RFID gets pitched as the modern replacement for barcodes. It is not a replacement, and treating it as one leads to expensive disappointment. It solves a specific problem that barcodes cannot solve at all, and it is worse than a barcode at nearly everything else.
The one real difference
A barcode requires line of sight and one item at a time. Someone has to see it, point at it and scan it.
RFID does not require line of sight and can read many tags at once. Walk a trolley through a doorway and everything on it registers, including the items at the bottom of the stack.
That is the entire value proposition. If your problem is not “I need to count many things quickly without touching them”, RFID is not your answer.
Where it clearly pays
Stocktaking at volume. Retail apparel is the classic case: a full-store count that took a team a night can take one person an hour. When that count moves from twice a year to weekly, stock accuracy improves enough to change how the business operates.
Assets that move and must be found. Tools, test equipment, medical devices, returnable containers. Anything expensive that walks and gets searched for.
Automatic checkpoints. Reading a pallet as it passes a dock door, with nobody scanning anything, removes a labour step and an opportunity for error simultaneously.
Items where scanning is awkward. Things in sealed cases, in bins, or in positions where getting a scanner to the label is genuinely difficult.
Where it does not
Point of sale. One item, one scan, and a barcode already costs nothing because it is printed on the packaging anyway.
Low-value, high-volume goods. A tag costs real money per item. On low-margin fast-moving stock, that cost lands on every unit and never comes back.
Where you need certainty about a specific item. RFID is excellent at reading many tags and imperfect at guaranteeing it read all of them. If you need to confirm this exact item, a barcode scan gives an unambiguous yes.
The physics people discover late
Passive UHF RFID, which is what most inventory deployments use, has real physical limits that do not appear in the sales material.
Metal reflects and detunes. A standard tag stuck directly on metal often will not read at all. On-metal tags exist, cost more, and need specifying deliberately.
Liquid absorbs. Water absorbs the frequency band used. Tagging bottles, containers of liquid, or anything largely made of water behaves very differently to tagging cardboard.
Read zones are not tidy rectangles. Signals reflect off walls, racking and floors. A reader intended to cover one doorway will happily read tags in the next aisle, producing stock movements that never happened. Containing a read zone is a real piece of engineering.
Dense tag populations interfere. Reading twenty tags is easy. Reading two thousand on a full pallet, reliably, every time, is not.
The cost that gets missed
People compare the tag price to a printed barcode and stop there. The full cost is tags, readers, antennas, installation, site survey, integration into the system that will hold the data, and the process change to make anyone act on it.
That last item is where most RFID projects underdeliver. Reading stock more often produces more data about discrepancies. If nothing in the business changes as a result, you have bought an expensive way to be better informed about a problem you are not fixing.
How to approach RFID or barcode sensibly
Pilot in one area, with a site survey done first, on the actual materials and the actual racking. The behaviour of a tag on your product in your building is not predictable from a datasheet, and it is cheap to find out before committing.
Then be honest about what decision will change once the data exists.
Our AIDC and mobility work covers both technologies, and a fair share of that work ends with a recommendation to fix the barcode process rather than replace it.
Related reading
- Specifying rugged mobile devices for warehouse and field work — Consumer phones look cheaper than rugged mobile devices until you count replacements and downtime. What the ratings mean and which specifications matter.
- Receipt printing or label printing: buying the right thing — Direct thermal and thermal transfer label printing are not interchangeable. Choosing wrongly means labels that fade before the goods even arrive.
- A backup is not a backup until you have restored it — The failures that matter are not missing backups. A backup restore that has never been tested is a belief, not a safeguard. Find out before you need it.