By Ryan MitchellEvery event photographer eventually gets the same question from an organiser: "How do we actually get the photos to people?" It sounds simple. It is not. The delivery method you choose determines whether your guests actually retrieve their photos, or whether the gallery link disappears into an unread inbox and 85% of attendees never see a single shot from the event.
Two approaches dominate the market: QR code access and AI selfie search. Both have genuine use cases. Neither is universally superior. The right choice depends on your event type, guest profile, photographer workflow and how much you care about retrieval rate versus implementation simplicity.
How QR code photo delivery works
QR code delivery comes in two flavours and they behave very differently in practice.
Per-event QR codes give every guest the same code, which links to a shared gallery containing all event photos. This is the simplest implementation: one code, one URL, no individual setup. Guests scan it at the event, bookmark the URL and can browse all photos at any time. The obvious limitation is that with 600 guests and 3,000 photos, each person has to find themselves in an unfiltered gallery. Most do not bother.
Per-person QR codes solve the discovery problem by assigning each guest a unique code that links directly to a pre-curated gallery of their photos. This requires either manual curation by the photographer (time-intensive) or a photo-tagging workflow integrated into the registration and photo upload process. When it works, it is excellent. The logistics of printing and distributing unique codes at scale, especially for walk-in events, make it operationally complex.
Retrieval rate reality check: Per-event QR codes (shared gallery) typically achieve 40–60% scan rates at the event itself, but only 12–20% of those who scan actually download a photo. Per-person QR codes achieve higher download rates among those who receive them. But code delivery failure (lost badge, badge not scanned) erodes the effective rate significantly. Real-world end-to-end retrieval rates for QR-based systems hover between 25–45%.
How AI selfie search works
With selfie search, the guest is the key. They visit the event guest page, submit a selfie photo and the AI matches their face against every photo taken at the event, returning only the photos that contain them. No badge, no unique code, no prior registration required.
The matching typically takes under three seconds. The guest sees a personalised gallery, can download individual photos or the full set and share directly to social media. They never have to scroll through hundreds of photos of people they do not know.
The retrieval rate difference is significant. Well-implemented selfie search systems achieve 85–92% guest access rates when the gallery link is promoted actively during the event, via screens, compere announcements and follow-up SMS or email. The reason is frictionless personalisation: every guest gets a gallery that is immediately, obviously relevant to them.
The numbers side by side
| Factor | QR Code (per-person) | AI Selfie Search |
|---|---|---|
| End-to-end retrieval rate | 25–45% | 85–92% |
| Setup complexity | Moderate–High | Low |
| Works without a smartphone | No (needs camera) | No (needs camera) |
| Works without pre-registration | No (for per-person) | Yes |
| Handles 1,000+ guests | Operationally complex | Scales automatically |
| Personalised results | Only with per-person codes | Always |
| Data privacy footprint | Minimal (no biometrics) | Requires GDPR consent flow |
| Cost per event | Low | Moderate |
Which works better at large events?
At events over 500 guests, QR code logistics become the dominant failure mode. Printing 800 unique badge inserts, ensuring every badge gets the right code and handling lost badges or walk-ins who did not pre-register, each of these is a solvable problem individually, but together they create a meaningful operational overhead on the day. Any breakage in the code delivery chain (badge not scanned on arrival, guest mislaid their lanyard) means that guest gets nothing.
Selfie search scales cleanly past 500, 1,000 and 5,000 guests without additional operational complexity. One URL, promoted consistently throughout the event, serves the entire attendee list. The guest does not need anything except their phone and their face.
For conferences and corporate events where registration is managed months in advance and badge production is already part of the workflow, per-person QR codes can be built into the registration pack without meaningful extra cost. If you already have an individualised badge production process, adding a unique photo QR code is incremental work rather than a new system.
The hybrid approach: when to use both
Some events benefit from running both methods in parallel. The QR code provides instant, zero-friction access at the event itself, guests scan a table card and get to the guest page immediately, without having to find the URL themselves. The selfie search then handles the personalisation once they arrive at the page.
This is especially effective for gala dinners and award ceremonies where each table already has printed materials. A QR code on the menu card gets guests to the gallery. The selfie matching does the work of finding their photos. The result is better than either method alone: low friction entry, high-quality personalised delivery.
When QR codes still make more sense
There are scenarios where QR codes remain the better choice:
- Small events under 100 guests where manual curation of per-person galleries is feasible and the organiser wants to avoid any biometric data collection entirely.
- Events with no consent infrastructure - selfie search requires explicit, logged GDPR consent from each guest. If your event has no mechanism for collecting and recording consent, QR is the correct default.
- Events where guests may object to facial recognition on principle, the technology is well understood and privacy-safe when properly implemented, but some audiences (certain professional communities, privacy advocates) are sensitive to facial recognition use regardless of the underlying data architecture. Forcing the issue is not worth it.
- Scenarios where no face photos are taken - product launches or exhibitions where the photography is of the products or installations rather than attendees have no use for selfie search at all.
The honest answer on "which is better": If your primary goal is getting the highest percentage of attendees to actually receive and keep their photos, selfie search wins for any event over 100 people. If your primary constraints are budget, data privacy simplicity or a guest profile where facial recognition consent is sensitive, per-person QR codes are a respectable alternative, provided the logistics are handled carefully.
Data privacy: a meaningful difference
QR code delivery has essentially no biometric data footprint. A URL is a URL. No face data is generated, stored or processed. For organisers whose guests or legal counsel are particularly cautious about facial recognition, regardless of how well the GDPR compliance is architected, this is a real differentiator.
AI selfie search collects and processes biometric data (face embeddings) and therefore requires a proper consent flow under GDPR Article 9. This is not a reason to avoid selfie search, a well-designed consent flow takes 15 seconds and can be embedded in the guest page naturally. But it is a requirement, not optional. Any selfie search vendor who does not surface this as a primary concern in their implementation guidance is one to approach with caution.
A third option gaining ground: WhatsApp-first delivery
Since this post was first published, a third approach has emerged as the dominant preference in GCC markets and is growing rapidly in South Asia and Southeast Asia: WhatsApp-first photo delivery. Rather than requiring guests to scan a QR code or submit a selfie to a web portal, the platform sends their photos directly to their WhatsApp number. The guest receives a notification, taps it and sees their photos immediately. No QR code required. No portal to navigate. No gallery link to bookmark.
How it works in practice: guests register their WhatsApp number at event registration. When their photos are ready - matched by AI face recognition from their selfie or from photos taken at check-in - the platform sends them directly via the WhatsApp Business API. The photos arrive as a native WhatsApp message, which means they sit in the same conversation thread as other messages the guest has with real people. The open rate for WhatsApp messages is consistently above 90% compared with 20-30% for email links.
The comparison then becomes three-way:
- QR code galleries suit events where guests are likely to be seated and browsing - award ceremonies, conference sessions and gala dinners where guests have time to scroll and download.
- Selfie search suits events where guests want to find their own photos without any pre-registration - walk-in conferences, public events and festivals where a simple selfie delivers a personalised result instantly.
- WhatsApp delivery suits events where guests are on-the-move, particularly in regions where WhatsApp is the primary communications channel. Photos arrive without the guest taking any action beyond their initial registration, making it the lowest-friction option of the three.
Many events in the GCC now use all three in parallel: guests who registered with WhatsApp receive their photos directly, guests who prefer to browse use the selfie portal and a public QR code at exits gives late registrants access to the full gallery. The choice does not need to be exclusive.
Not sure which method suits your event?
Book a demo and we will walk through your specific event type, guest profile and photographer workflow. Recommend the right delivery setup for you.
Book a free demo