AR for events and brand activations
Events are where AR earns its money and where it most often disappoints. The difference is rarely the creative. It is whether the experience survives contact with a crowd, a queue and venue wifi, and whether anyone thought about how it would be measured before it was built.
Why do AR activations underperform their storyboard?
Because the plan assumes the audience will do something they will not do. The storyboard shows a visitor scanning a code and being delighted. The reality is a visitor who has to install an application first, on congested venue wifi, while queueing, with a colleague waiting. Most of them will not, and the ones who do arrive at the experience already impatient.
The second reason is device assumption. An experience signed off on the agency's current iPhone behaves differently on a mid-range Android handset three or four years old, which is what a large share of any real audience is carrying. If nobody tested that case, the activation degrades in exactly the moment it needs to impress.
Both problems are decided before design starts, which is why they are so expensive to discover on the day.
How many people can actually experience it?
This is the question that should shape the format, and it usually is not asked until afterwards. A headset experience is limited by hardware and staff: plan on four to six people per headset per hour once fitting, briefing and resetting between users are counted. Two headsets across a two-day show is a few hundred people at most.
Browser AR has no such ceiling, because every visitor uses their own device simultaneously. The constraint moves from hardware to bandwidth, which is a solvable engineering problem rather than a physical one. It also does not stop when the show does: the same link keeps working when a visitor shows a colleague on the train home.
If the objective is a small number of high-value conversations, headsets and staff are defensible. If the objective is reach, they are the most expensive way to limit it.
| Browser AR on visitor phones | Headset VR at the stand | |
|---|---|---|
| People per hour | Unlimited, in parallel | About 4 to 6 per headset |
| Staff needed | None for the experience itself | One per one or two headsets |
| Hardware to ship and charge | None | Headsets, spares, chargers |
| Floor space | Anywhere in the stand | Dedicated safe area |
| Works after the event | Yes, link still live | No |
| Hygiene and reset between users | None | Every user |
What breaks at a venue that does not break in testing?
Lighting is the first. Exhibition halls are lit brightly and unevenly, with spotlights and large dark gaps, and tracking that behaved perfectly in an office can lose its anchor under a gantry light. Testing has to happen under stage lighting, not desk lighting.
Wifi is the second, and it is usually worse than anyone expects because every stand is competing for it. An experience that loads 40MB of assets will fail at the exact hour the hall is busiest. Assets have to be budgeted in kilobytes, with the first meaningful frame arriving before everything has downloaded.
Crowds are the third. Body and face tracking that works with one person in frame degrades when three strangers walk behind them. Surface tracking struggles on the reflective and featureless flooring that exhibition halls are built from. Both are designable around, but only if someone raised them before the creative was locked.
Which activation formats work best?
Packaging and print activation travel furthest, because the entry point already exists and the visitor keeps it. A code on a pack, a flyer or a badge opens the experience, and the object stays in their bag as a reminder.
Product visualisation works where the real product cannot be at the stand: too large, too expensive, too many variants, or not built yet. It replaces the brochure conversation with something the visitor controls.
Shareable moments work when the output is an image or short video the visitor wants to post. That is the only format that reaches beyond the people physically present, and it is worth designing the output frame deliberately rather than treating it as a by-product.
Gamified experiences with a leaderboard suit stands that need dwell time and a reason for staff to start a conversation. They need more staff attention than the others, so they trade reach for engagement depth.
How should an activation be measured?
Instrument the funnel, not the total. Scans, sessions that reached a working tracking state, time in experience, completions, shares, and click-throughs to whatever the next step is. The number that matters is where people dropped out, because that is the only one you can act on next time.
Tracking-lock rate is the metric almost nobody captures and the most diagnostic one available. If half of sessions never reach a stable tracked state, the problem is environmental and fixable, and no amount of creative iteration will touch it.
Total scans on their own describe an activation rather than measure it. They also flatter the format that had the most footfall, which tells you about stand position rather than about the experience.
What has to be ready before the day?
A tested route from code to experience on at least one older Android handset and one older iPhone, over a deliberately throttled connection. If it has only been tested on the team's own phones on office wifi, it has not been tested.
A fallback for when the camera is denied or tracking fails. That can be as simple as a static 3D viewer or a short video, but the visitor should never reach a dead end, because a dead end at a stand is a conversation lost in front of an audience.
Printed codes at the size and height they will actually be scanned from, tested at that distance. Codes designed on screen and printed small at the last minute are a routine cause of failure that has nothing to do with the AR at all.
Common questions
- Do visitors need to download an app for AR at an event?
- Not if the experience is built for the browser. A visitor scans a code, grants camera permission and the experience opens as a web page. Requiring an app install at an event is the single largest cause of low participation, because it asks someone standing in a busy hall with a queue behind them to navigate an app store first.
- How many people will use an AR activation at a trade show?
- With browser AR there is no per-person ceiling, because every visitor uses their own phone at the same time, so participation is limited by footfall and by how visible the invitation is. With headsets, plan on four to six people per headset per hour once fitting, briefing and resetting are included.
- Will AR work on the venue's wifi?
- Only if the experience is built for it. Exhibition wifi is congested and slow at exactly the busiest hours, so assets have to be measured in kilobytes and the first usable frame has to arrive before everything has loaded. This is a build decision, not something that can be fixed on site.
- How far in advance should an AR activation be commissioned?
- Four to six weeks is a realistic minimum for a focused single-product experience, and longer where 3D assets have to be produced or cleaned up. Asset preparation is normally the critical path rather than the AR work, so the earliest useful conversation is about what product files already exist.
