RFID stands for Radio Frequency Identification. It uses radio waves to read a tag attached to an item without needing line of sight, so hundreds of tags inside a sealed carton can be read at once. A system has three parts: tags, readers, and software that interprets the reads.
That last part is the whole proposition. A barcode has to be found, aligned and scanned one at a time. An RFID reader can pick up hundreds of tags inside a sealed carton as it passes through a doorway. Where that capability matches the problem, RFID transforms the work. Where it does not, it is an expensive way to do what a printed label already does.
What RFID is
An RFID system has three parts.
The tag carries a microchip holding an identifier, and an antenna. It is attached to the item, the carton, the pallet or the returnable container. Tags come as adhesive labels, hard plastic housings, or units embedded into a product during manufacture.
The reader transmits a radio signal and listens for tags responding. Readers are handheld, mounted in doorways and dock portals, built into conveyors, or fitted to forklifts.
The software turns a stream of tag reads into something meaningful. This is the part most often underestimated. A reader at a dock door produces thousands of reads a minute, including tags on a nearby shelf and tags on a pallet passing in the other direction. Turning that into “this shipment left at this time” is a filtering problem, and it is where RFID projects succeed or fail.
How RFID works
A passive tag has no battery. The reader’s radio signal energises the tag’s antenna, which powers the chip just long enough to transmit its identifier back. That is why passive tags are cheap, thin, and last indefinitely — there is nothing to go flat. Most warehouse RFID is passive.
An active tag carries its own battery and broadcasts continuously or on a schedule. It reads at much greater range — tens of metres — and can carry sensors for temperature or shock. It also costs considerably more per tag, and the battery eventually dies. Active tags suit high-value assets, vehicles, and returnable containers where the tag is reused for years.
Semi-passive (or battery-assisted passive) tags sit between the two: a battery powers the chip but the reader still supplies the transmission energy. Longer range than passive, cheaper than active.
Frequency matters as much as tag type:
- Low frequency (LF, ~125–134 kHz). Short range, a few centimetres. Reads well near metal and liquid. Used in animal tagging and access control.
- High frequency (HF, 13.56 MHz). Range up to about a metre. This is the band NFC uses, which is why a phone can read an HF tag. Common in libraries, ticketing and payment.
- Ultra-high frequency (UHF, 860–960 MHz). Range of several metres, fast bulk reading. This is what supply chain and retail RFID means in practice.
UHF is the interesting one for a warehouse, and it comes with a physics problem worth knowing before you commit: UHF signals reflect off metal and are absorbed by water. A pallet of bottled drinks or a cage of metal parts will read far less reliably than a pallet of clothing. Specialist on-metal tags and careful reader placement help, but the constraint is real and no vendor demo will feature a pallet of liquid.
RFID vs barcodes: the key comparison
The differences that actually change how you work:
- Line of sight. Barcodes need it. RFID does not. This is the fundamental divide.
- Read rate. A barcode is one scan at a time. A UHF reader can read hundreds of tags in seconds.
- Range. Barcodes are read from centimetres to a couple of metres. Passive UHF reaches several metres; active tags reach tens.
- Cost per item. A printed barcode costs a fraction of a penny. A passive UHF tag costs meaningfully more, and an active tag far more again. At scale this is usually the deciding factor.
- Data. A barcode holds a fixed printed value. Many RFID tags can be written to as well as read, so status can be updated on the tag itself.
- Durability. A barcode fails when the label is torn, smeared or obscured. A tag inside a housing survives conditions that destroy print.
- Certainty. A barcode scan is a deliberate act with a clear result. RFID reads can be probabilistic — a tag might be missed, or an unintended tag read. Processes have to be designed for that.
That final point is the one that catches people. A picker scanning a barcode knows they scanned it. An RFID portal reporting 47 of 48 items gives you a discrepancy with no obvious cause, and reconciling it manually can cost more than the scanning saved.
Where RFID shines — and where it doesn’t
It earns its place when:
- You need to count many items at once. Reading a full pallet or cage in one pass rather than item by item is where the labour saving is real.
- Items are inside packaging. Verifying carton contents without opening them removes an entire step.
- Stocktakes are frequent and painful. Fashion retail is the clearest success case — a full shop count that took a night with barcodes takes an hour with a handheld RFID reader, which means counting weekly instead of quarterly, which drives inventory accuracy up sharply.
- Assets are high-value and reusable. Returnable containers, tools, equipment. The tag cost amortises across years.
- The environment destroys labels. A tag in a sealed housing keeps working where print does not.
It struggles when:
- Margins are thin and volumes high. Tag cost per unit against margin per unit is usually the whole argument.
- Products are metal or liquid. Achievable with specialist tags, but expect lower read rates and more engineering.
- One item is identified at one point. If a picker handles items individually anyway, a barcode does the job for a fraction of the cost.
- Nobody owns the data. RFID generates far more reads than barcode scanning. Without someone responsible for what those reads mean, you get an expensive source of noise.
What an RFID project actually involves
The tags are rarely the hard part.
Site survey. Radio behaves differently in every building. Racking, cladding, other equipment and the products themselves all affect read rates. A survey before purchase is cheaper than discovering a dead zone after installation.
Tag selection and placement. The same tag reads differently depending on where it sits on the carton. Getting this right usually takes testing with your own products rather than accepting a generic recommendation.
Reader placement and tuning. Too much power and you read tags you did not intend. Too little and you miss some. Portal geometry, antenna angle and power settings all need tuning against real traffic.
Integration. Your WMS needs to accept tag reads and resolve them to stock movements. If it cannot, you are buying a parallel system and a reconciliation job.
Process redesign. This is the step most often skipped. RFID only pays back if the process changes to use it. Reading a pallet in one pass and then still scanning each carton captures the cost and none of the benefit.
Exception handling. Decide in advance what happens when a portal reports 47 of 48. Who investigates, how, and how quickly. Without an answer, staff invent one, and it is usually to ignore the discrepancy.
A sensible sequence is to pilot on one process with a defined success measure — stocktake duration, dock accuracy, count frequency — rather than tagging everything and hoping the benefits appear. Most operations that get value from RFID run it alongside barcodes rather than replacing them, using each where it is stronger.
Standards: EPC and why they matter
An RFID tag holds a number. For that number to mean anything outside your own systems, it has to follow a standard.
The dominant one in supply chains is EPC — the Electronic Product Code, administered by GS1, the same organisation behind retail barcodes. An EPC identifies not just the product but the individual item. Where a barcode says “this is a size 10 blue trainer”, an EPC says “this is a specific size 10 blue trainer, distinguishable from the other four hundred in the building”.
That distinction changes what is possible. Serialised identity means you can trace one unit through a supply chain, prove which specific items were in a shipment, and confirm that the item a customer returned is the one you sent.
The relevant standards in practice:
- EPC Gen 2 (also ISO 18000-63) — the UHF air interface standard. Effectively universal for supply chain RFID, which means tags and readers from different vendors work together.
- SGTIN — a Serialised GTIN. It takes the GTIN from your barcode and adds a serial number, so barcode and RFID identify the same product without a translation layer.
- ISO 18000 series — the broader family covering other frequencies.
Two practical implications. First, if you are tagging goods that leave your building, use EPC rather than an internal numbering scheme — otherwise every trading partner needs a mapping table. Second, because SGTIN builds on GTIN, an operation that already has clean barcode data has done most of the hard work.
What it costs, honestly
RFID business cases fail more often on the unglamorous costs than on tag price.
Tags are the visible cost and the one vendors quote. Passive UHF tags are cheap at volume; specialist on-metal and hardened tags cost several times more.
Readers and antennas are a larger capital line than most people expect. A single dock-door portal needs multiple antennas, a reader, mounting and cabling. Multiply by the number of doors.
Installation and tuning is skilled work. Portal geometry and power settings need adjusting against real traffic, not a test pallet.
Integration is usually the biggest single line. Getting tag reads into a WMS in a form it can act on may mean middleware, custom development, or a WMS upgrade.
Process change and training costs time rather than money, and it is the cost most often left out entirely. Staff need to work differently for the benefit to appear.
Ongoing — tag supply, reader maintenance, and someone who owns read-rate performance. A portal quietly dropping from 99% to 94% will not announce itself.
Against that, the returns worth counting are labour hours removed, stocktake frequency increased, shrinkage reduced through better visibility, and errors avoided at dispatch. The strongest cases usually come from counting: an operation counting quarterly because counting is painful, moving to weekly because it no longer is, and finding inventory accuracy climbing as a result.
Pilot before committing. One process, one measurable outcome, real products in the real building.
Advantages and disadvantages of RFID
The advantages.
- No line of sight. Tags read through cardboard, plastic and around corners. A sealed carton can be identified without opening it, and a pallet can be verified without breaking it down.
- Bulk reading. Hundreds of tags read in a single pass. This is the property that changes what is operationally possible, because it turns counting from a task into a by-product of movement.
- Speed. A full stock count of a shop floor or a storage area in minutes rather than a day.
- Tags carry unique identities. Every item can be individually serialised, not just identified as a product type, which makes item-level traceability and anti-counterfeiting practical.
- Tags are rewritable. Data can be updated in place — a status, a location, a reading — where a printed barcode would need reprinting.
- Durability. Encapsulated tags survive washing, freezing, paint, heat and physical abuse that would destroy a label.
- It removes the operator from the transaction. A read happens because stock moved past a portal, not because somebody remembered to scan it. That eliminates a whole class of human error.
The disadvantages.
- Cost per tag. Orders of magnitude more than a printed barcode. At scale on low-value goods, this is usually the whole argument, and it is why RFID took hold in apparel and high-value retail long before general merchandise.
- Physics. Metal reflects and liquid absorbs radio energy. Tagging tinned food, drinks, tools or anything metallic needs specialist tags, careful placement and testing, and sometimes it simply does not work well.
- Over-reading. A reader that sees everything within range will happily count the adjacent aisle, the next dock door, or a pallet still on the vehicle. Controlling the read field is a real engineering task rather than a setting.
- Infrastructure. Readers, antennas, cabling, mounting and network. The tags are the visible cost and the fixed infrastructure is frequently larger.
- Integration effort. A stream of reads is not the same as a stock figure. Something has to interpret duplicate reads, direction of travel and stray reads before any of it reaches your inventory system.
- It does not fix bad process. RFID tells you what is where far faster. It does not tell you why the wrong thing was put there.
RFID, privacy and RFID-blocking
RFID has a public reputation slightly separate from its industrial use, and it is worth addressing directly because customers and staff ask about it.
The security concern. A tag that can be read without line of sight can, in principle, be read without the holder’s knowledge. Applied to retail goods, that raises the question of whether a tagged item leaving a shop can be tracked afterwards, or whether the contents of a bag can be read by someone standing nearby.
What is actually true. Passive tags of the sort used on retail goods have no power of their own and respond only when energised by a reader, typically within a few metres at best and often far less. They carry an identifier, not personal data — the identifier means something only to the system that issued it. And most retail deployments deactivate or discard tags at the point of sale precisely to close the question.
Where the concern is more reasonable. Contactless payment cards, passports and access cards use related technology at short range, and those do carry data worth protecting. This is the origin of RFID-blocking wallets and sleeves, which work by putting a conductive layer between the card and any reader — a Faraday cage in miniature. They are effective for what they do, and their relevance to a warehouse tagging cartons is essentially nil.
What this means operationally. If you tag consumer goods, decide in advance what happens to the tag at the point of sale, and be able to answer the question when a customer asks. If you tag staff access or use readers where people work, be explicit about what is being recorded and why. The technology is benign; opacity about how it is used is what creates the problem.
Where RFID is used in practice
Fashion and apparel retail is the clearest adoption story. Garments are non-metallic, non-liquid and tag well; margins carry the tag cost; and stock accuracy directly drives sales, because an item recorded as in stock but actually missing is a lost order. Retailers moving from quarterly to weekly counts have reported inventory accuracy moving from the eighties into the high nineties, which in turn supports ship-from-store and click-and-collect.
Warehousing and distribution uses it mainly at choke points rather than everywhere: dock-door portals confirming what left on which vehicle, and handheld readers for cycle counting. Tagging at pallet or carton level rather than item level keeps the tag cost proportionate.
Returnable transit items — pallets, cages, totes, kegs, gas cylinders. The tag lives on the asset for years, so cost per read is negligible. Knowing where your containers are, and who has had them longest, usually pays for itself on asset replacement alone.
Healthcare tracks surgical instruments, implants and high-value consumables where expiry and recall traceability matter and the cost of getting it wrong is severe.
Manufacturing uses it for work-in-progress tracking, confirming that an assembly passed each station with the right components.
Libraries were early adopters for the same reason as fashion retail: frequent full counts of many similar items, done by hand, was miserable work.
A pattern runs through all of these. RFID wins where counting many things is the job, where the items tolerate radio, and where someone will act on the resulting data. Where any of those three is missing, the project tends to deliver a dashboard nobody uses.
Where to go deeper
- Types of barcodes — the alternative, and still the right answer for most item-level identification.
- What is a barcode — how the cheaper option works.
- Barcode scanning for inventory — building scanning into a live process.
- What is a stocktake — the job RFID most often transforms.
- Cycle counting — the counting discipline RFID makes cheap enough to do often.
- What is a SKU — the identifier the tag ultimately resolves to.
A sensible starting point
If RFID looks like it might fit, the cheapest way to find out is a contained pilot rather than a strategy document.
Pick one process where counting is the bottleneck — usually stocktake or dispatch verification. Define a single number you expect to move, and record what it is today: hours per count, dock errors per week, count frequency. Tag one product family in your real building, with your real racking, and read it with the hardware you would actually buy. Run it for long enough to see the exceptions, because the first week always looks better than the fourth.
Then decide on the measured result rather than the demonstration. Vendors demonstrate on cardboard boxes of clothing in an empty room, and that is not a lie so much as an unrepresentative sample.
The one-line takeaway
RFID reads many tags at once without line of sight, which makes it powerful for bulk counting and sealed cartons — but the tag cost, the physics around metal and liquid, and the process redesign mean it only pays back where those specific strengths match your problem.
Frequently asked questions
What does RFID stand for? Radio Frequency Identification. A tag on the item responds to a reader’s radio signal with its identifier, without needing to be seen or aligned.
What is the difference between active and passive RFID? Passive tags have no battery and are powered by the reader’s signal — cheap, thin, indefinitely long-lived, with a range of a few metres. Active tags carry a battery, broadcast on their own, reach tens of metres, and cost considerably more.
How far away can RFID be read? Low frequency, a few centimetres. High frequency, around a metre. Passive UHF, typically several metres. Active tags, tens of metres. Real-world range is always lower than the specification once products and racking are in the way.
Does RFID work through metal and liquid? Not well by default. UHF signals reflect off metal and are absorbed by water, so both reduce read reliability. On-metal tags and careful placement improve it, but expect lower read rates than with cardboard or textiles.
Is RFID replacing barcodes? No. It has been described as imminent for two decades. Barcodes remain far cheaper per item and give a definite result for a single deliberate scan. Most operations that adopt RFID run both, using each where it is stronger.
How much do RFID tags cost? Passive UHF tags are the cheapest by a wide margin and get cheaper at volume; active tags cost many times more. The number that matters is tag cost against the margin on the item being tagged, which is why RFID reached fashion retail long before it reached grocery.
What is the difference between RFID and NFC? NFC is a subset of high-frequency RFID, designed for very short range — a few centimetres — and two-way communication with consumer devices. Any modern phone can read an NFC tag. Warehouse RFID is almost always UHF, which reads much further but cannot be read by a phone.
Do I need to replace my barcode system to use RFID? No, and most operations do not. The two run alongside each other, and because a serialised EPC is built on the same GTIN your barcodes already use, clean barcode data makes RFID adoption considerably easier rather than redundant.
Can RFID tags be reused? Yes, where the tag is in a durable housing on a returnable asset — containers, totes, equipment. Adhesive label tags on disposable packaging generally are not worth recovering.
Warehouse Wise is a free, independent library of guides on inventory management, warehousing and fulfilment, written for the people actually doing the work — new warehouse and inventory staff, small and growing retail and ecommerce businesses, and anyone who holds stock and wants to understand it properly. Every guide explains the term, shows a worked example with real numbers where it helps, and stops. We are independent: we do not sell software or equipment, we take no commission on referrals, and our guides are not ranked by who pays us.