How to Run a QR & NFC Scavenger Hunt at a Conference

Conference attendees have a short attention span and a long list of sessions to skip. The booths in the back corner get ignored. The sponsor who paid for premium placement watches people walk past without stopping.

A scavenger hunt is one of the few things we have found that reliably moves people around a venue, and the reason is not that it is fun. It is that walking to a specific spot becomes worth points, and points become a t-shirt.

We ran one across the ETHDenver 2026 MakerSpace from 18 to 21 February with 59 checkpoints and 207 players. This is how the format works and what the data from that deployment showed.

The checkpoint model

You place NFC tags or QR codes at specific locations. Each one is a checkpoint worth a reward. An attendee taps their phone on the tag or scans the code, and the reward lands immediately, with no verification step in between.

The two checkpoint types are not interchangeable. An NFC tag takes a physical tap, which suits tables, walls and hidden spots, and it is the one that feels like picking something up. A QR code works at a distance, which suits signage and booth banners, and it is easier to produce at scale. We used a mix of both at ETHDenver, across everything from the Main Stage entrance to the Zen Zone to sponsor booths.

The app download is the expensive part

Every step between hearing about the game and playing it costs you players, and installing an app is the largest single step: finding the store, searching, downloading, opening, creating an account, granting permissions.

Treasure Hunt runs in the phone browser. Attendees open a URL, sign in through unicorn.eth with one tap, and are playing in under 30 seconds. We do not have a credible industry figure for how many players an app download costs you, and we are not going to invent one, but our own onboarding is the thing we chose to optimise first and it is the thing we would change last.

At ETHDenver nobody was required to play, and 207 people did. The game spread by word of mouth and by people seeing tags on walls.

What the leaderboard actually does

The top 10 players at ETHDenver finished within 3,350 tokens of each other, on an average of 7,830 each. A gap that narrow is what keeps a leaderboard working: players check it, see they are one or two finds off the next rank, and go back out.

It does different work at different levels of engagement. A casual player looks at the board and discovers they found 5 checkpoints out of 59, which is an invitation rather than a score. A serious one plans a route; the most active hunter, zkprof, logged 67 separate mint events across four days. Groups race each other, which is where most of the noise on a show floor comes from.

The board also recruits. When people see others tapping tags and comparing numbers, they ask what it is.

The merch store is what makes the points mean anything

Points you cannot spend do not motivate anyone for long. Treasure Hunt includes a store where players spend earned tokens on physical items. At ETHDenver:

• Night Market Claw Play (1,000 BUFFI): 78 purchases
• Nova Blockchain Lab T-shirt (1,500 BUFFI): 15 purchases
• Nova Blockchain Lab Socks (700 BUFFI): 10 purchases
• 2026 ETHDenver T-Shirt (5,000 BUFFI): 8 purchases
• Pink and black hats (1,500 BUFFI each): 6 purchases between them

137 purchases from 73 shoppers, which is 35% of everyone who played, and 213,820 BUFFI burned on the way through. The burn matters more than it sounds: it keeps the token supply from inflating past the point where a reward feels like it cost something.

Placement is the organiser's real lever

You decide where the checkpoints go, which means you decide where people walk. That is the whole product.

At ETHDenver the most-found checkpoint was the Merch Night Market at 62 finds, in an area that was busy anyway. The more useful result was further down: a custom sneaker vendor in a back corner got steady traffic once a checkpoint went in at his booth, traffic he was not getting before.

For a sponsor this is a number instead of an estimate. You can say how many people came to a specific spot and when, and at ETHDenver every one of those interactions was recorded on-chain, so the sponsor does not have to take your word for it.

Running one yourself

Map the venue first and pick the locations that matter: sponsor booths, the areas nobody walks to, food stations, stage entrances. Each becomes a checkpoint. Reward tiers do not have to be flat, and making the hard-to-find ones worth more is the cheapest way to push people into the corners of a floor plan.

Stock the store with things people will carry home. Launch browser-based with a one-tap sign-in, because a barrier at the door costs you players you never hear about. Then put the leaderboard on screens and read the top names out during sessions, and it largely runs itself.

At ETHDenver the busiest hour across the four days was 5 PM, with 500 transactions, and the game held up through all four days without any further push from us after launch.

The infrastructure runs on Nova Cidade Chain, a Layer 3 network that records every action, with a paymaster covering the fees so players never pay anything. None of that is visible from the phone. You tap a tag and get a reward.

For the two checkpoint types in detail, see QR scavenger hunt for events and NFC treasure hunt for events. The full dataset is in the ETHDenver 2026 report.