UX case study · Service design

One room becomes many, the moment a wall moves

Some big rooms are built to change shape: a hall splits into smaller rooms for the afternoon, then opens back up at night. This is the design that let the people who run those spaces set the lighting up once, in plain language, and have it follow the walls on its own.

Read the full case study See it move
One big room, or five smaller ones, from the very same floor space
What it is

A room that changes shape through the day

Some spaces are divided by walls that slide open and shut, so one big room becomes several smaller ones and then opens back up. You have felt it in hotel event halls, conference centres, school gyms. The catch is the lighting: each smaller room should light on its own when the wall is closed, and the whole space should light as one when it opens.

I led this end to end: framing the idea, shaping how the setup is structured, testing it with the people who actually configure these rooms, and delivering it for phone, tablet, and desktop.

Role
Lead Experience Designer
Scope
Concept & structureUX · delivery
Surfaces
Phone · TabletDesktop
Process
Research · prototypeTesting · build
Domain
Flexible spaces · lighting
Worked with
Product · engineering
How we framed it

From a real room to a simple model

Before any screens, the idea had to be explained: take a room that divides, and turn it into a handful of named rooms, the walls between them, and what each wall does. This is the framing I shared with the team.

Concept-framing slides that map a real venue onto the model of rooms, walls, and what each wall joins
The concept, mapped from a real space to a simple model
The problem

It only works if it's set up right

Flexible rooms are valuable: the same floor space earns its keep in different shapes. But teaching the lighting to follow the walls was complex and easy to get wrong. The person setting it up had to hold a tangle in their head, which rooms, which walls, which wall joins which rooms, which lights belong where, and what tells the system a wall just moved. In early testing, people got lost.

01
A wall is a relationship, not a gadget
What matters about a wall is the two rooms it joins or separates. The old setup had no way to say that, so a wall was just one more thing in a long list.
02
It has to follow the walls by itself
Open a wall mid-event and the lights should merge on the spot; close it and they should split again, with nobody hunting for a control.
03
Set it up once, get it right
There is a lot to define, and one wrong link lights the wrong room. So the setup had to guide you, and be hard to get wrong.

“The wall moves. The lights should just know. I don't want to go find a laptop and open a program, I'm holding a phone.”

Someone who sets up these rooms · research interview
The service, end to end

From an empty room to one that runs itself

Four very different people touch this before a guest ever notices the lights. Mapping the whole path is how each step earned its design.

01Plan
Who
Space planner
What happens
Sketch the room and the ways it can divide.
The design job
Give them a picture that matches how they already think.
02Install
Who
Electrician
What happens
Fit the small sensors that notice when a wall moves.
The design job
Keep the hands-on part simple and forgiving.
03Set up
Who
Setup tech
What happens
Name each room; say how many walls and sections.
The design job
Turn a complex space into a few plain choices.
04Connect
Who
Setup tech
What happens
Link each wall to its sensor; give each room its lights.
The design job
Make the step people got wrong impossible to fumble.
05Use
Who
Staff & guests
What happens
Walls move through the day; the lighting follows on its own.
The design job
After setup, it runs itself.
Behind the glass

One moment, top to bottom

Take the simplest thing that happens all day: a wall slides shut for the afternoon sessions. A service blueprint shows every layer that makes the lights do the right thing, from the room down to the sensor.

01A wall closes
02The system notices
03The room splits
04Each side runs alone
In the room
Staff slide the wall shut
Guests carry on, either side
Someone dims the lights on this side
Only this side responds
What people feel
Two rooms now, not one
No menus, no setup
Controls here affect here
Presence and daylight behave per side
System, unseen
A sensor reports the wall is shut
The space is split into two rooms in logic
Controls are scoped to each side
Light commands stay within each side
What holds it up
The link made at setup: this wall joins these rooms
Each room's lights, assigned once
Presence sets tied to each room
The plain floor-plan model underneath
The approach

Model the room the way people think about it

The key decision was to build the setup around rooms, not wiring. The people who run these spaces think in rooms: this room, that room, the wall between them. Get that right, and all the technical plumbing can hide behind a simple floor plan anyone can read.

01
Rooms come first
Each room owns its own lights and settings, the way staff describe the space out loud.
02
A wall is just a link
A wall is modelled as "these rooms join when this opens," not another gadget in a flat list.
03
It senses its own shape
Each wall is watched, so the room always knows whether it is open or split, with a manual override.
04
Guided once, then automatic
A short, walked-through setup removes the guesswork; after that the room re-lights itself as walls move.
Thinking it through first

Map the real room before the app

The setup only feels simple because the thinking happened up front: every wall, every room, and every light mapped from a real floor plan onto the model the software needs.

A floor plan with each wall drawn around it, mapped to the rooms it joins
1 · Map each wall to the rooms it joins
The plan broken into rooms with groups of lights marked out
2 · Break the space into rooms and light groups
Lights and inputs assigned into each room's groups against the plan
3 · Assign the lights and inputs into each room
A finished room's light list checked against the plan
4 · Check the finished room against the plan
Try it

Open a wall. The rooms merge.

That same model, running. Tap any wall to open or close it. Sections joined by an open wall light as one room. This little interaction is the whole idea the setup exists to protect.

From the product

The shipped setup, five guided steps

The real screens that take a room from "this one divides" to fully wall-aware. Guided once, then it runs itself.

The setup summary screen: each wall listed with the rooms it joins, and a wall's sensor input assigned
Every wall matched to the rooms it joins, and pointed at its sensor
Setup screen: creating a new space
01
Create the space
Start a new space, the entry point for the whole setup.
Setup screen: choosing the divisible type and the number of walls and rooms
02
Describe the room
Set how many walls and rooms the space can divide into.
Setup screen: naming each room
03
Name the rooms
Give each room a name people recognise.
Setup screen: choosing which rooms each wall joins
04
Match walls to rooms
Say which rooms each wall joins, the link that drives merge and split.
Setup screen: pointing each wall at the sensor that watches it
05
Point each wall at its sensor
Connect each wall to the sensor that watches it, so open or shut is known live.
Research & testing

We watched people get stuck, then fixed exactly what stuck

Before building, we put an early version in front of the people who would really set these rooms up and watched where they hesitated. Every point of confusion became a specific change, and we tested again.

They got stuck on

Linking a wall to the sensor that watches it.

So we

Made it its own walked-through step, instead of a setting buried out of sight.

They got stuck on

Moving between the whole room and its temporary sub-rooms.

So we

Put that navigation where people expected it, and made phone and desktop behave the same way.

They got stuck on

An icon that was supposed to mean "sub-room."

So we

Dropped it for a pattern people already recognised.

They got stuck on

Landing on the details page and not knowing the next move.

So we

Gave each room an expand-and-collapse, so the next action is obvious and only what is needed loads.

They got stuck on

Too much on screen, too much scrolling, unsure what was already set.

So we

Put the counts up front and added checks that block an invalid setup before it happens.

Evidence over opinion

When the setup grew to several steps, the team debated how to show progress. Rather than argue it, I researched the options against our design guidelines, built a prototype of the contested version, and let the next user test settle it. The point of a test is to end a debate, not to confirm a hunch.

Early prototype → moderated usability sessions → redesign → validation round
Inside a room

Fill each room, one at a time

After the guided setup, each room is filled in from a shared pool: pick a room, assign its lights and presence sensing, and confirm each fixture. The same simple pattern, repeated per room.

Configuring a room: selecting several lighting zones to assign at once
Assign the room's lights in a batch
Configuring a room: the pool of unassigned devices, grouped by type
The unassigned pool, grouped by type
Configuring a room: devices assigned, each showing its zone
Assigned, each light carrying its zone
Configuring a room: presence sets assigned to the room
Presence sensing, the same way, per room
The whole space

What it all adds up to: one space holding its rooms, with any room opened to its own lights and presence sets, and the walls one edit away.

The assembled space: its rooms listed, one expanded to show its zones, devices and presence sets, with edit actions for walls and rooms
The assembled space, every room in its place
On a phone

The same guided model on a phone, so a room can be set up standing at the panel, no laptop required.

The same setup shown as a sequence of phone screens, from creating the space to naming rooms, matching walls and assigning devices
Outcome

Confusion in, clarity out

Every problem we saw in testing had a specific answer in the next version, and the reworked setup was approved to build. The same plain model, rooms first, walls as links, held up on a phone at the wall and on a desktop at a desk. The figures below are what the design is built to handle, not numbers I can verify for you here.

At a glance

The shape of it

Not outcomes I can measure for you here, just what the setup is built to handle, and how it's meant to be used.

0
Rooms one space can be split into
0
Movable walls the setup can track
0
States a wall can be in, open or shut
0
Surfaces at parity, phone, tablet, desktop
Plan → install → set up → it runs itself · phone · tablet · desktop

The breakthrough wasn't a screen. It was deciding that rooms, not walls, should be the thing you set up, and letting all the wiring hide behind a floor plan anyone can read.

View Quick tour Deep dive