Guerrilla Projection: How It Actually Works on the Street

Guerrilla projection is a crew in a vehicle, a wall that was not expecting them, and a window of time that closes.

That is the whole format. Everything else is detail about how you make those three things line up often enough to call it a campaign.

I want to separate this from the version of projection most buyers have seen. A booked architectural projection is a planned thing: a surveyed building, an agreed owner, a rigged position, a set run time, a crew that stays until two in the morning. It is beautiful and it is slow.

Guerrilla projection is mobile, short duration and opportunistic. The projector lives in a vehicle. The crew moves between positions through the night. Each image lives somewhere between a few minutes and an hour, then the van moves. The output is not dwell time on one wall, it is a sequence of appearances across a city and the footage and photographs that come out of them.

Different tool, different brief, different failure modes. Here is how we run it.

The vehicle is the format

Everything about guerrilla projection follows from the fact that the projector is in a vehicle.

A projection vehicle is fitted with the projector, a power source, a media machine, and a mounting arrangement that lets the crew aim through a side window, a rear door or a roof position without unloading anything onto the street. The point is to be able to arrive, aim, run and leave without a rigging operation.

That constrains you in specific ways. The throw is whatever the parking position allows, which means the image size is determined by the street, not by the creative. You are working at the angle the kerb gives you, so keystone correction is doing real work on nearly every position. Your power is onboard, which caps output. And you are parked, which means you are subject to all the ordinary realities of a parked vehicle in a city at night.

It also gives you the thing that makes the format work: you can be at six positions in a night instead of one. If one wall is wrong, you drive two blocks and try another. A fixed rigged install has no such option.

Some crews work on foot with a battery projector for very small scale work, typically onto sidewalk surfaces, alley walls and shuttered storefronts. That is real and it has its place, mostly for tight, low, close range images. Output is modest, so the image has to be small and the surroundings dark. We use it as a supplement, not a main plan.

Picking positions on the move

We do not improvise the whole night. We scout.

Before a run we drive the city after dark and build a position list. For each candidate we note the surface, the throw position, the angle, the ambient light, what the audience situation looks like at that hour, and anything that would make the position awkward, such as a bus stop, a loading bay, a hydrant, a doorway in use, or a venue with security on the pavement.

What we look for is a light coloured, matte, reasonably unbroken surface that is dark enough around it for the image to read, with a parking position roughly opposite at a distance that gives a usable image size, and people nearby who are walking rather than driving.

The reliable categories are familiar once you have done it. Blank side elevations of buildings on a corner. Shuttered retail frontages after closing. Construction hoarding and site boarding, which is often pale, flat and large. Parking structure walls. The end wall of an older building where a neighbour has been demolished. Underpass and bridge abutments, which are usually dark and often have a pedestrian route beneath.

A good night's list has more positions on it than we intend to use, because perhaps a third will be unusable when we get there. Something is parked where we need to be, a shutter is up, a street is closed, a new light has been installed. Redundancy is the plan.

What a night actually looks like

A guerrilla projection run uses three people. A driver who stays with the vehicle, an operator who aims and runs the image, and a spotter who stands where the public sees the wall, judges the image, shoots the stills and video, and is the person who talks to anyone who comes over.

The spotter is the one most clients would cut and the one you should never cut. You cannot evaluate a projection from behind the projector, you cannot photograph it from there, and you need somebody between the vehicle and the public.

The rhythm goes like this. Arrive, assess the position in under a minute, aim and focus, run the image, shoot the content, move. Typical dwell at a position is somewhere between five and thirty minutes. Longer than that and the value curve flattens, because the people walking past have changed over, and the longer you sit the more likely a conversation happens.

A productive night covers four to eight positions. More than that and the driving eats the night, especially in a city with slow cross town routes.

Timing matters more here than in any other format we run. You need real darkness and you need people. Those two overlap in a window that is shorter than clients expect. In most cities the sweet spot runs from full dark until somewhere around midnight on a weeknight and later at weekends, concentrated wherever people are leaving venues, waiting for transport or walking between bars and restaurants.

In summer in a northern city that window does not open until late. In winter it opens early and lasts, which is why I would rather run guerrilla projection in November than in June.

Cities and streets where this works

The format needs pedestrian density and dark surfaces at the same time, which is a narrower combination than it sounds.

In Brooklyn the industrial stretches in Bushwick and the blocks around Williamsburg have large blank masonry elevations with nightlife traffic beneath them, which is close to ideal. In Chicago the area under and around the elevated structure gives dark, dense, pedestrian heavy positions, and the older masonry buildings in the near north and west take light well. In Philadelphia the blocks around Fishtown and Northern Liberties have the right mix of blank walls and foot traffic.

Los Angeles is a positive for surfaces and a problem for audience, since the pedestrian clusters are specific and separated by long drives. The arts district downtown is the clearest exception, with plenty of plain warehouse walls and people on foot.

Austin's downtown nightlife core and the areas around the east side bars are dense enough to work. Nashville's lower broadway surroundings have the crowds but a very high ambient light level, which pushes positions a block or two off the main strip. Miami gives you warm humid air that often leaves haze in the beam path, which some creatives like and which does soften the image.

Denver, Seattle and Portland all have good building stock and early winter darkness. Seattle and Portland also have the rain problem, which for this format is fatal more nights than people plan for.

Surface and weather truth

Pale and matte wins. Painted render, light brick, concrete, pale hoarding, roller shutters that have been painted a light colour. These give you an image with real saturation.

Dark brick, dark stone, glass and polished metal all lose. Glass is close to useless, since most of the light goes through it. Heavily textured surfaces break up fine detail but can look excellent with bold simple artwork, because the texture reads as deliberate.

Ambient light is the constant fight. A wall under a streetlight is a wall you cannot use. Positions a block off the brightest strip almost always outperform positions on it, which is counterintuitive to clients who want to be where the crowd is thickest.

Rain is the end of a night. A wet wall scatters the image, the beam path lights up, and the street empties. We watch the forecast and we would rather postpone than go out and get nothing, because a wasted crew night with no footage is the worst outcome in this format.

Haze and mist soften the image and make the beam visible. Depending on the creative that is atmosphere or it is a problem.

Wind is less of an issue than with rigged positions, since the vehicle is heavy and the projector is inside it, but gusts still produce visible shake on long throws.

Cold is manageable. Condensation on optics when kit goes from a warm vehicle into cold air is a real thing and the crew manages it by letting the equipment settle.

The permission picture, framed honestly

This is where guerrilla projection is most often misrepresented, so I will be careful.

Projecting an image onto a surface without the owner's agreement sits in a grey area in the United States. It varies by municipality. Some cities have signage ordinances broad enough to reach projected images, some have rules written entirely around physical signs that do not obviously apply, and some have nothing specific at all. Enforcement is inconsistent, and the view taken often depends on surface ownership, the content, the time of night and whether somebody complains.

Nothing here is legal advice, and I am not going to tell you that this format is either legal or illegal, because that depends on the city, the wall and the circumstances.

What I will tell you is how we reduce exposure.

We prefer permissioned surfaces. On a large share of our guerrilla projection nights, a meaningful number of positions are walls where somebody has agreed: a cooperating business, a venue, a developer with hoarding, a property manager who likes the idea. Those positions are the ones where we can dwell longest and shoot the best material.

We keep dwell short on positions where we are relying on the general tolerance of a dark wall at midnight.

We never leave anything behind. Nothing is applied to any surface, nothing is attached, nothing needs removing. When the projector goes off, the wall is exactly as it was. That single fact changes how most conversations go.

We stop when asked, immediately, with no argument. The crew is instructed on that and it is non negotiable.

We avoid categories entirely. Anything aimed into traffic sightlines or at drivers. Residential windows. Anything projected skyward. Government, religious and sensitive buildings. Transit infrastructure. Hospitals. Content that would be inflammatory on a stranger's wall.

We document positions, times, images and permission basis for everything.

If a client asks us to project content that is designed to provoke, onto buildings belonging to people who have not agreed, the answer is no. Not because of squeamishness, but because that is the kind of job that ends a campaign in week one and damages everybody's ability to work in that city afterwards.

What drives scope up or down

The driver of scope in guerrilla projection is crew nights, not technology.

It grows with the number of nights, the number of cities, and the amount of scouting required in a market we have not worked before. It grows when the client wants specific named walls rather than a corridor, because specific walls mean permission conversations and fallbacks. It grows with content complexity and with any requirement for same night delivery of edited footage. It grows when weather contingency nights are held.

It shrinks when we have worked the market before and already hold a position list. It shrinks when the brief is a corridor or a neighbourhood rather than a named building. It shrinks when the creative is one bold asset that works at any size. It shrinks when the dates fall in a season with early darkness. And it shrinks when nights run consecutively, since scouting carries over.

The least efficient version of this format is one night, in a city we do not know, on a specific building, with a fixed date. The most efficient is three or four consecutive nights in a neighbourhood with a flexible position list.

What goes wrong, and what we refuse

Positions evaporate. A third of any list will be unusable on the night. Crews that do not over scout end up standing around.

The wall was glass, dark, or lit. The scouting pass exists to catch this, and skipping it to save a night always costs more than it saves.

The artwork was built for a screen. Fine type, gradients and dense layouts disappear outdoors. The content that works here is a single bold mark, big legible words, and slow movement if any.

Nobody shot the material properly. This is the big one. Guerrilla projection's primary output is content. If there is no competent spotter with a camera, you have spent a night projecting to a few dozen passers by and have nothing to show for it. The footage is the deliverable as much as the projection is.

The weather was ignored and the night was lost with no fallback date.

The content list was too long. Six messages across a night means no single image gets established anywhere.

Now the pushback. The brief I decline most is guerrilla projection as the main coverage vehicle for a campaign. Clients ask for projection to carry awareness across a city for a month, and it cannot. The in person audience for any given position is small, the format only runs on dry nights in a limited window, and there is nothing left on the wall in the morning.

What it is superb at is making something that looks and feels like it was not bought. The photographs and the footage travel. A projection night produces assets a brand uses for weeks, and it reads as presence rather than media spend.

So when the brief is genuinely about coverage, I steer the coverage work into Wheatpasting & Poster Campaigns, where the message sits on many surfaces for weeks and the city actually feels covered, and keep the projection for the nights and the images that are worth filming. Projection creates the moment, posters hold the ground. Sold that way, both do their job. Sold as a substitute for each other, neither does.

The other thing I talk clients out of is insisting on a single iconic building. Those buildings are lit, surveilled, usually glass, and often the single worst projection surface in the city. The wall that photographs best is almost never the famous one.

What you get back from a night

Position by position records with locations, times and the permission basis. Stills from the public viewing position. Video of the projections in situ with people in frame where we can get it. Notes on positions that worked and ones that failed, which is what makes the next night in that city faster.

If a planned position was abandoned, we say why. Street closed, parked over, shutter up, light installed, somebody asked. That record is genuinely useful, because the position list is an asset that improves every time we use it.

Frequently asked questions

Request a campaign quote

Tell us the city, the neighbourhood or corridor you care about, the nights you have available, and what the single strongest image is. If you want the footage as a deliverable, say that early, because it changes how we staff the night.

We will scout, build a position list, tell you honestly how many nights the weather risk justifies, and recommend what should run alongside it so the campaign has ground as well as a moment.

Request a campaign quote and we will plan the night properly.