Stadiums

A ticketing platform sells the seat. An access control system decides whether the person standing at gate 7 gets through the turnstile. Those are usually two different products from two different suppliers, and they have to agree with each other tens of thousands of times inside a two-hour window before kick-off.
Integration is the contract between them. It is also where most entry problems actually live: the ticket that looked valid in the app and was refused at the reader, the transferred ticket that admitted two people, the lane that stalls because every tap waits on a round trip to a server sitting behind a congested network.
This covers how that link is built, which credentials survive event day, what the law now asks of a ticketed event, and the questions worth asking before a contract is signed.
What is a stadium ticketing system with access control integration?
It is a ticketing platform whose issued credentials are checked and redeemed by the venue’s entry hardware, so that a scan at the turnstile is settled against the same record that sold the seat.
Three jobs are involved, and they are often confused with each other. Issuance creates the credential and binds it to a seat, a price class and a buyer. Validation decides, at the reader, whether a presented credential is genuine and admissible right now. Redemption records that it has been used, so it cannot be used again somewhere else.
The third one is the job venues underbuy. A credential can be perfectly genuine and still be a duplicate. If two readers at opposite ends of the stadium do not share redemption state, both of them will say yes.
Which credential should the ticket actually be?
The credential is whatever the reader checks: a static barcode, a rotating code, an NFC pass, a payment card token, or a biometric template. Each choice moves the risk to a different place rather than removing it.
| Credential | Reader checks | Where the risk moves |
|---|---|---|
| Static QR code | Barcode against a local allowlist | Screenshots, resale, duplicate presentation |
| Rotating barcode | Short-lived code generated in the app | Device clocks, battery, app open |
| NFC wallet pass | Pass held on the device, tapped | Reader retrofit cost, phone dependency |
| Payment card as token | Card token captured at first tap | Same card needed on every visit |
| Biometric verification | Enrolled template matched one to one | Consent, enrolment, alternative lane |
QR remains the default because it is cheap and universal: the symbology is a published standard, currently ISO/IEC 18004:2024, the fourth edition, published in August 2024. Any camera can read it, which is exactly the weakness. A static barcode is a bearer instrument, and a bearer instrument can be photographed.
Rotating barcodes fix the photograph problem by making the code expire. They introduce a quieter one. The code is time-derived, so a phone whose clock has drifted presents a code the server considers invalid, and the fan in front of you is told their real ticket is fake. Validation windows and server-sourced time are not optional details here.
NFC passes move the credential into a secure element and the tap into the contactless standards, with the air interface specified in ISO/IEC 14443-2:2020. They are harder to copy and faster at the gate. They also need readers you may not own yet.
How do ticketing and access control systems talk to each other?
Through one of four patterns: a file export, a live lookup on every scan, a cached list with delta sync, or a signed credential the reader can verify entirely on its own.
| Pattern | How a scan is decided | Breaks when |
|---|---|---|
| Scheduled file export | Access control holds the full ticket list | Late sales, upgrades, re-issues |
| Live API call per scan | Reader asks the ticketing system each tap | Network at the gate degrades |
| Cached list plus delta sync | Local list refreshed every few minutes | A void lands inside the sync gap |
| Signed offline credential | Reader verifies a signature locally | Revocation needs a separate channel |
The file export is still common and it is the one that produces the angriest queue. Tickets sold after the export, seats upgraded at the box office and credentials re-issued after a transfer do not exist as far as the reader is concerned.
Live lookups are correct and fragile. Correctness is easy to demonstrate in a half-empty building in March. The test is forty thousand phones all fighting for the same spectrum at 19:00, at which point a 300 millisecond lookup becomes a three second lookup and a lane that was clearing fans comfortably becomes a bottleneck with a crowd behind it.
Signed credentials are the pattern that holds up. The reader checks a signature against a public key, offline, in milliseconds, and never asks anyone’s permission. The cost is honest and should be stated: revocation is no longer implicit. A voided or transferred ticket stays valid until something actively tells the reader otherwise, so the revocation feed becomes a first-class piece of the integration rather than an afterthought.
What goes wrong at the gate, and why?
Nearly every entry failure traces back to three causes: stale data at the reader, redemption state that is not shared between readers, or a validation path that depends on the network at the exact moment the network is worst.
| Failure | Cause | What removes it |
|---|---|---|
| Duplicate entry admitted | Readers do not share redemption state | One central redemption ledger |
| Valid ticket refused | Stale cached allowlist at the reader | Delta sync under one minute |
| Rotating code rejected | Device clock drift | Wider window, server-sourced time |
| Lane stalls at peak | Validation round trip over loaded Wi-Fi | Local signature verification |
| Transferred ticket bounces | Previous barcode still cached | Pushed revocation, not nightly export |
Notice that only one of these five is a fraud problem. The rest are integration problems that look like fraud problems to the person being turned away, and they are what staff spend the pre-match hour resolving by hand.
There is a related measurement point worth building in from the start. If redemption is centralised, you get an accurate live count of who is inside the building, by gate and by minute. If it is not, you get five partial counts that do not reconcile, which is the wrong state to be in when a safety officer asks a direct question.
What does the law require of a ticketed event?
More than it did two years ago. A ticketed event is now a regulated object in several jurisdictions at once: UK counter-terrorism duties, US accessibility rules covering both the seats and the app that sells them, and EU limits on biometric entry.
| Obligation | What it means at the gate |
|---|---|
| Martyn’s Law, enhanced tier | Ticketed events of 800 or more carry protection duties |
| ADA ticket sales rule | Accessible seats sold on the same terms and channels |
| ADA Title II web rule | Public-entity ticketing apps meet WCAG 2.1 Level AA |
| GDPR Article 9 | Biometric entry needs a lawful basis and an alternative |
| EU AI Act Article 5 | Live remote biometric identification is tightly restricted |
In the UK, the Terrorism (Protection of Premises) Act 2025 received Royal Assent on 3 April 2025. The definition that matters here is the qualifying event. Section 3 of the Act sets its conditions jointly: the public has access to the event, 800 or more people may be present at the same time, and access is by express permission, with tickets, passes or membership checked at entry. Ticket checking is part of what pulls an event into scope, and a qualifying event carries the same requirements as enhanced tier premises. The Act’s commencement provision leaves the start date to regulations, so the duties are not live yet.
In the United States, accessible seating is governed by 28 CFR 36.302(f). The Department of Justice’s ticket sales guidance sets out what that means for a ticketing system: accessible seats are sold during the same hours, through the same methods of purchase and the same sales outlets, at the same stages of sale including presales, at a price no higher than other tickets in the same seating section, and a purchaser may buy up to three additional companion seats in the same row. An inventory model that treats accessible seats as a phone-only exception does not satisfy that.
If the venue is owned or operated by a state or local government, including a public university or a municipal arena, the ticketing website and app are also in scope of the ADA Title II web rule, which adopts WCAG 2.1 Level A and AA. The deadlines moved: an interim final rule published on 20 April 2026 extended compliance to 26 April 2027 for public entities serving 50,000 or more people, and 26 April 2028 for smaller entities and special district governments. The DOJ’s own fact sheet on the rule carries the current dates.
Should facial recognition be part of entry?
Sometimes, for a small enrolled group such as season ticket holders or accredited staff, and only with a non-biometric lane of equal speed running beside it.
The distinction that decides the legal question is verification against identification. Verification is one to one: a person claims to be the holder of a specific ticket, and a template confirms it. Identification is one to many: a camera searches a crowd against a watchlist. Regulation (EU) 2024/1689, the AI Act, prohibits real-time remote biometric identification in publicly accessible spaces for law enforcement purposes outside narrowly listed cases, and those prohibitions have applied since 2 February 2025 per the European Commission’s AI Act page. The Commission’s own guidance is clear that biometric verification, confirming a person is who they claim to be, is not caught by that prohibition.
Data protection is the harder constraint. Under Article 9(1) of the General Data Protection Regulation, biometric data processed for the purpose of uniquely identifying someone is a special category, and processing is prohibited unless an Article 9(2) condition applies. Explicit consent is the usual route for a stadium. The European Data Protection Board’s Guidelines 3/2019 on video devices add the operational consequence at paragraph 86: a controller cannot make access to its services conditional on accepting biometric processing.
Read that as a design requirement rather than a legal footnote. If the biometric lane is fast and the alternative is a single staffed desk with a queue, consent stops being freely given in practice. The honest assessment is that biometric entry earns its place for repeat, pre-enrolled attendees, where enrolment happens weeks ahead and the template replaces a credential people lose. For a one-off capacity crowd, enrolment is the bottleneck, and a signed NFC pass is faster to deploy and far easier to defend.
What happens after the turnstile?
The ticket stops being useful as a locator. A scan proves that somebody entered at gate 7 at 18:42, and says nothing at all about where they are at 19:10, which is when concourse crowding, restricted zones and the walk to an unfamiliar seat actually matter.
This is the gap between access control and indoor positioning, and the two are frequently sold as though they were one thing. Access control is a boundary decision at a fixed point. Indoor positioning is a continuous estimate of where a device is inside a building, computed from signals such as Bluetooth Low Energy, Wi-Fi fine timing measurement under IEEE 802.11, ultra-wideband under IEEE 802.15.4z, or the phone’s own inertial sensors. Different physics, different accuracy, different failure modes.
Three things in a stadium depend on the second layer rather than the first. Wayfinding from the gate to the correct seat, aisle and restroom, which is the single largest source of staff questions in a building most attendees enter twice a year. Zone awareness for credentialed staff and contractors, where a ticket scan at the perimeter tells you nothing about who is currently in the players’ tunnel. And concourse flow, where knowing which stairwell is filling is more actionable than knowing how many people entered in total.
The integration point is the same credential. A ticket that already identifies a seat and a zone entitlement can drive both the turnstile decision and the in-app route to that seat, provided the two systems were designed to share it.
What should you ask a vendor before signing?
Ask how validation behaves with no network, where redemption state is held, how quickly a voided ticket reaches a reader, what the non-biometric lane looks like, and who owns the scan data.
| Question | A good answer sounds like |
|---|---|
| How does a reader validate offline? | Signature verified locally, no server lookup needed |
| Where does redemption state live? | One shared ledger, written by every reader |
| How fast does a void reach the gate? | Seconds, pushed to readers, not exported nightly |
| What is the non-biometric lane? | Staffed lane, comparable speed, no extra cost |
| Who owns the scan data? | The venue, exportable, with a stated retention period |
The last question is the one that is easiest to skip and hardest to undo. Entry data is the record of how your building behaves, and a contract that leaves it inside someone else’s platform in a format you cannot export has quietly decided your next five years of operational analysis for you.
Frequently asked questions
What does access control integration mean for a stadium ticketing system?
It means the credential the ticketing platform issues is the credential the entry hardware validates and marks as used. Three functions have to line up: issuance binds a credential to a seat, validation decides admissibility at the reader, and redemption records the use centrally so the same credential cannot be accepted twice at two different gates.
Can a ticket be validated when the gate loses network connectivity?
Yes, if the credential is cryptographically signed. The reader verifies the signature against a stored public key without contacting a server, so admission decisions continue through an outage. The trade-off is revocation: voided, refunded or transferred tickets stay valid at the reader until a separate revocation feed reaches it, which makes that feed a required part of the design.
Does Martyn’s Law apply to a ticketed stadium event?
It applies to qualifying events. Section 3 of the Terrorism (Protection of Premises) Act 2025 sets the conditions jointly: the public has access, 800 or more people may be present at the same time, and entry is by express permission with tickets, passes or membership checked. A qualifying event carries the same duties as enhanced tier premises. The Act received Royal Assent on 3 April 2025, and commencement is left to regulations.
Can accessible seats be sold through a separate channel from general tickets?
No. Under 28 CFR 36.302(f) and the Department of Justice ticketing guidance, accessible seats must be sold during the same hours, through the same purchase methods and outlets, and at the same stages of sale as other tickets, including presales. Pricing cannot exceed other seats in the same section, and a buyer may purchase up to three companion seats in the same row.
Is facial recognition entry legal at a European stadium?
Biometric verification, matching a person against their own enrolled template, is not prohibited by the EU AI Act, whose Article 5 bans target real-time remote biometric identification for law enforcement. GDPR Article 9 still applies, so explicit consent is the usual lawful basis, and EDPB guidance says service access cannot be made conditional on it.
Entry design depends on your own building: your gates, your readers, your ticketing platform. Book a scoping call and we will walk your gate flow together.
