NFC product authentication cost has six components: the chip, its form factor, application on the line, serialisation and encoding, the verification platform, and integration with your systems. In most premium programmes the chip is the smallest of these. Application and integration decide the total, and chip class decides whether the programme is actually secure.
The first question in almost every NFC authentication conversation is about tag price. It’s a reasonable question with an unhelpful answer, because the tag is the most visible cost and usually the least important one. Programmes that stall rarely stall because the chip was too expensive. They stall because nobody budgeted for changing a filling line, because a co-packer was never asked whether it could apply a tag at speed, or because the platform fee was built on a scan volume nobody could forecast.
This guide is for the people who have to defend the budget: the Head of Supply Chain who will own the line changes, the Head of Digital who will own the integrations, and the brand protection lead who needs the programme to actually stop something. It breaks the cost of NFC product authentication into the lines that make it up, explains which ones move the total, and gives you a model to compare vendor quotes on the same basis.
We don’t publish a price list, and we’d be wary of anyone who does without seeing your product. A tag in a 70 cl spirits closure, a tag sewn into a leather lining and a label on a fragrance carton are different products with different tooling, yields and failure modes. What we can do is show you how the money is spent.
Why the tag price is the wrong place to start
Three numbers that matter more than tag price
- Encoding yield: Every tag that fails during encoding or commissioning is a unit you paid for that can’t be verified. On a well-run line the loss is small. On a line that was never tested, with a closure that shields the antenna, it can be large enough to change the economics.
- Line impact:If applying the tag slows a line, the cost is measured in lost capacity, not cents per unit. Programmes that put tags into components upstream (the closure, the label, the lining) before they reach the main line usually avoid this.
- Coverage: A programme that protects 20% of a range protects 20% of a range. Counterfeiters and diverters move to the unprotected SKUs. Cost per unit is only meaningful alongside the share of the range it covers.
The six cost lines in an NFC authentication programme
Where NFC authentication budgets actually go
| Cost line | What drives it | What makes it swing | Usually owned by |
|---|---|---|---|
| Chip | Chip class (static UID vs cryptographic vs tamper-detecting) | Volume, security level, memory | Brand protection |
| Form factor | Inlay, label, cap, closure, sewn-in, embedded | Substrate (metal, liquid), custom tooling, finish | Packaging |
| Application | Where and how the tag is fixed to the product | Manual vs in-line, co-packer changes, line speed | Operations |
| Serialisation and encoding | Writing keys and identifiers, registering each unit | Yield, key management, link to batch data | Operations / IT |
| Platform | Verification, event history, dashboards, consumer experience | Volume of units and scans, modules used | Digital / brand protection |
| Integration | ERP, MES, WMS, CRM, e-commerce connections | Number of systems, data ownership, in-house capacity | IT |
Two things stand out. First, four of the six lines sit outside the team that usually sponsors the programme. Second, the lines that swing most (application and integration) are one-off or step costs, while the chip is a per-unit cost. That changes how the programme scales: the second million units are far cheaper per unit than the first.
One-off, step and per-unit costs
It helps to sort the six lines by how they behave over time. Per-unit costs (chip, form factor, application labour) scale linearly with volume. Step costs (tooling, line integration) arrive each time you add a new line, site or packaging format, then stay flat. One-off costs (systems integration, key management set-up, consumer experience design) are paid once and reused for every subsequent product.
| Behaviour | Lines | What it means for scaling |
|---|---|---|
| Per-unit | Chip, form factor, application labour | Volume discounts help, but this line never disappears |
| Step | Tooling, line integration, new co-packers | Adding a new format costs more than adding volume to an existing one |
| One-off | Systems integration, key management, experience design | Should be amortised across the whole range, not charged to the pilot |
The practical consequence: a pilot always looks expensive per unit, because it carries one-off costs on a small volume. Judge a pilot on whether its one-off work is reusable, not on its unit cost.
Chip class decides security before it decides price
| Chip class | What a scan proves | Can it be cloned? | Relative unit cost | Fit |
|---|---|---|---|---|
| Static UID | That a tag with this ID exists | Yes. The UID and URL can be copied to a blank tag | Lowest | Engagement only, not authentication |
| Cryptographic (dynamic message) | That this specific chip holds a secret key, at this scan | Not practically. Each tap generates a new signed message | Medium | Authentication for most premium goods |
| Cryptographic + tamper loop | The above, plus whether the closure has been opened | Not practically | Highest | Spirits, fragrance, anything refillable |
A cryptographic tag (NXP’s NTAG 424 DNA is a common example) holds an AES key that never leaves the chip. On each tap it produces a URL containing a counter and a message authentication code computed with that key. The platform recomputes the code server-side. A copied URL fails because the counter has already been seen.
GET /v?uid=04A23B1C6F7380&ctr=000127&cmac=8F3A1C0D9B44E2A7
-> recompute CMAC with key for UID 04A23B1C6F7380
-> check ctr 127 > last seen ctr 126
-> result: AUTHENTIC, first scan in this marketWhy the tamper loop changes the business case for spirits and fragrance
For sealed products, the main fraud isn’t always a full counterfeit. It’s a genuine container with the wrong contents: a premium bottle collected after use, refilled with something cheaper and sold again in a bar or a market stall. A cryptographic tag alone will happily confirm that the bottle is genuine, because it is. A tamper-loop tag adds a conductive loop across the closure. When the seal is broken, the loop breaks, and every scan afterwards reports the bottle as opened.
That one extra bit of state turns authentication into evidence of refill fraud. It also costs more per unit and constrains the closure design, so it belongs on products where refill is the real threat, not on everything.
A cheap tag that can be cloned costs you the whole programme the first time someone copies it.
Form factor and application: the costs that live on the line
The same chip can arrive as a wet inlay, a paper or PET label, a moulded closure insert, a woven label, a hard tag or a component embedded in a product. The choice is driven by the product, not the chip.
| Product situation | Typical form factor | Cost driver |
|---|---|---|
| Glass bottle, premium closure | Tag inside the closure or capsule, often with tamper loop | Closure tooling and supplier integration |
| Leather goods, textiles | Woven or sewn-in label, or embedded component | Manual application in production; durability through use |
| Metal or liquid-filled product | On-metal or ferrite-backed inlay | Specialised inlay, tuning for read range |
| Carton or rigid box | Label applied in-line | Label applicator, line speed, carton recycling |
| Watches, jewellery | Micro tag in case, box or certificate card | Integration with product design; often on accompanying item |
Application is where operations teams earn or lose the programme. Three questions decide it: can the tag be applied by a supplier upstream rather than on your line; can encoding and commissioning happen at line speed; and who is accountable when a batch of units leaves without working tags. Selinko’s ILM Station hardware exists for the second question, encoding and commissioning items on the line, but the first and third are organisational and need answering before any equipment is ordered.
A cost model you can fill in
units_in_market = units_produced * encoding_yield
per_unit = chip + form_factor + application_labour
one_off = tooling + line_integration + systems_integration
recurring = platform_fee + support
cost_per_auth_unit = (units_produced * per_unit + one_off + recurring)
/ units_in_market
Run it for year one and year three. If year three isn’t dramatically cheaper per unit, the one-off costs are being charged as recurring, or the programme isn’t designed to scale beyond the pilot SKU.
Worked example: how the model behaves
Take two programmes with identical tag prices. Programme A applies tags by hand at the end of the line, on one SKU, with no ERP link. Programme B embeds tags in the closure at the supplier, rolls out across a range, and connects shipment data from day one. Programme A has a lower year-one invoice. Programme B has a lower cost per authenticated unit by year two, because its application cost is near zero on the main line, its integration is shared across every SKU, and its scans can detect diversion as well as counterfeits.
The model doesn’t tell you which programme to buy. It makes the trade-off visible so the decision isn’t made on the tag line alone.
How platform pricing models change the total
Platform fees are usually priced on one of four bases, and each one behaves differently as the programme grows.
| Pricing basis | Predictable? | Risk | Works when |
|---|---|---|---|
| Per unit tagged | Yes, follows production | Pays for units that are never scanned | Coverage across a range is the goal |
| Per scan | No, follows consumer behaviour | Penalises successful engagement | Scans are rare and high-value (e.g. resale checks) |
| Per module or feature | Yes | Paying for features you don’t use | Clear, stable use cases |
| Flat platform fee | Yes | Over- or under-sized as volume changes | Early stage or fixed scope |
Per-scan pricing deserves particular scrutiny. Scans are where the programme generates traceability data and consumer relationships. A pricing model that makes each additional scan cost money creates an incentive to keep engagement low, which is the opposite of what the programme is for.
What the money buys: three programmes, three cost profiles
| Scenario | Threat | Mechanism | Outcome |
|---|---|---|---|
| Refill fraud on premium spirits | Genuine empty bottles are bought back, refilled and resold in on-trade venues. | A tamper-loop tag in the closure reports ‘opened’ once the seal breaks. A scan of an opened bottle presented as sealed is flagged. | Cost weighted towards chip class and closure tooling. Integration light. |
| Resale authentication for leather goods | Counterfeits passed off as pre-owned originals on resale platforms. | A cryptographic tag embedded at manufacture plus a transferable ownership record. Resellers verify chip and history. | Cost weighted towards form factor (embedding) and platform. Chip mid-range. |
| Grey market diversion in beauty | Stock allocated to one region surfaces at discount in another. | Each unit is serialised against its intended market. Consumer and retailer scans reveal where it actually is. | Cost weighted towards serialisation and integration with distributor allocation data. |
Sizing a pilot that tells you something
Pilots fail for two reasons: they’re too small to produce meaningful scan data, or they’re designed so that nothing they build is reusable at scale. A useful pilot answers specific questions.
- Choose a SKU with a real threatA product with known counterfeiting, refill or diversion exposure, so results mean something.
- Run it on the production line you’ll use at scaleNot a manual workaround that hides application problems.
- Connect intended market from the startEven one ERP field turns scans into diversion evidence.
- Give consumers a reason to scanWithout content or service, scan volumes will be too low to learn from.
- Measure yield, line impact and scan behaviourThose three numbers decide the scale-up budget.
At the end of a well-designed pilot you should know your encoding yield, your line impact, the cost of adding the next SKU, and what the field scan data shows. That’s enough to build the year-three model with real numbers rather than vendor assumptions.
Questions to ask
| Question | Description | How Selinko answers |
|---|---|---|
| Which chip, exactly? | ‘Secure NFC’ is a marketing phrase, not a specification. The chip part number tells you whether a scan proves anything. | We name the chip in every proposal and pick its class from your threat: cryptographic NFC for counterfeits, tamper-loop for refill fraud. Because Selinko also works with secure QR, RFID, ACM-secured barcodes and Diamond Dust, we have no reason to push NFC where another carrier fits your product better. |
| What is the encoding yield assumption? | Every tag that fails on the line is a unit you paid for that can’t be verified, and you usually find out from a consumer. | Selinko’s ILM Station encodes and commissions each item on the line, so a failed tag is caught before the product ships, not after. Only commissioned items verify as genuine, which also means rejected or stolen labels can’t be passed off later. |
| Who applies the tag, and at what line speed? | If the answer is ‘your co-packer’, the real question is who has done it before, on a line like yours. | We design application around your production setup, ideally upstream in the closure, label or lining so your main line doesn’t slow down. As part of TOPPAN, we have printing, labelling and packaging expertise in the same group, and we’ve done this across 40+ brand programmes. |
| Where do the keys live? | Whoever controls key generation, injection and storage controls whether your programme can be trusted. | Keys are generated and managed within Selinko’s platform and written to chips in a controlled encoding process. They’re never handed to co-packers as files, and each chip has its own key, so one compromised tag can’t compromise a product line. |
| What does the platform fee scale with? | Units, scans, modules or markets: each model behaves very differently by year three, and per-scan pricing penalises the engagement you want. | We model your cost at year one and year three before you sign, using your volumes. Scans are where your traceability and consumer data come from, so we don’t design pricing that makes success expensive. |
| What happens to the scan data? | If scans can’t feed your systems, you’re paying for authentication only. | SelinkoTrack compares every scan with the item’s intended market, so the same tags that stop counterfeits also reveal grey market diversion, down to the shipment. Data flows to your ERP, CRM and brand protection tools through our integration options, and it stays yours. |
For how the chip-level security works in detail, see How Secure NFC Tags Work. For sector specifics, NFC Product Authentication for Luxury Brands and NFC Authentication for Wine & Spirits: How It Works and Why It Matters cover application constraints that move cost.
Scope your programme cost on real numbers
FAQs
How much does an NFC authentication tag cost?
It depends on chip class, volume and form factor. Static UID tags are cheapest but can be cloned. Cryptographic tags cost more per unit and are the minimum for real authentication. Tamper-detecting tags cost most. In premium programmes the tag is usually a small share of total cost compared with application and integration.
Is NFC authentication more expensive than QR codes?
Per unit, yes: a printed QR code costs almost nothing to add. But a static QR code can be photocopied, so it does not authenticate. The relevant comparison is NFC against secured alternatives such as protected QR or barcodes with a security feature, and that depends on the threat you need to stop.
What is the biggest hidden cost in an NFC programme?
Application on the production line. Adding a tag can affect line speed, involve co-packers and require tooling. If the quote does not cover who applies the tag and at what yield, that cost lands on operations later.
Can we start with a pilot to control cost?
Yes. A pilot on one SKU or one market tests application, encoding yield and scan behaviour before committing. Design it so the integration work is reusable, otherwise you pay for it twice.
Discuss NFC feasibility and implementation for your project
