Perfume Rework Process | Quality Defects, Corrective Action, Reinspection

Perfume can be reworked when the defect is known and the correction can be measured. A batch may be filtered again, adjusted by calculation, refilled, relabeled, or fitted with new pumps. If the formula, contamination source, or batch history cannot be confirmed, the batch should stay on hold or be rejected instead of being adjusted again and again.

Stop the Batch

Once a defect is found, stop the affected material from moving.

• bulk perfume in the tank;
• filled bottles;
• packed cartons;
• unused pumps, bottles, labels, or cartons from the same lot;
• retained samples;
• stock already moved to another warehouse.

Record the batch number, quantity, production stage, and the time the problem first appeared.

Say 18,000 bottles have already been packed. One filling nozzle started running low at 14:20 and was corrected at 15:05. If the line was producing about 30 bottles per minute, the first group to check is about 1,350 bottles:

45 minutes × 30 bottles/minute = 1,350 bottles

The remaining 16,650 bottles should not automatically be treated as defective if the production records show they were filled outside that time window.

FDA's cosmetic GMP inspection checklist includes records for processing, filling, sampling, adjusting, and reworking cosmetic products.^[1]

Tanks, pallets, and component lots also need a clear hold status. Otherwise, good and suspect material can easily get mixed during a busy production shift.

If some units have already shipped, check where they went. Further shipment may need to stop while the defect is reviewed.

Confirm the Defect

A batch needs to fail against something specific.

"Looks different" or "smells weak" is too vague.

• approved formula and formula version;
• approved fragrance concentrate;
• approved reference sample;
• color limits;
• fill specification;
• pump and bottle specification;
• approved artwork;
• finished-product specification;
• stability requirements.

For odor, compare the suspect batch with the approved reference under the same conditions. Use the same blotter, a similar spray amount, and the same evaluation times. A fresh spray should not be compared with a sample that has already been sitting for 30 minutes.

For appearance, use the same light and container type.

For fill quantity, measure the actual amount. Do not rely on the liquid line alone. Glass thickness and internal bottle shape can make equal fills look different.

XUELEI's perfume QC specifications define the test method and pass limit before production starts.

• Safety: harmful contamination, sharp glass, or another condition that may injure the user.
• Regulatory: wrong ingredient information, noncompliant materials, or an incorrect market label.
• Functional: leakage, poor spray, blocked pumps, or incorrect fill.
• Cosmetic: minor scratches, print marks, or carton defects that do not affect use or safety.

A cracked bottle neck is a good example of overlap. It may leak, and it may also create a safety problem.

Find the Cause

Start with the batch record.

Compare target quantities with the quantities actually used. Then check:

• raw-material lot numbers;
• formula version;
• weighing and dispensing records;
• tank identity;
• transfer losses;
• mixing time;
• processing temperature;
• filtration conditions;
• storage history.

Do not adjust the formula if the batch composition cannot be confirmed.

Before a major rework starts, write down the material and quantity to be changed, the process conditions, and the tests required afterward.

Keep a sample of the failed batch before rework.

For a large tank, one sample from the easiest valve may not be enough after a formula adjustment. Samples can be taken from the top, middle, and bottom after mixing. The actual sampling points should follow the factory procedure for that tank.

Odor Problems

Is the whole perfume weaker? Is the opening different? Is one note too strong? Does the opening match while the dry-down does not?

The dry-down is the smell left after the lighter opening materials have evaporated.

A weak overall scent can point to dilution or a fragrance-concentration error. An unfamiliar smell may come from oxidation, the wrong raw material, contamination, or residue in production equipment.

XUELEI's perfume sample comparison checks samples under the same application conditions and at several points after spraying.

Check the weighing record before adding more fragrance.

Take a 100 kg batch containing 18 kg of fragrance concentrate. The target is 20% by weight.

Adding only 2 kg is not enough because the total batch weight also increases.

(18 + x) ÷ (100 + x) = 20%

x = 2.5 kg

• new batch weight: 102.5 kg;
• fragrance concentrate: 20.5 kg.

20.5 ÷ 102.5 = 20%

This is a w/w calculation using the same fragrance concentrate as the original formula.

Use this calculation only after confirming that fragrance was under-dosed. If the original additions cannot be confirmed, adding more fragrance makes the final formula harder to verify.

Repeated opening, pumping, filtration, or transfer can also lose ethanol and volatile top-note materials.

IFRA Limits

A higher fragrance concentration may change the IFRA limit that applies.

IFRA Standards can prohibit a fragrance material, restrict how much may be used, or set conditions such as purity requirements.^[2]

If the IFRA document covers the perfume at 18%, check it again before increasing the fragrance level.

Check the new concentration against the IFRA information and the requirements of the market where the perfume will be sold.

An IFRA Certificate of Conformity does not replace the finished-product safety assessment. IFRA also states that its Standards do not replace national or local law.^[3]

Haze and Sediment

Cloudiness often comes from solubility, excess water, temperature, natural waxy materials, filtration, contamination, or contact with the package.

Take a sample before filtering anything.

If the perfume is clear at room temperature but cloudy when cold, check solubility first. Hard particles at the bottom need a different investigation.

• Does the haze change with temperature?
• Does the sediment dissolve in ethanol?
• Is the material already present in the bulk tank?
• Does it appear only after filling?
• Do several bottles contain the same material?

If 8 of 20 affected bottles came from one filling head while the rest were spread across five other heads, inspect that filling head first.

If the bulk reference stays clear but filled bottles become cloudy, check the bottle, pump, gasket, and filling process. XUELEI's perfume separation and solubility checks compare bulk perfume and filled packs separately for this reason.

If the batch is filtered again, record the temperature, holding time, filter type, filter grade, quantity before filtration, and quantity recovered.

A perfume that looks clear after filtration but turns cloudy again after cold storage has not been fixed.

Color Changes

Color changes usually point back to the raw materials, oxidation, heat, light, color dosing, or contamination.

• approved color range;
• fragrance and colorant lots;
• actual colorant quantity;
• alcohol and water quality;
• processing temperature;
• storage and light exposure.

If records show that too little approved colorant was added, a calculated correction may be possible.

Do not use more colorant to cover oxidation.

Check the smell at the same time. Darkening plus a stale or unusual odor points more toward chemical aging than a simple color-dosing error.

Wrong Fragrance Strength

Take a 100 kg batch containing 22 kg of fragrance concentrate. The target is 20% by weight.

If the approved dilution material contains no fragrance concentrate:

22 ÷ (100 + x) = 20%

x = 10 kg

The corrected batch weighs 110 kg:

22 ÷ 110 = 20%

Do not mix kilograms, liters, % w/w, and % v/v in the same calculation. Density may be needed when converting between weight and volume.

After adjustment, mix the batch before taking a sample. A sample taken from the top of a large tank immediately after adding material may not represent the whole batch.

If the batch grows from 100 kg to 110 kg, record the extra 10 kg and the additional finished units it produces.

Water and Alcohol

Adding water changes the alcohol level and fragrance solubility.

• clarity;
• alcohol level;
• spray feel;
• stability;
• microbial risk.

Use the same solvent system as the approved formula. A different ethanol grade, denaturant system, or water quality can change the product.

Ethanol can also evaporate while a tank is open or during repeated transfers. That can raise the fragrance concentration without anyone adding more fragrance.

Foreign Matter

Do not classify an unknown particle as a cosmetic defect until its source is known.

• bulk perfume;
• empty bottles;
• filters;
• transfer hoses;
• filling nozzles;
• pumps and gaskets;
• caps;
• the filling area.

If most particles come from one filling head or one packaging lot, inspect that source first.

Keep the defective bottle or particle sample before cleaning the line.

Do not use a normal packaging sampling limit to accept unidentified contamination.

Microbial Failure

Simple filtration does not prove that a batch with a microbial failure is safe.

FDA lists contaminated raw materials, water, poor manufacturing conditions, packaging, and storage among the ways cosmetics can become contaminated with harmful microorganisms.^[4]

• sampling;
• laboratory method;
• water;
• raw materials;
• tank and hose cleaning;
• filling equipment;
• staff hygiene;
• packaging.

A high-alcohol fine fragrance does not have the same microbial risk as a water-rich body mist.

Repeat the test only when there is a documented reason to question the original sample or method. Do not keep testing until one result passes.

Keep the failed result in the batch record even if the repeat test passes.

Filling and Refilling

Measure underfill and overfill. Do not judge them only from the liquid line.

A six-head filler gives these results:

• Nozzle 1: 49.9 mL;
• Nozzle 2: 50.1 mL;
• Nozzle 3: 49.8 mL;
• Nozzle 4: 48.8 mL;
• Nozzle 5: 50.0 mL;
• Nozzle 6: 50.1 mL.

The average is about 49.8 mL, but nozzle 4 is clearly low.

Glass bottle weight can vary too. A single tare weight may distort the result if the bottle lot has enough variation.

A 50 mL perfume with a measured density of 0.86 g/mL contains about 43 g of liquid:

50 mL × 0.86 g/mL = 43.0 g

Use the actual density of the formula when the filling line controls product by weight.

If 2,000 bottles enter a refilling operation but only 1,960 finished units and rejects are counted, the factory must account for the missing 40 bottles.

Keep recovered perfume linked to its original batch. Do not pour consumer returns or mixed bottles with unknown storage history back into normal production bulk.

Pumps and Leakage

Write down the real pump defect, not just "pump failed."

• no spray;
• weak spray;
• side spray;
• uneven spray;
• dripping;
• slow priming;
• neck leakage;
• loose crimp.

A blocked pump may be faulty, or particles in the perfume may be blocking it. Leakage may come from the bottle neck, gasket, crimp setting, or material compatibility.

Check spray amount, pattern, priming, leakage, closure fit, and dip-tube position against the bottle-and-pump specification.

On a 5,000-bottle rework run, one internal sampling plan might check 10 bottles from the start, 10 from the middle, and 10 from the end:

10 + 10 + 10 = 30 units

The actual sample size should come from the factory's inspection plan.

A bottle can pass at room temperature and leak during warm storage. Heat changes headspace pressure and can affect gaskets, plastics, and closures.

In XUELEI's perfume bottle compatibility testing, the perfume is checked with the bottle, pump, gasket, dip tube, and closure together.

Bottles and Cartons

• A light decoration scratch may be cosmetic.
• A neck defect may affect sealing.
• A sharp chip or crack can create a safety risk.

Use clear defect samples or photos so inspectors are working to the same standard.

If bottles are cleaned or solvent-wiped during rework, check again for coating damage, lost printing, adhesive failure, or surface cracking.

When cartons are replaced, confirm the bottle SKU, fragrance, fill size, batch, target market, barcode, insert, and carton version.

If 3,000 wrong cartons were issued but only 2,940 were removed, locate the missing 60:

3,000 - 2,940 = 60 cartons

After repacking, check the bottle, label, insert, and carton as one finished SKU.

Labels

• product name;
• net contents;
• ingredient list;
• batch code;
• barcode;
• required warnings;
• responsible-party information;
• target-market language;
• fragrance-allergen information.

Raising the fragrance concentration can change the allergen declaration.

Under EU fragrance-allergen rules, fragrance allergens covered by Annex III must be listed individually when they exceed 0.001% in leave-on products or 0.01% in rinse-off products.^[5]

0.001% = 0.00001 = 10 ppm

This is a labeling threshold, not a universal safety limit.

Fine fragrance used on skin falls into the leave-on category. Products that did not meet the newer requirements could be placed on the EU market only until 31 July 2026 under the transition rules. Qualifying products already placed on the market may continue to be made available until 31 July 2028.^[5]

XUELEI's perfume sample approval process checks IFRA documents, allergen data, safety documents, and final artwork separately.

Reinspection

Reinspection needs to show that the original defect is gone and that the rework did not create a new one.

After refiltration:

• clarity;
• particles;
• odor;
• color;
• return of haze or sediment.

After formula adjustment:

• final calculation;
• odor;
• appearance;
• needed physical or analytical tests;
• IFRA and allergen review if the formula changed;
• stability risk.

After reopening bottles:

• cleanliness;
• fill quantity;
• closure;
• leakage;
• spray;
• appearance.

After relabeling:

• correct artwork;
• position and readability;
• batch code;
• barcode;
• correct market version.

Define the pass limits before testing.

If 5,000 bottles were reworked because of defective pumps, the final inspection should cover more than the first few bottles after setup. The sampling plan should cover the whole rework run.

Finish the required mixing or holding time before taking the sample. Keep a pre-rework and post-rework sample when comparison is needed.

Sampling

AQL is used for countable packaging defects such as scratches, carton damage, print position, or cap alignment.

It is a sampling method. An accepted lot can still contain some defective units.

Do not use AQL to accept unknown contamination, a wrong formula, or uncertain product identity.

If 4,000 bottles may carry one of two visibly different labels, and the wrong label can be identified by sight, checking all 4,000 bottles may make more sense than relying on a small sample.

Stability

ISO/TR 18811:2018 states that one stability test does not suit every cosmetic. The manufacturer has to choose and justify the conditions used for the actual product.^[6]

• dilution change: check clarity, odor, and package compatibility;
• color adjustment: check color drift and light response;
• refiltration: watch for haze or sediment returning;
• new pump or gasket: check leakage and material compatibility.

XUELEI's perfume manufacturing process checks liquid stability and finished-package compatibility separately.

Do not copy a temperature or test period from another perfume. Use the conditions set for that formula and package.

Rework does not reset the age of the batch. Use the original manufacturing date, storage history, and stability data when deciding how much shelf life remains.

Release or Reject

Before release, check:

• the original defect;
• affected quantity;
• identified cause;
• approved correction;
• materials added or removed;
• final quantity;
• reinspection results;
• regulatory review where needed.

EU Regulation (EC) No 1223/2009 requires cosmetic manufacturing to follow good manufacturing practice and requires a safety assessment before a cosmetic product is placed on the market.^[7]

ISO 22716:2007 gives GMP guidelines for cosmetic production, control, storage, and shipment.^[8]

In the United States, MoCRA requires the responsible person to ensure and maintain records supporting adequate safety substantiation for a cosmetic product.^[9]

Keep the batch on hold if the contamination source, product identity, or formula cannot be confirmed. Reject it if the remaining risk cannot be controlled.

Repeated "adjust, test, fail, adjust again" adds transfer loss, evaporation, oxidation, handling, and contamination risk.

Root Cause and CAPA

Correction fixes the current batch. Corrective action fixes the process that caused the problem.

Weak root cause: "Operator error."

Better: "The operator used an obsolete formula sheet that was still available at the weighing station."

Weak root cause: "Machine problem."

Better: "Filling nozzle four delivered 1.8 g below target because its valve seal was worn."

CAPA has four parts:

• Correction: repair or sort the affected product.
• Root cause: find why the defect happened.
• Corrective action: change the process that allowed it.
• Effectiveness check: check whether the same cause returns.

With the worn valve example, correcting bottle fill fixes the current batch. Replacing the seal, adding it to preventive maintenance, and checking nozzle output during setup address the cause.

If nozzle four stays within its fill limit in later runs, the fix is working.

Records

The record should show exactly what happened to the batch.

• original batch number and SKU;
• original and affected quantities;
• the defect;
• date and production stage;
• investigation and root cause;
• rework method;
• materials added or removed and their lot numbers;
• equipment and operators;
• start and finish time;
• final quantity and losses;
• sample IDs;
• tests and results;
• release, hold, or rejection decision.

If a 500 kg batch receives 10 kg of solvent and loses 2 kg during sampling and transfer:

500 + 10 - 2 = 508 kg

If only 492 kg remains, another 16 kg is missing from the record. That needs to be explained before release.

Keep the original failed result even if the reworked batch later passes.

Buyer Checks

Ask the supplier:

• What exactly failed?
• How many units were affected?
• What caused it?
• Did the formula change?
• What was added or removed?
• What was tested again?
• Who released the batch?

If 10,000 bottles were affected but the report covers only one 200-bottle production segment, ask how the other 9,800 bottles were checked.

For a formula rework, the records may include the deviation report, rework instruction, calculation, test results, and release record.

XUELEI's fragrance quality-control process links formula records, raw materials, production, packaging, and test results to the same batch.

Does Rework Restart Perfume Shelf Life?

No. If bulk perfume was made 60 days ago and spent another 20 days on hold, it is already about 80 days old after rework.

Repacking or correcting the batch does not turn it back into a day-zero product.

Can Reworked Perfume Be Sold as Normal Stock?

Yes, if the reworked batch meets the approved specification, the needed safety and regulatory checks are complete, and QA releases it.

Can a Failed Batch Be Mixed With a Good Batch?

Do not use a good batch to dilute a failure that has not been explained.

Mixing 100 kg of failed material with 900 kg of good product may reduce the average concentration of a known component by a factor of ten. It does nothing to explain unknown contamination, the wrong raw material, or an off-odor.

If blending is allowed, calculate the new formula, keep both batch numbers traceable, and test the combined batch again.

Can Returned Perfume Go Back Into Production?

Consumer and distributor returns have a different history from bottles kept under factory control.

Do not reuse returned perfume unless its batch, storage history, condition, and handling are known and the factory has an approved recovery procedure.

Conclusion

A 100 kg batch containing 18 kg of fragrance concentrate needs 2.5 kg more concentrate, not 2 kg, to reach 20% by weight. That correction only makes sense when the original weighing error is proven. After rework, mix the batch, take the right sample, repeat the affected tests, and keep the original failure in the batch record. If the formula, contamination source, or batch identity cannot be confirmed, stop adjusting the product and keep it on hold or reject it.