
When a customer reports a defect, the questions are immediate: which batch, which material, which machine, which shift, which other customers received parts from the same lot? Without traceability, answering takes days of searching registers, and the safe response is to recall far more than necessary. With it, the answer takes minutes. For background, read manufacturing traceability: lot genealogy and recalls.
This guide explains how custom traceability software is designed and built, the decisions that shape it, and what it costs to develop.
Start with the questions you must answer
Traceability is only as detailed as it needs to be. Before any design, list the questions customers, regulators and your own quality team ask:
- Which raw material lots or heats went into this finished lot or serial number?
- Which customers and shipments received products made from this material lot?
- Which machine, mould or die, operator and shift produced it?
- What were the inspection results and process parameters?
- Which job work vendor processed it?
- Were any deviations or concessions approved?
Each question implies data that must be captured at a specific step. That is the traceability design.
Identity: lot, batch, heat or serial
- Supplier lot or heat — the identity of incoming material, from the supplier's certificate. See heat number traceability.
- Internal lot or batch — a quantity produced under the same conditions: a production batch, a shift's output on one machine, a mixing batch.
- Serial number — unit-level identity for products that need it.
- Container or package ID — each bag, coil, bin or pallet labelled uniquely, which makes scanning precise.
Define rules for when a new lot starts: a new material lot, a new shift, a tool change, a process change. Smaller lots mean narrower recalls but more records.
Capture points
- Goods receipt — supplier lot or heat recorded, certificates attached, labels printed.
- Issue to production — material labels scanned against the production order.
- Each operation — machine, operator, time, parameters and output lot recorded; where lots merge or split, the relationship is captured.
- Inspection — results linked to the lot.
- Job work — lots sent out and received back with the same identity.
- Packing — finished lots or serials linked to packages and labels.
- Dispatch — packages scanned against invoices and customers.
Every capture point must be quick — a scan, not a form — or it will be skipped.
Genealogy: splits, merges and rework
Real processes are not straight lines. A coil is slit into several narrow coils; two batches of compound are blended; a lot is partly reworked and partly scrapped; components from several lots go into one assembly. The data model must record these relationships as a graph so that a trace follows every branch. This is the technically most important part of traceability software, and the most common weakness in simple systems.
Tracing and recalls
- Backward trace — from a finished lot or serial to all inputs, operations and results.
- Forward trace — from an input lot to all outputs, stock locations and customer shipments.
- Recall workflow — affected stock blocked automatically, shipments listed by customer, notifications prepared, and actions tracked.
- Mock recalls — run periodically to prove the system works.
Industry requirements
- Automotive — customers commonly expect lot traceability, part marking and fast response to complaints. See handling OEM rejections.
- Pharma and nutraceuticals — batch records and full material genealogy. See electronic batch records.
- Food — one-step-back, one-step-forward traceability at minimum, with expiry.
- Steel and forging — heat numbers from mill certificates through to finished parts.
- Electronics — component reels and serial-level assembly records.
Integrations
- Inventory — lot-level stock and status.
- MES — operations, machines, operators and parameters. See MES software development.
- Quality — inspection results, NCRs and complaints. See QMS development.
- ERP / invoicing — customer shipments.
- Printers and scanners — labels and scanning at each capture point.
What drives development cost
- Depth: lot-level versus serial-level.
- Number of capture points and the complexity of splits and merges.
- Labelling and scanning hardware.
- Integration with inventory, MES, quality and invoicing — or building those parts too.
- Customer and regulatory report formats.
- Data volumes for high-speed lines.
Rollout
- Define questions, identities and lot rules.
- Start with receipt labelling and issue scanning — the foundation of backward traceability.
- Add packing and dispatch scanning for forward traceability.
- Add operations, parameters and inspection detail.
- Run a mock recall and refine.
A worked example: a forging and machining company
A company forges steel parts from bar and machines them for automotive and industrial customers. Its customers require traceability from finished part back to the steel heat. Before the system, heat numbers were written on job cards and copied by hand between departments, and a trace took a day or more.
With the new system, bars are labelled with their heat at receipt. The forging shop scans the heat when bars are issued to a forging lot; forged lots carry the heat through heat treatment, shot blasting and machining as scanned labels on bins; and packing links finished lots to boxes and invoices. A backward trace from a customer's part number and lot to the heat and mill certificate now takes minutes, as does a forward trace from a suspect heat to every customer shipment. See forging shop software.
Label design tips
- Include a unique ID as a barcode or QR code, plus human-readable item, lot and quantity.
- Choose label material for the environment: heat, oil, washing, outdoor storage.
- Print labels at the point of creation — receipt, production, packing — not in advance.
- Place labels consistently so they are easy to scan.
- For parts that must carry identity themselves, use direct part marking such as dot peen or laser marking.
Common mistakes
- Lots too large — a recall then covers far more product than necessary.
- Identity lost at mixing points — bins combined without recording both sources.
- Manual entry of lot numbers — typing errors break the chain; scan instead.
- Traceability only on paper at job work vendors — challans must carry lot identity.
- Never testing — run mock recalls regularly.
Working with Corposol
Corposol builds traceability software together with inventory, MES and quality systems for manufacturers. We visit factories on-site across India, work with clients abroad remotely or on-site depending on the project, deliver in phases, and transfer all custom code on final payment.
Frequently asked questions
What is the difference between backward and forward traceability?
Backward traceability starts from a finished product and finds the materials, processes, machines and people involved. Forward traceability starts from a material lot or event and finds every product and customer it reached. A recall needs both.
Do we need serial numbers for every part?
Not always. Many factories trace by lot or batch, which is simpler and sufficient for most customers. Serial tracking is used where customers or regulations require unit-level history, or for high-value products.
How quickly should we be able to trace?
With traceability software, a backward or forward trace should take minutes. Many customers and auditors now test this with mock recalls.
Can traceability work if part of our process is done by job work vendors?
Yes. Material sent to vendors keeps its identity on the challan, and returned material is received against the same lot, so the chain is not broken.
