Scanning replaces "the driver says they delivered it" with "the system recorded that this specific item was delivered at this place and time". That difference is worth money in three areas: inventory accuracy, dispute resolution and load verification.
It also costs seconds per item, and if you apply it everywhere without thought, the seconds add up faster than the benefits.
Where scanning pays
| Point | Value | Cost |
|---|---|---|
| Pick verification | Prevents wrong items reaching the van | Warehouse seconds |
| Load-out onto vehicle | Confirms the load matches the plan | Seconds per item |
| Delivery at the stop | Confirms what was actually handed over | Seconds per item |
| Return to depot | Reconciles unsold and returned stock | Seconds per item |
| Asset and container tracking | Tracks returnables per customer | Seconds per asset |
The highest return is usually at load-out and at the stop for high-value or dispute-prone items. Scanning every low-value case at every point is where operations create a productivity problem and then blame the technology.
A pragmatic default: scan at load-out for everything, scan at the stop for high-value, serialised or frequently disputed items, and use quantity confirmation for the rest.
Time ten scans in real conditions — a dirty label on a wrapped pallet in dim light, not a fresh label on a desk.
Hardware options
Camera-based scanning on a phone or tablet. No extra hardware, adequate for occasional scanning, slower and less reliable in poor light or on damaged labels.
Bluetooth ring or wearable scanners. Fast, hands-free, and a large productivity difference in high-volume scanning. Common in warehouses, increasingly used on vehicles.
Rugged handheld with integrated laser or imager. Fastest and most reliable, purpose-built, but an additional device to carry, charge and support.
Sled attachments that turn a phone into a proper scanner. A reasonable compromise where the phone is already the primary device.
For anything above a few dozen scans per day, dedicated scanning hardware pays for itself in time saved and frustration avoided. Camera scanning is adequate for exception verification, not for volume work.
Symbologies
- 1D linear codes (Code 128, EAN, UPC) — universal, compact, and hold limited data.
- 2D codes (QR, DataMatrix) — hold much more data, readable when partially damaged, readable from any angle. Better for asset labelling and for encoding multiple fields.
- GS1 standards — structured data (product, batch, expiry, serial) in a single code. Essential in food and pharmaceutical distribution, and increasingly expected by retail customers.
- RFID — no line of sight, bulk reading of many tags at once. Excellent for pallet and container reading at gates; higher cost per tag and more infrastructure.
If your customers are large retailers or you handle regulated goods, the symbology is often dictated to you. Confirm requirements before designing labels.
Label quality: the invisible failure
Most scanning failures are label failures.
- Thermal labels fade with heat and sunlight, and freezer condensation destroys adhesion
- Labels wrapped in stretch film distort under the wrap
- Printers with worn heads produce codes that scan intermittently
- Labels placed over seams or on curved surfaces fail regularly
- Low print contrast on recycled packaging
Practical controls: a print quality check on a sample per shift, a defined label placement standard, printer head replacement on schedule, and label materials appropriate for the environment. A scanning programme that never audits label quality will blame the scanners.
Introducing scanning without a productivity collapse
- Start at one point, usually load-out, and stabilise before extending.
- Scan high-value first. Prove the benefit where it is largest.
- Design the exception path. When a code will not scan, the driver needs a fast, controlled manual entry with a reason code — otherwise the whole process stalls at the first damaged label.
- Measure scan failure rate by location, product and printer. Failures cluster, and the clusters are fixable.
- Do not require scanning for items where quantity confirmation is sufficient.
- Train on the failure cases, not the happy path.
Questions readers send us
Does scanning slow down deliveries? Per item, yes — a second or two with good hardware, longer with camera scanning. Whether that is worth it depends on the value of the accuracy gained. For high-value or dispute-prone items it clearly is; for low-value bulk cases it often is not.
What do we do when a barcode will not scan? Provide a controlled manual entry with a reason code, and report on manual entry rates. Blocking the workflow entirely causes drivers to abandon scanning; allowing unrestricted manual entry defeats the purpose. Track the rate and fix the underlying label problems.
Is RFID worth it instead of barcodes? Where you need to read many items at once without line of sight — pallet-level reading at a gate, returnable asset pools — RFID is transformative. For item-level scanning at a delivery, barcodes remain cheaper and entirely adequate.
Should we scan every item at every stop? Rarely necessary. Match the scanning point to the risk: serialised and high-value items at the stop, bulk cases at load-out with quantity confirmation at delivery. Blanket scanning policies are usually a sign that nobody analysed where the errors actually occur.
How do we handle customer-supplied labels? Confirm their symbology, data structure and placement standards during onboarding, and test scanning before the first delivery. Large retail customers frequently mandate specific formats, and a scanning failure at a retail dock costs far more than the label.
