RFID Wristbands, Loyalty Cards, and Metal Cards: The 9-Point Spec Checklist I Use Before Approving a Run

Who this checklist is for (and who should skip it)

I review custom print and specialty runs before they ship: rfid wristbands, retail loyalty cards, rfid smart cards, ultra high frequency rfid tags, custom badge reels, custom metal cards. Over four years on the compliance side, most of the reprints I've had to authorize trace back to one of nine things that never made it onto the spec sheet.

Here's the list, roughly in the order the decisions need to be made. Nine steps.

One caveat up front. If you're ordering 250 stock loyalty cards with no inlay and no variable data, steps 3, 4, 6 and 7 largely don't apply to you. Buy the stock item and move on. Everything past that line involves a component you cannot change once production starts.

Step 1: Decide whether you need RFID at all — and which frequency

“RFID” is several different technologies wearing one name. They are not interchangeable. The wrong pick doesn't degrade gracefully; it just doesn't work.

• HF / NFC, 13.56 MHz — reads at a few centimeters. The right choice for a retail loyalty card or an RFID smart card you want customers to tap deliberately with a phone.
• UHF (ultra high frequency RFID tags), 860–960 MHz — reads at meters, and reads many tags at once. The right choice for RFID wristbands at a gate or an inventory dock.
• LF, 125 kHz — short range, tolerant of metal and water. Niche, but it's the one that survives a wet pocket.

Frequency. Standard. Inlay position. In that order. Everything else on the spec sheet depends on those three.

Honest limitation: the mistake I see most often is UHF specified for a wallet card. You don't want a loyalty card that reads when someone walks past a doorway. You want a tap. That's HF.

Step 2: Put the frequency and the standard on the spec, not the word “RFID”

Write: “UHF, EPC Class 1 Gen 2 (ISO/IEC 18000-6C), 860–960 MHz.” Or “ISO/IEC 14443-A, 13.56 MHz, NTAG 213 or equivalent.”

Why it matters: vendors hear “RFID” and ship whatever is already in their bin. Two suppliers can both be technically correct and still deliver inlays that behave completely differently in your reader. This is the step most buyers skip, and it's the cheapest one to get right.

Step 3: Fix the inlay position before anyone touches the artwork

The antenna is not movable later. A 3 mm shift can cut read range in half, and a foil stamp or a metal logo plate laid over the antenna will kill it outright.

Put it in writing: inlay centered, X mm from the left edge, minimum 5 mm clearance between the antenna footprint and any metallic or dense ink element.

I assumed “same specifications” meant identical results across vendors. Didn't verify. Turned out each vendor had a slightly different reading of “centered,” and two of them placed the inlay under the logo — which is exactly where the metallic ink was. We caught it at the sample stage. That time.

Step 4: Specify construction, not adjectives

“Premium feel” is not a specification. Write the actual stack-up:

• Retail loyalty cards / RFID smart cards: 30 mil PVC, matte laminate both sides, 4-color process, spot UV on the logo only, inlay per step 3.
• Custom metal cards: stainless 0.8 mm, brushed finish, laser-engraved on both faces. If RFID is required, specify a ferrite layer between the inlay and the metal. Metal detunes an RFID antenna (i.e., shifts its resonance out of the readable band), and without shielding you get a card that reads at 5 mm or not at all.
• Custom badge reels: cord length, reel diameter, attachment type (swivel clip vs. bulldog clip vs. belt clip), and where the print goes — face plate only or full wrap. Face-plate-only print is cheaper and usually enough.
• RFID wristbands: substrate (silicone, woven, Tyvek), closure type, single-use vs. reusable, and whether the tag is laminated inside or applied to the surface. Surface-applied tags peel. That's it.

Step 5: Give tolerances, not just targets

A target is what you want. A tolerance is what you'll accept. Without the second, every deviation becomes an argument.

Reasonable numbers to put in the spec: trim ±0.5 mm, color within 3 Delta E of the approved target, card thickness ±5%, and a defined tolerance on any punched hole or badge slot. A slot that's off by 2 mm makes every card hang crooked in the badge reel, and nobody notices until 5,000 of them are sitting in a warehouse.

Step 6: Approve a physical sample, not a PDF

The most expensive assumption in this whole process is that a proof represents the final product. It does not represent substrate, laminate, metallic ink, foil, or finish.

Ask for a production-intent sample using the actual inlay, the actual laminate, and the actual finish. Then abuse it. Bend it. Drop it. Sit on it in a wallet for a week. Read it with the exact reader, at the exact distance you'll use in the field.

So glad we insisted on a physical sample for a custom metal card run. We were one signature away from approving a “champagne gold” finish from a screen, which would have meant 6,000 cards the color of a penny.

I'm not 100% sure of the exact split, but my rough guess is that two-thirds of the RFID failures I've seen were visible in a physical sample and completely invisible on a screen.

Step 7: Test the data, not just the card

Encoding is usually the printer's job. Verifying it is still your risk.

Read the tag with the reader that will be in the field, at the distance it will be in the field, on the phone model your customers actually own. If the card is serialized, confirm the sequence has no gaps and no duplicates before the run is packed. Ask for the encoded value list as a CSV shipped alongside the cartons — that one line in the spec has saved me more hours than any other.

Step 8: Assign ownership of variable data and mail prep

If loyalty cards are being mailed, they fall under USPS rules, not just print rules. Per USPS Business Mail 101 (pe.usps.com), a large envelope or flat runs from 6.125" × 11.5" up to 12" × 15", with a maximum thickness of 0.75". If your card is going into a window envelope, the window position and the card's clear zone belong in the spec — not in a follow-up email three days before the mail date.

Also settle here: who owns the variable data file, who does the matchback, and who pays if the list is wrong.

Step 9: Lock packaging, count, and delivery before the run starts

• Bulk packed or individually poly-bagged?
• Quantity tolerance. ±2% is standard in commercial print; on 1,000 pieces that's 20 either way. If you need exactly 1,000, say so on the RFQ — it changes the quote.
• Sequential or random packing.
• Drop-ship address, receiving hours, and any appointment requirement.

None of this is glamorous. All of it shows up as a delay if it's decided late.

Common mistakes that still get through

1. Approving a PDF as a proof. Step 6 covers why this one is expensive.
2. Letting the card design dictate the inlay position. It works the other way around.
3. Assuming “RFID” means one thing.
4. Skipping the real-reader test.
5. Never asking about quantity tolerance.
6. Environmental claims. Per FTC Green Guides (16 CFR Part 260), a “recyclable” claim has to be substantiated — a PVC card is not recyclable in most curbside programs, while most metal cards are. Don't let “eco-friendly loyalty card” language onto a PVC spec sheet.
7. Rush fees. Next-business-day turnaround typically runs +50–100% over standard pricing; 2–3 business days usually runs +25–50%, based on major online printer fee structures as of 2025. Worth it for a deadline-critical launch, rarely worth it for a trade show three weeks out.

On cost, for reference: setup fees in commercial printing generally include die cutting at $50–200 depending on complexity and custom Pantone matching at $25–75 per color, though many online printers now fold setup into the quoted price. Base card pricing as of January 2025 — 500 cards, 14pt cardstock, double-sided, standard turnaround — sits roughly in the $20–35 range at the budget tier and $60–120 at the premium tier. Expect anything with an inlay or a metal substrate to land well above that band. Verify current rates before you budget.

When to ignore this checklist

This is overkill for a stock order. If you need 250 plain loyalty cards with a barcode and a logo, steps 3 through 7 won't earn back the time you spend on them.

It is roughly the minimum for anything with an RFID inlay, a metal substrate, a ferrite layer, or a serialized encode. Those are the projects where a half-hour on the spec sheet is cheaper than the reprint.

And the one I'd underline twice: if your reader infrastructure can't handle the frequency you picked, nothing else on this list matters. Check that before you check anything else.