Retained Samples and Batch Traceability for Candy Orders: How Buyers Build an Evidence Chain

14 min read

Short Answer

A retained sample is useful only when the Buyer can prove what it is, which finished-goods lot it represents, how it was selected, where it was stored and which production records belong to it. A sealed bag kept in a cabinet without lot identity is not a reliable evidence system.

For candy orders, the strongest evidence chain connects:

approved product standard → raw-material and packaging lots → processing batch → packing run → finished-goods lot/date code → retained production sample → shipment/pallet records → warehouse receipt → downstream customer or complaint sample.

Three evidence types must not be confused:

Evidence type Main question it answers What it does not prove by itself
Approved sample / reference What should the product be like? What a later commercial lot actually was
Retained production sample What did this specific production/packing lot look, taste and behave like when retained? Complete upstream/downstream traceability or transport history
Traceability records Where did the lot come from and where did it go? Sensory or physical quality of the retained product

A professional Buyer should therefore define the retained-sample and traceability system before production, not after a complaint. The system should be risk-based: a first commercial run, a new factory, a filled or layered gummy, a moisture-sensitive freeze-dried candy or a product with a sensitive certification claim usually deserves stronger controls than a mature, stable reorder.

Buyer takeaway: The goal is not to keep more candy. The goal is to preserve enough representative product and linked records to reconstruct a commercially important event months later.

Buyer Decision Summary

Decision Main variables Primary risk Verification required
Which lots require retained samples? risk, first vs repeat run, line/packing structure, customer requirements Important lot has no physical reference Written retention matrix linked to production plan
How many samples and from where? Run length, start/middle/end variability, multiple lines, coating/filling/packing complexity Sample is not representative Defined sampling points and sample IDs
Where should samples be stored? Product sensitivity, commercial packaging, storage claim, dispute risk Retained sample changes because storage was uncontrolled Storage conditions and location record
How long should samples and records be retained? Law, customer/scheme requirements, , complaint window, internal risk Evidence is destroyed before it is needed Documented retention rule; no invented universal duration
Does the Buyer need a second retained set? Import route, long transit, claim value, dispute likelihood Factory sample cannot distinguish factory vs downstream damage Factory + destination control sets for selected projects
Is lot coding sufficiently granular? Cooking batch, packing date, film/artwork revision, multiple lines Affected population cannot be isolated Lot architecture mapped to actual production flow
Can traceability be demonstrated? Supplier lots, process records, packaging materials, shipment records Lot code exists but evidence chain breaks upstream/downstream Periodic traceability test and mass-balance exercise

1. Retained Samples Are an Evidence-Control Tool, Not a Souvenir

Factories, importers and buyers often use the term "sample" for very different objects. That creates problems when a dispute appears six months after delivery and people discover that the only surviving reference is an R&D sample from before mass production.

A retained production sample should normally be taken from actual commercial output, identified and protected so that it can later support comparison, complaint investigation, shelf-life review, change control or traceability verification.

It should not be confused with:

AXTIMES sample-control work already treats sample identity as a controlled project element: product code, version, purpose, preparation/production date and batch reference should travel with the sample, rather than relying on informal descriptions such as "the new sour one." (AXTIMES Knowledge Base: Shipping Candy Samples from China, internal source)

2. Batch, Lot and Packing Run: Map the Factory's Real Language

There is no useful buyer control in arguing whether a factory "should" call something a batch or a lot. The Buyer needs to understand the factory's actual boundaries.

A typical candy flow may include:

  1. raw-material receiving lots;
  2. one or more cooking / mixing / depositing batches;
  3. conditioning, drying or curing periods;
  4. coating, oiling, sour sanding, filling or secondary processing;
  5. temporary work-in-progress holding;
  6. packing run on one or more packaging lines;
  7. finished-goods lot/date code;
  8. pallet/carton allocation and shipment.

One cooking batch may be packed across two days. Two cooking batches may be combined into one packing campaign. A film roll or artwork revision can change during the same day. A third party may perform coating or repacking. If the printed lot code does not distinguish the point that matters during a failure, the Buyer may have to hold far more product than necessary.

Practical lot-architecture questions

Question Why it matters
What event creates a new production batch? Establishes upstream processing boundary
What event creates a new finished-goods lot? Defines what the consumer/warehouse can identify
Does a new packing date create a new code? Helps separate packaging-line failures
Are different lines or shifts distinguishable? Important if one line or shift creates the problem
Is the packaging-material lot recorded? Supports seal, migration, printing and artwork investigations
Is rework linked to the finished lot? Prevents invisible mixing of evidence histories
Are outsourced stages traceable? Maintains chain through coating, repacking or other third parties

IFS Food Version 8 requires a documented traceability system that identifies product lots and relates them to raw-material batches and relevant food-contact packaging materials, with records through receipt, processing, rework and distribution until delivery to the customer. It also requires periodic traceability testing, including mass balance. This is a useful benchmark for professional Buyer expectations even where IFS certification is not contractually required. [Source: IFS Food Version 8, section 4.18.]

Buyer takeaway: A printed lot code is only the visible key. The real control is the database or record set behind that key.

3. Build the Evidence Chain Before the Purchase Order Is Released

The Buyer should define the chain before production because some evidence cannot be recreated later.

A practical evidence chain is:

Stage Evidence to preserve
Approved standard Current product specification, approved physical reference ID, artwork revision, approved claims
Raw materials Supplier/manufacturer identity where relevant, internal material code, incoming lot, certificates/specifications required by the project
Packaging materials Film/jar/closure/carton supplier or internal code, material lot, print/artwork revision
Processing Production date, batch number, line, shift where relevant, critical process records
Secondary operations Conditioning/coating/filling/rework/outsourced-step records
Packing Packing date, packaging line, finished lot/date code, weight and seal controls
Finished-goods release Quantity, /inspection, agreed /test reports, deviation status
Retained sample Unique sample ID, exact lot, sampling point, quantity, storage location/condition
Shipment Carton/pallet allocation, quantity by lot, handover/shipment document
Destination receipt Received quantity by lot, date/expiry capture, damage/hold status, storage condition where relevant
Downstream Customer/channel allocation by lot where commercially or legally required

The traceability chain should be proportionate. A Buyer does not need to copy every factory record into its own system, but it should know which records exist, who controls them, how long they are available and how quickly they can be retrieved.

4. What a Retained-Sample Record Should Contain

A physical pack without metadata loses most of its value. At minimum, the retained-sample record should allow an independent person to identify the sample later without relying on memory.

Core fields

For a first run or disputed product, add more granularity. For a stable routine SKU, the system can be leaner, provided traceability remains intact.

5. How to Select Representative Retained Samples

"Keep one bag" is not a sampling strategy.

The selected units should be representative of the commercial lot and of the process risks the Buyer wants to preserve. Sampling only the most attractive pack from a finished pallet creates a biased reference.

Risk-based selection logic

Production situation Stronger sampling approach may include Reason
First commercial run Start, stable middle and end of packing run Captures setup and end-of-run drift
Long run over multiple shifts Samples across relevant shifts Shift/process variation may matter
Multiple packing lines Sample from each line or distinguish line code Seal/fill performance can differ
Sour-coated candy Samples across coating/conditioning and packing periods Moisture/coating equilibrium may drift
Filled/layered gummies Samples across run, especially after setup changes Filling ratio, centering and leakage can vary
Freeze-dried candy Samples after final drying and from final sealed packs Moisture barrier and breakage are critical
New packaging material Samples before and after roll/material change Isolates packaging-related failures
Routine stable reorder Reduced but documented representative set Avoids unnecessary burden while preserving evidence

The article does not prescribe a universal number of bags per lot. That number should follow the product risk, customer requirement, company policy and likely investigation needs.

6. Keep the Product in Its Real Commercial Pack Where Possible

For buyer-facing quality evidence, the most useful retained sample is often the finished, sealed commercial pack from the actual lot. This preserves the interaction between candy and packaging.

That matters because later defects may involve:

A loose candy sample in a laboratory jar can still be useful for specific tests, but it does not reproduce the commercial pack system.

Separate retained evidence from accelerated or abuse testing

Do not use the only legal/commercial retained sample for destructive shelf-life work. Keep separate, clearly marked groups for:

Once a retained pack is opened, record the opening date and purpose. An opened pack should not later be presented as if it were an untouched control.

7. Storage Conditions Can Destroy the Evidence

A retained sample can become misleading if it is stored differently from the claimed product conditions or under uncontrolled heat, humidity, light, odor or compression.

The retention procedure should define:

For shelf-stable candy, the correct system is not automatically "keep it in an office cabinet." The storage method must be appropriate for the product and the evidence question.

8. How Long Should Retained Samples Be Kept?

There is no universal AXTIMES rule such as "two bags for 24 months." A retention period must be built from the applicable obligations and commercial risk.

Inputs can include:

IFS Broker guidance provides a useful example of how retained samples can be treated as an agreed control: it asks whether customers request retained samples, whether representative sampling is agreed with the manufacturer, where samples are kept, under what conditions, and whether an up-to-date retained-sample list can be compared with deliveries. [Source: IFS Broker Guideline.]

The Buyer should therefore define a retention rule, not copy a random duration from another product category.

Planning framework: Retention end date should normally be no earlier than the longest applicable legal/customer/internal evidence requirement. Where requirements conflict, the stricter applicable requirement should be reviewed with the responsible QA/legal team.

9. Factory Retained Sample vs Destination Control Sample

For selected orders, keeping the same evidence at two locations is extremely valuable.

Factory set answers questions such as:

Destination warehouse / Buyer set answers different questions:

A good Factory retained sample does not automatically prove that the Factory is not responsible. It may indicate that the defect emerged later, but the evidence still needs to consider manufacturing, packaging, transport and storage hypotheses. Conversely, a defective destination control does not automatically prove transport damage.

When a dual set is most useful

The physical sample is strongest when it is the index to a digital record set.

A practical lot evidence folder can contain:

  1. purchase order / order confirmation;
  2. current product specification;
  3. approved sample/reference ID;
  4. approved artwork and packaging revision;
  5. ingredient/material declarations required by the project;
  6. relevant supplier/ingredient certificates where controlled;
  7. factory batch/packing record reference;
  8. raw-material and packaging-material lot links where required;
  9. in-process and finished-goods QC;
  10. COA or agreed laboratory reports;
  11. pre-shipment inspection report;
  12. deviation/concession/CAPA documents;
  13. finished quantity by SKU and lot;
  14. shipment/pallet/carton allocation by lot;
  15. warehouse receipt record;
  16. complaint or return records if later created;
  17. retained-sample register entry and photos.

The Buyer does not need to receive every proprietary factory production record automatically. The important control is that the agreed evidence can be retrieved and the relationships are known.

11. Traceability Is Upstream and Downstream

Codex defines traceability/product tracing as the ability to follow the movement of food through specified stages of production, processing and distribution. Codex CXG 60-2006 treats traceability as a tool that should be designed around defined objectives rather than assumed to solve every food-control problem by itself. [Source: Codex CXG 60-2006.]

For a candy Buyer, practical traceability therefore has two directions.

Upstream traceability — from finished lot back toward:

Downstream traceability — from finished lot toward:

The EU's general food-law framework is one example of a market where traceability is a legal requirement across food businesses. Article 18 of Regulation (EC) No 178/2002 requires operators to be able to identify suppliers and the businesses to which products have been supplied, with systems/procedures available to competent authorities. Exact implementation and additional sector rules must be checked for the destination market. [Source: EUR-Lex, Regulation (EC) No 178/2002, Article 18.]

Important boundary: The AXTIMES evidence-chain framework is broader than minimum legal "one-step" traceability. It is designed for commercial control of a private-label order, so it may link more manufacturing and quality evidence than the legal minimum requires.

12. Traceability Testing: Do Not Wait for a Real Incident

A traceability system that has never been tested is only a hypothesis.

A mock traceability exercise should select a real finished-goods lot and ask the Factory or supply-chain team to reconstruct the chain in both directions within a defined target time.

Test questions

IFS Food v8 requires the traceability system, including mass balance, to be tested at least once within a 12-month period or whenever significant changes occur, with test samples reflecting the complexity of the product range. This is a strong operational benchmark for factories using that standard. [Source: IFS Food Version 8, 4.18.2.]

13. Mass Balance: The Quantity Check Behind Traceability

Traceability is not complete simply because a spreadsheet contains matching codes. Quantities should make commercial sense.

A simplified mass-balance framework is:

input / produced quantity → released finished goods + retained samples + QC/test use + rework + scrap/destruction + remaining stock.

For a finished-goods shipment, a Buyer can perform a narrower reconciliation:

packed quantity by lot → shipped quantity + warehouse/held quantity + inspection/retained units + documented loss.

The exact model depends on process yield and factory accounting. The goal is not mathematical perfection to the gram for every gummy process. The goal is to identify unexplained quantity gaps large enough to undermine the claimed traceability chain.

14. Packaging Traceability Matters as Much as Candy Traceability

Candy complaints are often packaging complaints in disguise.

Examples:

IFS Food v8 explicitly links finished-product lots to relevant food-contact packaging materials. A Buyer should also preserve the artwork revision and material identity, because two visually similar pouches may not be functionally equivalent.

Packaging sub-lot controls may include

15. Warehouse Receipt Is Part of the Evidence Chain

Traceability should not stop when the goods leave China.

At destination receipt, capture at least:

AXTIMES reorder planning already treats lot/date capture, remaining shelf life, damage segregation and open complaints/deviations as operating data for the next order rather than as an archive-only exercise. (AXTIMES Knowledge Base: Candy Reorder Planning, internal source)

If a warehouse strips pallet/carton identity, mixes lots into one bin and ships without lot-level records, a good Factory traceability system can become useless downstream.

16. Worked Buyer Example A — Dual Retained Sets for a First Commercial Run

Illustrative numbers only; they are not an AXTIMES universal sampling rule.

Inputs

A first commercial order contains:

Calculation

Factory retained units: 6 lots × 6 packs = 36 packs.

Destination retained units: 6 lots × 6 packs = 36 packs.

Total controlled retention: 72 packs.

Constraint

This is more administrative work than keeping one sample per SKU. However, it allows comparison of Factory and destination controls by exact finished lot.

Buyer Decision

Use dual-location retention for the first two commercial runs. After complaint history and repeatability are reviewed, reduce the routine quantity if risk supports it.

Why this example matters: The decision is based on evidence value, not on a universal "six packs" standard.

17. Worked Buyer Example B — Isolate the Affected Lot Instead of Holding Everything

Illustrative numbers only.

Inputs

A Buyer receives 8 complaints about sticky gummies. Complaint codes show:

Analysis

The retained sample of B-2407 and destination control from the same lot can be compared with adjacent lots. Warehouse storage and shipment history can be checked specifically for B-2407.

Constraint

Concentration in one lot does not prove the Factory root cause. It only narrows the affected population and investigation scope.

Buyer Decision

Place remaining B-2407 inventory on hold while investigating, rather than automatically blocking every lot in the entire product family. Expand the hold only if evidence shows a broader affected population.

18. Worked Buyer Example C — One Cooking Batch, Two Packaging Revisions

Illustrative scenario.

One gummy cooking batch is packed over a long run. The first half uses pouch revision R3. After a film-roll change, the second half uses R4. Weeks later, seal complaints occur almost exclusively in R4 packs.

If the finished-goods lot code only identifies the cooking batch, the Buyer cannot quickly isolate the packaging change. If the packing record links finished sub-lots to R3 and R4, the investigation can focus on packaging material, seal settings and changeover timing.

Buyer decision: when packaging changes during one processing batch, preserve a packing sub-lot or revision link even if the candy formula itself has not changed.

19. Buyer Decision Matrix

Situation Retained-sample intensity Traceability detail Buyer control
New Factory / first commercial run Enhanced Full lot + line/packing/date links Factory and destination sets; first-run review
Stable routine reorder Standard risk-based Lot/date + current material/spec links Periodic verification
New formula/gelling system Enhanced Formula/material revision + lot Link change-control approval and sample
New pouch/film/closure Enhanced around change Packaging lot/revision + finished sub-lot Sample before/after change where practical
Filled/layered gummy Enhanced Processing + filling/packing records Preserve samples across run
Sour-coated product Enhanced if moisture/coating sensitive Coating/conditioning + packing timing Link defects to process period
Freeze-dried candy Enhanced Drying + final pack material/lot Preserve sealed finished packs
Short seasonal item Risk-based Enough to support complaint/recall period Avoid excessive archive but do not destroy evidence too early
Sensitive Halal/Kosher/allergen claim Enhanced documentary chain Ingredient/material/certification scope + lot Link evidence to exact SKU/site/lot route
Outsourced coating/repacking Enhanced Third-party lot/handover link Do not allow traceability to break at subcontractor

20. What Changes the Answer?

The retained-sample and traceability design should become stronger when any of these variables increase:

The system may be simpler for a mature, low-risk SKU with strong lot coding, stable packaging and demonstrated traceability performance.

21. When This Rule Does Not Apply or Needs Modification

This guide is a commercial quality-control framework, not a universal legal retention schedule.

Special cases include:

Do not dispose of evidence after a complaint, hold, recall, legal notice or unresolved deviation merely because the routine retention period has expired.

22. Failure Scenarios

Initial assumption What was overlooked Operational consequence Correct control
"The Factory keeps samples." Samples have no finished-lot identity. They cannot prove which commercial lot they represent. Unique sample ID linked to finished lot and record set
"One nice-looking pack is enough." Selection was biased and not representative. Later comparison hides start/end or line variation. Risk-based representative sampling points
"The lot code proves traceability." Upstream raw/packaging records are not linked. Cause and affected population cannot be reconstructed. Upstream/downstream traceability map + test
"The Factory sample is good, so transport caused the issue." Factory sample storage and package history differ from destination. Wrong party may be blamed. Compare factory, destination and complaint evidence; reconstruct timeline
"We stored the sample for two years." Storage was hot, humid or exposed to light. Retained sample deteriorates independently. Defined storage conditions and monitoring where needed
"Warehouse has the goods, so lot data exists." Lots were mixed after depalletizing. Downstream isolation becomes impossible. Lot/date capture and allocation discipline
"Keep everything forever." No disposal ownership or record lifecycle. Archive becomes expensive and unreliable. Written retention/disposal rule + incident hold override

23. Buyer Control Table

Control item Owner When checked Evidence required
Lot architecture Factory + AXTIMES Before first PO / after process change Written process/lot map
Sample-retention matrix Buyer QA + Factory Before production SKU/lot retention rule
Approved reference link Buyer + AXTIMES Before mass production Sample ID + specification revision
Raw-material traceability Factory QA Production/release Batch/material records available
Packaging-material traceability Factory QA / Packaging Packing/release Material lot + artwork revision link
Retained sample selection Factory QC / inspector During/after packing Sample register + lot/selection point
Storage Factory QA / Buyer QA During retention Location/condition record
Finished quantity by lot Factory / AXTIMES Pre-shipment Lot quantity reconciliation
Shipment lot allocation Logistics / Factory Handover Packing/pallet/shipment record
Destination lot capture Importer/Warehouse Receipt Receiving record
Traceability test Factory QA / Buyer audit Periodic / after major change Upstream/downstream result + mass balance
Disposal or extended hold QA Retention end / incident Disposal record or evidence-hold instruction

24. Information Needed Before Designing the System

The Buyer should collect:

25. Questions the Buyer Should Answer Internally

  1. Which decisions do we expect retained samples to support: quality comparison, complaint investigation, shelf-life review, legal claim, retailer audit or all of these?
  2. Which SKUs are high enough risk to justify dual-location retained sets?
  3. Can our warehouse receive and ship by lot, or do we destroy downstream traceability after receipt?
  4. Who owns the retained-sample register?
  5. Who can authorize opening a retained pack?
  6. What event stops routine disposal and places evidence on hold?
  7. How long must commercial and quality records remain retrievable?
  8. Do we need line/shift/packing-revision granularity beyond the consumer lot code?
  9. What is our target time for reconstructing an affected lot during a complaint?
  10. Can we identify customers or channels that received an affected lot?
  11. Which legal/customer requirements apply in the destination market?
  12. How will we review traceability performance before a reorder?

26. Questions to Ask the Factory

  1. How do you define a processing batch and finished-goods lot for this SKU?
  2. What creates a new lot code: cook, date, shift, line, packaging run or another event?
  3. Can one finished lot contain more than one cooking batch?
  4. Can one cooking batch be divided into more than one finished lot?
  5. Which raw-material lots are linked to a finished lot?
  6. Are food-contact packaging material lots and artwork revisions linked?
  7. How is rework recorded and linked?
  8. Which steps are outsourced and how is traceability preserved through them?
  9. What retained samples do you normally keep, and for how long?
  10. Are retained samples sealed finished commercial packs?
  11. How are representative sampling points chosen?
  12. Where and under what conditions are samples stored?
  13. Can you provide a retained-sample register for our lots without disclosing unrelated customers?
  14. How quickly can you retrieve upstream and downstream records for one lot?
  15. How often do you test traceability and mass balance?
  16. What happened in the most recent traceability test: scope, retrieval time and corrective actions if any?
  17. How are lot/date codes linked to pallets and shipments?
  18. What happens to records and samples when a complaint or legal hold is open?

27. Implementation Workflow

  1. Define the evidence objective. Decide what the system must prove later.
  2. Map the production chain. Identify in-house and outsourced steps.
  3. Map batch and lot boundaries. Document the Factory's real coding logic.
  4. Classify SKU risk. First run, sensitive packaging, filling/coating, long shelf life, claims and complaint history increase control needs.
  5. Define the retained-sample matrix. Specify which lots, sample points, quantities and locations.
  6. Define sample IDs. Make them unique and linkable to digital records.
  7. Define storage and access. Protect normal-condition controls from uncontrolled use.
  8. Define record links. Specification, materials, processing, packaging, QC, deviations, shipment.
  9. Run production and select samples. Do not choose only after results are known.
  10. Reconcile finished quantity by lot. Preserve a usable mass-balance view.
  11. Link lots to shipment. Pallet/carton records should identify what was handed over.
  12. Capture destination receipt by lot. Record date/expiry and damaged/held stock.
  13. Test traceability periodically. Reconstruct upstream and downstream chain and quantity balance.
  14. Use evidence in complaint/reorder review. Do not treat retention as an archive-only activity.
  15. Dispose or extend hold under written rules. Suspend disposal whenever an incident is unresolved.

28. Buyer Retained-Sample & Traceability Record Template

Field Entry
Project / Buyer code
SKU / product
Variant / flavor / color
Finished-goods lot code
Production batch(es)
Production date
Packing date
Best-before / expiry
Production line / shift
Packing line
Specification revision
Formula/material revision
Artwork revision
Packaging material lot
Relevant raw-material lot references
Outsourced process reference
Rework used?
Retained-sample ID
Sampling point
Quantity retained
Sealed commercial pack?
Selected by / date
Storage location
Storage condition
Normal / accelerated / lab purpose
Linked QC / COA / test report
Linked inspection
Deviation / concession / CAPA
Shipment / pallet reference
Destination receipt reference
Retention/disposal date
Evidence hold status
Notes / opened date

29. AXTIMES Operational Insight

Amanda XUN, Head of Sourcing, AXTIMES: In export projects, the most useful evidence is rarely one photograph or one retained pack. The value comes from preserving the relationship between what the Buyer approved, what the Factory actually produced, which lot was shipped and what later appeared at the destination. We therefore treat lot/date capture, retained references and open complaint/deviation status as inputs to reorder control, not as paperwork that can be reconstructed after a problem.

AXTIMES has also found that sample identity can be lost surprisingly early: different development revisions, production samples and packaging samples can arrive in the same parcel. The control should begin before courier dispatch and continue through commercial production. The existing AXTIMES sample-shipping guide therefore calls for unique sample identification and batch/line references where available. (AXTIMES Knowledge Base: Shipping Candy Samples from China, internal source)

30. Buyer Checklist

Before production

During production / packing

Before shipment

At destination

Periodically

31. Relationship to Other AXTIMES Controls

This article owns the evidence infrastructure decision. It should not duplicate the full content of adjacent guides.

Related guide Its primary question
How to Investigate Candy Quality Complaints After Delivery How should a specific complaint be contained, investigated and assigned a root cause?
Candy Quality Inspection Before Shipment What should be inspected before release and how should defects be classified?
Candy Factory Change Control What must be revalidated after formula, supplier, process, line, site or packaging changes?
Candy Shelf-Life Evidence What evidence supports the declared shelf life?
Candy Reorder Planning When and how much should be reordered after reviewing previous-lot performance?
Laboratory Testing for Candy Orders Which laboratory tests should be selected, when, and for what decision?

This scope separation prevents search-intent cannibalization and makes the Knowledge Base useful as a decision system rather than a collection of overlapping "complete guides."

32. Source Notes and Evidence Boundaries

Externally verified framework facts

AXTIMES operational observations

Planning frameworks in this article

33. FAQ

Is a retained sample legally required for every candy lot?

Not universally. Legal, certification, customer and company requirements vary by market and product. The Buyer should determine applicable obligations and then add commercial controls where useful.

Is one retained pack per SKU enough?

Sometimes, but often not. A first run spanning multiple lots, lines or packing dates can require more representative sampling. The correct number is risk-based.

Should the retained sample come from the start or end of production?

For higher-risk or first runs, using more than one process point can capture setup or end-of-run variation. A stable routine lot may use a simpler documented method.

Can an inspection sample become the retained sample?

Yes, if its exact lot identity and selection are controlled and it remains sealed/undamaged. An opened or destructively tested inspection pack should not be treated as an untouched retained control.

Should the Factory or Buyer keep the sample?

Either can, but selected projects benefit from both Factory and destination control sets because they answer different timeline questions.

Does a good Factory retained sample prove the defect happened during transport?

No. It is one piece of evidence. Storage conditions, packaging history, destination control, complaint sample and lot/process records still need to be compared.

What is the difference between traceability and chain of custody?

Traceability reconstructs product/material movement and relationships through the supply chain. Chain of custody focuses on who possessed or controlled a specific sample/evidence item and under what conditions. Formal laboratory or legal evidence may require stricter chain-of-custody procedures than routine commercial retained samples.

Do lot codes need to identify the production line?

Not always. But the system behind the code should be granular enough to isolate commercially meaningful differences. If multiple lines have different failure risks, line linkage is useful even if it is not printed directly on the consumer pack.

How often should traceability be tested?

Follow applicable law, certification scheme, customer and company requirements. As one benchmark, IFS Food v8 requires its certified sites to test the traceability system, including mass balance, at least once in a 12-month period or when significant changes occur.

Can we discard samples at the end of the normal retention period if a complaint is still open?

No. An unresolved complaint, hold, recall, legal dispute or investigation should trigger an evidence hold under the responsible QA/legal procedure.

Before the next candy production run, create a one-page lot architecture + retained-sample matrix for each SKU family. Link the finished lot/date code to the current specification, packaging revision, production/packing records, retained sample and shipment allocation. Then run one mock traceability exercise before relying on the system during a real complaint.