Back
Best Practices
Should you split packing and QA on your line?
Deciding whether to split or embed hinges on risk, takt time, and defect impact.
Rabot
Content Contributor
Jun 4, 2025
Short answer: not always. Whether you gain by splitting them depends on what you are trying to control, how risky an error is, and how fast your line must run.
What “separation” really means
Model | Who looks for defects | When it happens | Typical tools |
|---|---|---|---|
Integrated QA (built-in) | The packer (or an in-line device) | During the standard work cycle | Barcode scans, in-line scales, vision or poke-yoke fixtures |
Separate QA (offline gate) | A dedicated inspector or audit team | After packing, before ship | Sampling plans, drop-tests, label audits, AQL tables |
Toyota’s jidoka principle in the TPS Handbook calls this building quality into the process and warns that:
inspection carried out by off-line inspectors yields no added value, so efforts are necessary to find ways to manufacture quality products with fewer inspectors
In practice it means wiring simple verifications straight into the operator’s work so that defects are trapped instantly rather than traveling downstream.
Pros of keeping QA inside the packing step
Instant feedback and correction: The operator fixes the mistake before the carton is sealed, so defects never move on. This aligns with “building in quality at each process brings the inspector’s function into each process”.
Shorter lead time: No separate queue or re-handling. That is pure waste in lean terms.
Higher ownership: Packers see the results of their own work; error data drives their kaizen.
Low marginal cost: If you already touch every item, adding a scan or vision check is pennies per order. Rabot’s barcode-plus-vision feature is a good example.
Pros of a separate QA gate
Independence and credibility: A neutral auditor can resist schedule pressure and certify compliance, a common customer requirement for regulated or high-value goods.
Broader inspection scope: Drop damage, torque, certificate-of-conformance paperwork, or allergen verification often need fixtures or skills that a packer does not have.
Structured sampling: A standalone team can run statistically valid AQL or ISO 2859 plans across lots, catching patterns the line crew might miss.
Common pitfalls of each approach
Integrated QA | Separate QA |
|---|---|
Risk of self-certification bias: The person who made the mistake is the one who approves the fix. | Long feedback loop: Defects found after the station require rework loops and hide the real root cause. |
Takt-time drag: Adding a 1s scan to every one-line order is trivial, but on a 40-line pick ticket it is 40s. | Extra touches and space: Moving cartons to an audit lane adds handling, conveyors, and dwell inventory. |
Blind to non-instrumented defects: A barcode will scan even if the item is dented. | Costs scale with volume: As orders grow, the head-count to inspect X% of them grows too. |
How best-in-class operations decide
Map the real failure modes. If the top customer complaints are “wrong SKU” and “wrong quantity”, integrating a scan is worth every second. If they are “crushed box” or “missing MSDS sheet”, an offline audit may be unavoidable.
Run an FMEA or risk matrix. Weight the severity of each error against its likelihood and the cost of inspection time.
Blend the layers. Many e-commerce DCs now scan 100% for identity but only sample weights or run vision checks on the 20% highest-risk SKUs. Others reverse it for fragile items.
Automate wherever possible. Jidoka teaches that the cheapest inspector is a device that stops itself. Sensors, scanners, and in-line scales enforce quality without adding labor.
Practical guidance for Rabot clients
Start with in-station verification for the defect you can fix immediately. Wrong item, wrong count, wrong label. Automate it.
Add a lightweight outbound audit for defects that need separate skills or fixtures. For example, crush testing a cosmetics shipper or validating hazmat documentation.
Feed every QA result back into the pick-pack process. Even if auditors are separate, their findings should light an andon for the cell that created the error so root cause is attacked inside the takt.
Review the mix quarterly. As scanners, AI vision, or new pack materials roll in, move checks upstream and shrink the offline gate. Lean never stops iterating.
Bottom line: you do not have to choose either-or. The lean ideal is zero defects with zero inspectors, achieved by embedding simple, reliable checks inside the standard work and reserving a small, high-skill QA team for what technology or line operators cannot yet cover.
Key takeaway
Digital QA layers give you sight, memory, and analytics that manual sampling or barcode-only checks cannot match. They move inspection from a cost center to a real-time coaching and proof engine, closing the last meter between “we think we packed it right” and iron-clad evidence that you did. For operations already scanning 100% of barcodes, Digital QA is the logical next step to protect brand reputation and drive continuous improvement at line speed.
Next steps
Spend 30 minutes this week with your quality or operations lead and walk through the last ten customer claims. For each one, ask: Would a near real-time review of video clips of the pack step by your quality team have prevented or settled this issue? If the answer is “yes” more often than “no,” sketch a small-scale Digital QA pilot, one workstation, two cameras, thirty days and measure the impact. Your data will tell you whether scaling makes sense.
