Designing a photo scavenger hunt for a city you do not know
There is a particular kind of trip where the thing you want is not a list of sights. You have two or three days in a city, you have one free afternoon, and the children (or the friends, or the team) will not survive another museum. What you want is something that makes the unfamiliar streets feel like a place rather than a map you are reading.
That is the problem a photo scavenger hunt solves, and it is worth spending a little thought on the design, because a badly designed hunt turns an afternoon into a walk with errands attached.
Start from the streets, not from the landmarks
The instinct is to list the famous things: the cathedral, the castle, the market. That instinct is wrong for a hunt. Landmarks are already photographed by everyone who has ever been there, and a photo of the cathedral proves nothing - the point of a photo check is that the person taking it had to be standing in the right place at the right time.
Design the hunt around things that can be recognised but not faked. A red door is a red door. If your clue says "find something red", every red thing on the street unlocks the spot and the game collapses. What works is specific and verifiable: a particular shopfront, a fountain with a lion on it, a set of steps with a distinctive railing, a narrow alley you have to know to go down, a corner cafe with a blue awning. The clue can be generous in wording but it must be unambiguous in reality, because the person matching the scene has to be able to say yes or no without a second opinion.
Budget the time honestly
Most people get this wrong by over-packing. A hunt that takes two hours will not survive a group with tired legs and a hot day. Working out the length is arithmetic: pick a radius, pick a walking speed of about four kilometres an hour for a group with children, multiply, and then leave a margin. Fifteen to twenty minutes of actual walking for the whole hunt, spread over a sequence of stops, is the sweet spot. It should feel like a walk with a purpose, not a march.
Sequence the stops so the route does not double back. Draw the stops on a map first, then order them along a path rather than by how interesting they are. A hunt that sends people back across their own tracks is the fastest way to lose a group.
Keep the hints degrading, not failing
A hint should get less specific as the person uses it, not disappear. "Somewhere on this street" is the first hint. "On the side of the street with the blue door" is the second. "It is next to the bakery" is the last, and it should still be a place rather than a riddle. If the final hint is a pun, the hunt is over - somebody will be standing on the correct street with no idea what to look at.
Have a fallback for when the ground is wrong
This is the part that separate a working hunt from a broken one. Imagery goes out of date, a shop changes its sign, a spot becomes inaccessible. The design needs a ladder of degradation: widen the radius, widen it further, lower the difficulty by removing the photo check on the hardest stop, and only if all of that fails, say plainly that there is not enough coverage here and offer a different neighbourhood instead.
A hunt that cannot be completed is worse than no hunt, because it is a promise you break while the person is standing in the rain looking for a blue door.
Weather is the real enemy
Plan for the version of the day that is most likely, not the version you wish for. In most of northern Europe that means rain, and in most of southern Europe it means heat at two in the afternoon. The practical design answers are unglamorous: short stops, shade where possible, water nearby, and a route that can be abandoned and restarted later. The best hunts are the ones where stopping early does not break anything.
What this looks like in practice
We build exactly this kind of hunt. Ping Hunt generates one around whichever area you are staying in, from street-level imagery, and you match each spot with your phone camera to unlock the next one; you can also build your own hunt and share it as a link. The full description, including what it costs and where it works honestly (coverage depends on how fresh the imagery is in a given area), is on the docs page: https://www.pinghunt.com/docs/
If you would rather see it in a specific place, there are city pages for Copenhagen, Berlin and London.
The design principles above are the same whether you write the clues yourself or let a tool propose them. The part that matters is not the technology; it is the decision that every stop must be recognisable, that the route must not backtrack, and that there is always a way to finish the afternoon.

