How do you choose the RFID warehouse count? Efficiency-cost trade-offs
Let's get this whole thing off the record and then talk about technical options.
Before choosing the RFID option, you answer three questions: what's in the warehouse, how much is a miscalculation, and how high is the count? These questions determine whether you need RFID, not the other way around.
Disk metal, liquids, clothing, paper files, readers choose a completely different frequency. Metal surfaces block normal passive labels, and must use anti-metallic labels at a unit price that jumps from a few centimetres to $3 to $10; liquids are heavily absorbed into ultra-high frequency signals, and beverage warehouses and chemical stores are careful; clothing shoe-hatting is the most classic RFID scene, with tags placed on a laminate or wash sign, reading and writing distance of 3 to 8 metres, and several hundred items can be identified in one instance.
The frequency of the count is also a key variable. It's very different to take one and one a day, and it's completely different to be tolerant of the program. We met with a customer who was doing clothing distribution, and we started to do monthly counts with PDA, so it took hundreds of thousands of dollars a year for the manual cost to go back and assess the RFID fixed channel.
RFID scans against PDA. How much is the efficiency gap?
A skilled worker needs six to eight hours to count a midsize stockroom of 5,000 to 10,000 items by hand. A handheld PDA scanner cuts this to 1.5 to 2 hours. Fixed RFID portals or smart shelves can automate the whole process in the background, so the count itself takes practically no staff time.
This number seems exaggerated, but it is only if the labels and the readers match up. A common error zone is that RFID is "tweeted" -- in fact, without a light-pointing angle, labels are headed, power configuration, one-time recognition may be 70 to 80%, and the rest of them need to be read more difficult. We usually ask for a single batch of success at not less than 95%, otherwise the credibility of the program cannot be built.
The RFID is more valuable than the inventory itself. The garment industry uses the RFID portal to check for delivery and delivery, with a single-cargo pass recognition of over 98%, which takes two people 20 minutes to check a car's cargo, now sweeps + automatically read it for seconds.
This is the cost. Don't just read the label unit price.
The cost of the RFID storage project is shown to be four main items: labels, reading and writing equipment, antennas and wiring, intermediates and system integration.
Tags: Passive UHF tags cost CNY 0.50-2 each in volume. Specialty tags, including on-metal, rigid, and high-temperature tags, cost CNY 3-15 each. Active tags for tracking equipment with metal enclosures cost from several tens of yuan to over CNY 100 each. In a garment warehouse with 100,000 SKU units, tags alone cost CNY 50,000-200,000.
Reader/writer equipment: Fixed readers cost 2,000–8,000 yuan each, depending on the four- or eight-port configuration. RFID antennas cost 500–2,000 yuan each, and handheld PDAs cost 3,000–15,000 yuan. A receiving lane usually requires two to four antennas. Equipment costs for an entire warehouse can range from tens of thousands to hundreds of thousands of yuan.
Middle and Integrated: Many people underestimated the cost of this piece. The raw data from RFID readers are to be heavy, filtered, connected to SKU and storage positions, and connected with WMS, ERP. This part of the software development and commissioning often takes more time than hardware. A full RFID warehouse project, from hardware procurement to go online, with a budget below 300,000, clothing warehousing to go online, less than 300,000, clothing warehousing to do hardware to go up, and clothing warehousing to go up very hard.
Which warehouses are worth RFID and which aren't.
There are several typical scenarios for the warehouse that fits on RFID.
One is the SKU, the large, medium-value warehouse of a single piece, like clothing, shoes hats, book files, retail distribution. The cost of this scene is not very high but high in frequency, and the benefits of RFID mass reading can be translated directly into human savings.
The second is the parts and semi-finished swaths of manufacturing, especially sophisticated, medical devices, electronics. These are very valuable, one-time costs, and the precise positioning and protection of RFID can avoid direct economic loss.
The third is the electrical powerer's compliance warehouse, 3PL warehouses, which require high timeliness of inventory counts. SKU access is frequent, and traditional means of counting directly affect delivery efficiency, with RFID portal + smart shelf almost standarding.
There are several other things that are not appropriate: SKU has few small warehouses (less than a few hundred) and the bulk readability advantage of RFID is not available; pure metal raw materials dumps, with difficult labels, have poor recognition; very low budget and no continuous input programmes, but one-time inputs to incomplete solutions are more wasted.
A few pits that can easily be stepped on in the project.
The first pit is the reflection and shielding of UHF signals by metal shelves and metal commodities. We met with a customer who made a carbide, which was first designed to fit the antennas in the clothing warehouse, and the metal reflectors kept reading rates off, and finally added suction material, and lowered the antenna power, adjusting the antennas for polarization. The metal environment program needs to be designed individually, not directly using the parameters of the normal warehouse.
The second pitfall is tag orientation. Passive UHF tags are highly sensitive to alignment. They read farthest when tag and antenna polarization align and may not read at all when perpendicular. Applications where product orientation cannot be controlled, such as garments hung at different angles, are much more complex than fixed-orientation setups.
The third pit is multilabeled collisions and double reading. In a warehouse that is dense with SKU, dozens of labels are colliding at the same time, and RFID readers are reoriented by algorithms. If the middle is not well done, the same SKU could be read into two pieces, the inventory will go up. This Shanghai-based Cheekyuke is running small-scale pilot projects in several projects, running data for a week or two, then deciding on the type of reader and label to read and write, so as not to go up and fill the entire warehouse.
The fourth pit is connected to the existing WMS/ERP. The RFID system does not work independently, and eventually data will have to flow back into the warehouse management system to be closed. If the current WMS is an old one from a decade ago, the second development effort often exceeds expectations. This is clear at the selection stage with the IT department.
A simplified decision-making framework for the warehouse that prepares to be on RFID
First, to determine the human costs of the current inventory and miscalculation, including warehousing staff, time for correction, client claims; second, to estimate the total input of the RFID programme, including tagging, equipment, software, implementation, training and follow-up maintenance; And third, to calculate the return period for investment, general clothing, books, electrical-manufacturing projects that can be returned in 1-3 years, more of the manufacturing-retroactive projects are risk-averse value than direct cost savings.
If the payback period is longer than three years or hard to determine, do not rush into a full rollout. Pilot the system for three to six months in one receiving lane or one area, then use actual performance data to decide how far to expand. RFID involves more than an initial purchase. Tag losses, equipment maintenance, and software improvements all carry ongoing costs. It is much cheaper to uncover these issues during a pilot than to go back and fix them after rollout.

Ms. Lei
Mr. Pang