One-on-One Scene Design
One performer alone with one participant behind a closed door, the whole encounter resting on a power gap that has to be made safe before it can be made beautiful.
A one-on-one is the most concentrated thing a live experience can do to a single person. One performer takes one participant into a closed room for five to fifteen minutes and gives them undivided, scripted attention. There is no crowd to hide in and no companion to glance at. Done well it becomes the moment the participant remembers above everything else in the show. Done badly it runs from flat to genuinely harmful, because the same closed door that makes the scene intense also removes every ordinary check on what happens inside.
The design problem and the safety problem are the same problem. The scene works because the social architecture is tilted. The performer holds the script and the room and the lighting and owns the only way out, while the participant stands alone and a little disoriented. That tilt is what produces the vulnerability the scene trades on. It is also what makes the room dangerous if either person abuses it. You cannot design the encounter and then bolt safety on afterwards. The safety envelope is the structure that lets the intensity exist at all.
This article treats consent and the power gap as load-bearing rather than as a compliance footnote. It works through what a one-on-one actually contains, how participants are chosen, the consent vocabulary borrowed from Nordic larp, and the scale arithmetic that makes “everyone gets a one-on-one” a promise no large show can keep.
The Mechanism (How it works)
What the scene is made of
Strip the theming away and a one-on-one has a fixed skeleton. There is an opening action that signals the scene has begun, a compulsory physical task the participant performs, and a closing gesture that hands agency back and lets them leave. The participant should never have to ask to go. The exit is built into the choreography.
The canonical examples come from a small set of well-documented productions. At Punchdrunk’s Sleep No More (New York, March 2011 to January 2025) the scenes were short and tightly fixed, often only a minute or two long. In one, the witch Hecate gave the participant a liquid from a vial. She led them backwards through a dark maze and pressed their hands into a wall of branches, then whispered a demand to find her ring before replacing their mask and pushing them out through a hidden door. Third Rail Projects’ Then She Fell (Brooklyn, 2012 to August 2020) ran longer and quieter. An Alice might show a doll collection and ask about love and loss. A Lewis Carroll dictated a letter for the participant to transcribe. A White Queen told a bedtime story while they lay on a hospital bed. The task is doing real work in all of these. Drinking, transcribing, holding an object to the chest: the small action gives the participant something to do with their hands and their attention, which is what stops the asymmetry of the room from curdling into stage fright.
The closing gesture matters more than its size suggests. The performer might replace the participant’s mask for them. Or they open the door and point the way out. Either move hands a small piece of control back to a person who has spent several minutes with very little. A scene that ends by merely stopping leaves the participant stranded and unsure whether they are even allowed to move.
How performers choose who to take
Selection is a designed beat rather than a whim. The performer reads the room and extracts one person, and the cues they read are consistent enough to teach. At Sleep No More the documented signals were sustained eye contact beyond an ordinary glance, single-minded attention to the performer rather than to a group, and a position in the crowd that allowed a clean extraction. The telling detail is what performers avoided: anyone visibly waiting to be chosen.
That avoidance is not snobbery. It is consent doing its quiet work. A person who has relaxed into following a performer has, in effect, been opting in continuously, where someone lying in wait has front-loaded a demand the scene then has to satisfy. The healthier model makes the opt-in explicit. Strange Bird Immersive frame it plainly. Players who want engagement signal it and players who do not look away, and the performer imposes on nobody. The cleanest way to read selection is as the first calibration check of the scene, not as casting.
A handful of selection choices are traps, and each one fails for a reason worth naming.
| Tempting target | Why it goes wrong |
|---|---|
| The most reluctant, shrinking participant | The scene becomes the performer’s agenda; consent is fragile from the first second |
| The most enthusiastic and disruptive | Rewards boundary-testing, and may hand the scene to someone trying to take it over |
| The visibly intoxicated | An obvious welfare risk once the door closes on a private room |
| Whoever is physically separated from their group | Produces an encounter that feels transactional rather than genuinely personal |
Scripted, improvised, or the middle
How much the scene bends to the individual is a real design decision, and the field runs the full range. The tightly scripted scene, as at Sleep No More, holds the same sequence whoever walks in, with responsiveness limited to pacing and physical proximity. It buys reproducible quality and predictable timing inside a multi-performer show, all at a lower training cost. What it pays for that is the participant’s sense of being genuinely seen. The fixed-beats-with-improvised-fill scene, the Then She Fell standard, keeps the opening action and the closing gesture fixed while letting the emotional fill between them follow whatever the participant says and does. One elderly couple began directing a Mad Hatter scene themselves and the performer let them carry on. Both sides experienced the hijack as a gift rather than a failure. Pure improvisation around a held character produces the most memorable encounters at the highest variance, and it works only with heavily trained performers.
The bridge between scripted reliability and improvised life is a catalogue. Strange Bird keep what they call an “esprit de l’escalier” document: a growing library of recurring participant behaviours paired with ideal in-character responses, so a performer has a pre-thought reply ready for the predictable situations and is free to invent for the rest. The trap is wanting consistency and responsiveness at once without admitting the trade. Each mode sustains a real show; the mistake is pretending you can have both for free. This is the live cousin of the scripted-versus-emergent tension that runs through interactive work, where the same choice is settled in branch design rather than in a performer’s body: Branching Narrative Craft.
The consent vocabulary the scene needs
A one-on-one needs more than a general agreement that the show might get intense. It needs tools the participant can reach for in the middle of the scene, to dial it back without having to blow it up. The most developed set comes from Nordic larp, and the project treats this category of agreement under the canonical name Safety Contract. The individual tools are worth naming precisely, because each does a different job.
| Tool | What it does |
|---|---|
| Cut | Full stop, immediate, no explanation owed |
| Brake (or Largo) | Reduce the intensity without ending the scene |
| OK Check-In | A non-verbal welfare check: a thumb-and-finger circle, answered thumbs-up, thumbs-down, or a flat waver |
| Lookdown | A graceful exit: shield the eyes and walk away, signalling “I am not in this” |
These exist so the participant can steer the pressure rather than choose between enduring it and ending it. The principle that travels with them is permission. A participant who trusts they can stop will go further in, which means the safety layer is what lets the designer build more intensity rather than less. The full toolkit and the reasons it works are Safety Tools as Design Layer, and the discomfort it governs is genuine even when the fiction is doing exactly what it should, which is the throughline of Trust, Conformity & Social Pressure.
Physical contact gets its own protocol, because “we might touch” is not a usable agreement. Ars Amandi (Emma Wieslander, first used at the larp Mellan himmel och hav, 2003) represents intimacy through restricted touch: permitted contact is limited to the arms, shoulders, sternum, upper back, and the neck below the ears, using hands, arms, and neck only, with the intensity carried mostly by eye contact and breathing rather than by reach. It is a larp mechanic, but its logic ports directly to a one-on-one. The permitted zones are named in advance and the intensity is built up slowly, with a non-verbal way for the participant to call a halt. The conventions now treated as a baseline follow the same spirit. There is no restraint of any kind, whatever the story says, so the participant can always leave. Contact above the shoulders or below the knee happens only after a pre-show briefing has said it will. And any request to be let out ends the scene at once, whether it comes in character or out of it.
When does that agreement get made? Not once, and not only at the door. Fast Familiar deliver content warnings at three separate moments. The first comes at booking and the second at entry, and the last is repeated on the threshold just before the experience starts. The repetition is deliberate, because a warning absorbed at the door lands far harder than one skimmed weeks earlier at checkout. The up-front version of this agreement, where the group settles tone and boundaries before anyone is in a room alone, is the live pre-event briefing covered in Pre-Game Workshop Design and The Session Zero Playbook.
The power gap, and who it actually endangers
The asymmetry is the point and the peril together. The performer holds the script, the room, the light, and the exit; the participant is alone, often masked or freshly unmasked, sometimes disoriented, without an ally. That imbalance is what opens a person up to the vulnerability the scene is built to reach. The same imbalance is what makes the room a place where harm can happen unobserved.
The instructive surprise is which direction the documented harm ran. In January 2018 a BuzzFeed News investigation reported that audience members had been groping and sexually assaulting Sleep No More performers, confirming seventeen incidents, with eight former cast and crew on the record and the production acknowledging seven. The format itself supplied the conditions. The mask anonymised whoever did it. The dim light and sanctioned proximity normalised physical closeness. Pre-show alcohol lowered inhibition, and the house instruction to follow curiosity and touch everything was loose enough to be misread as licence to touch the performers too. The show had carried no explicit no-touching instruction until that same month. That was nearly seven years into the run, and a cast member had asked for exactly that line years earlier. None of this means participants are dangerous. It means that an under-specified envelope around an asymmetric room gets exploited. The person left most exposed is not always the one the designer had in mind.
The seductive mistake is to treat the one-on-one as a normal scene played for one. The intensity it generates then gets read as proof the design is working. That intensity is real either way; on its own it tells you nothing about whether the room is safe. A scene can feel electric and still be a place where a performer is groped behind a closed door. It can feel electric while a participant is held past the point they wanted out. The tells are the gaps. There is no stated no-touch baseline. No briefed staff watch for distress. Neither person has a trained way to halt the scene. There is alcohol in the room and an exit that depends on the participant working up the nerve to ask. None of that infrastructure is overhead to be trimmed when the production runs late. It is the load-bearing wall. Cut it and the asymmetry that makes the scene memorable becomes the asymmetry that makes it harmful.
A standing intervention closes the gap the participant cannot close themselves. At well-run shows, production staff can enter any space and step in the moment a participant looks distressed. They do not wait to be asked, because the whole problem with a closed asymmetric room is that asking is exactly what gets hard inside it.
Landing the scene
A one-on-one does not end when the door opens. The intensity it manufactures produces bleed, the leak between what the character felt and what the person feels, and a participant can walk out carrying something heavier than they expected. The design question is whether the threshold they walk into is staffed by someone briefed to meet a person who is still processing, rather than left as an empty corridor. Fast Familiar run guided debriefs after charged work. Clinical professionals are present when the material warrants it, a practice borrowed from larp de-roling and from aftercare in intimate-play communities. The full craft of the landing is Bleed, Aftercare & Debriefing.
Strange Bird Immersive (Houston) carried the one-on-one into the escape room without letting the actor become “just another obstacle on the path to winning,” which contaminates both the puzzle and the performance. In The Man From Beyond and THE ENDings the two interactive actor scenes bookend the open puzzle time rather than threading through it, so the performance and the problem-solving never compete for the same attention. The supporting insight is that legitimacy of environment precedes openness to intimacy: adults will only drop their guard and play-act with a performer once the space feels serious and beautiful enough to be believable, so the build quality of the room is part of the consent design rather than decoration laid on top of it.
Why “everyone gets a one-on-one” is a lie the maths exposes
The scarcity is structural, not a failure of generosity. At Sleep No More, with roughly 450 to 600 people a night and about twenty-five performers running three loops each, a performer managed perhaps nine one-on-ones in a night, which left any given first-time visitor with something like a one-in-twenty-two chance of receiving one. The show was deliberately built to satisfy without the scene and to transcend with it. Every production resolves the same arithmetic differently, and each resolution carries a cost the designer is choosing on purpose.
| Approach | Production | The cost it accepts |
|---|---|---|
| Ration by scarcity | Sleep No More | Most people never get one and feel they missed the show’s best moment |
| Cap the audience | Then She Fell (15 a night) | Premium pricing and long runs needed to stay viable; routes choreographed per person |
| Serialise one at a time | You Me Bum Bum Train | Hundreds of unpaid volunteers per night, with the labour questions that follow |
| Simulate with small groups | various | Two to four people per scene; close to the feeling of solo attention, at a fraction of the load |
There is no arrangement that hands a generous one-on-one to every person in a three-hundred-seat show with a small cast. Pretending otherwise only manufactures disappointment, so the honest move is to design for the ratio you actually have. How a confined cast and limited routes get parcelled out across a paying crowd is itself a structural problem, related to the audience-throughput thinking in Team Sizes & the Bottleneck Problem.
Where this fits
The bigger picture. The one-on-one is the most concentrated expression of the bargain this whole family runs on, which is real pressure made legitimate by real consent. Everything that makes the scene worth designing comes from the power gap between the two people in the room. That gap supplies the undivided attention as much as the genuine vulnerability, and it is also what makes the scene the family’s highest-stakes safety case. So this article sits downstream of the consent infrastructure rather than beside it. The Safety Contract tools and the pre-event briefing and the debrief are not adjacent concerns a one-on-one designer can consult later. They are the conditions under which a closed asymmetric room is allowed to exist at all. Get that pairing wrong and no amount of evocative scripting redeems the scene; get it right and the most intense five minutes a live show can offer becomes available.
Next steps. To build the moment-to-moment regulation the scene depends on, work through Safety Tools as Design Layer, and lay the up-front agreement with Pre-Game Workshop Design and The Session Zero Playbook. Plan the landing with Bleed, Aftercare & Debriefing, since a scene this concentrated produces residue that an empty corridor cannot absorb. To understand why the pressure is legitimate in the first place, go back to Trust, Conformity & Social Pressure, which explains why a participant’s discomfort can be a sign of correct play instead of a fault. Where the performer’s authority is the lever being tuned, GM Role Models & Stances covers how the person running the encounter can hold or release a room.
Referenced by
- Bleed, Aftercare & Debriefing
- Player Types & Casting
- Pre-Game Workshop Design
- Romance & Relationships
- Safety Tools as Design Layer
- Scene Structure & Transitions
- Sleep No More & Punchdrunk