quantity specification, destination rules, measurement approach and reporting scope with the project team. No factory accuracy or every-bottle guarantee is established by this guide.