You need to track every part through the line. The integrator suggests RFID. The salesperson says it’s the future. Six months later, the RFID readers are erroring on metal parts, the tags cost $3 each, and nobody trusts the data.

What barcode does well

A barcode is cheap, reliable, and works on almost any surface. It needs line-of-sight, but on a production line that’s easy to arrange: a fixed scanner above a conveyor reads the label as the part passes. Barcodes tolerate oil, dust, and vibration as long as the label is legible. They’re also passive: no battery, no maintenance.

The failure mode is obvious: if the label smudges or falls off, the scanner reads nothing. You can see this. The operator re-applies the label and moves on. Barcode failures are visible and fixable in seconds.

What RFID does well

RFID doesn’t need line-of-sight. A reader can read a tag through a box, through a bin, or from several feet away. On a tote that carries 50 parts, one reader inventory’s the whole tote in a second. You can’t do that with barcodes.

RFID also writes back. The tag stores data: part number, serial, time stamp, process history. As the part moves through each station, the reader updates the tag. The part carries its own history. Barcodes just point to a database record.

Where RFID falls apart

Metal. UHF RFID tags work poorly near metal because the metal reflects the signal and detunes the antenna. On a metal part, a metal bin, or a metal pallet, you need special on-metal tags, which cost more and read less reliably. Liquid causes similar problems.

Reading speed is another limit. A barcode scanner reads a label in milliseconds. An RFID reader inventory’s tags over 100 to 500 ms. On a high-speed conveyor, that delay matters. And if two tags are in the reader field at once, collision algorithms slow the read further.

The other trap is the infrastructure. RFID readers, antennas, gates, and wiring cost real money. A barcode scanner is $200. An RFID portal for a pallet is $2,000 to $5,000. If you’re tracking one part per tote on a low-speed line, a barcode scanner is cheaper and more reliable.

Hybrid is the usual answer

Most lines end up using both. Barcodes on individual parts for station-level tracking. RFID on totes, pallets, and racks for bulk inventory. The barcode tells you which part is at this station. The RFID tells you what’s in the tote without opening it.

Use the barcode as the authoritative serial number. Use RFID for location and bulk count. If the RFID says the tote has 12 parts but the barcode history says 10 were loaded, trust the barcode. RFID tells you where; barcode tells you what.

Tag placement

If you go RFID, put the tag where the reader can see it, not where it’s convenient. A tag on the bottom of a tote that rides on a metal roller conveyor is read through the metal. Put it on the side, at a consistent height, away from brackets and fasteners. Test the read range with the actual load, not an empty tote.

On-metal tags need a spacer between the tag and the metal surface. Without it, the read range drops to a few centimeters. This is the most common RFID integration mistake.

Bottom line

Barcode for parts, RFID for containers. Don’t put RFID on every part just because you can. The tag cost, the metal and liquid issues, and the reader infrastructure add up fast. Use barcode where line-of-sight is easy, and reserve RFID for the bulk-loading and inventory scenarios where it actually saves time.