What a Custom Immersive Project Needs to Start

A custom immersive project has no template, which means it also has no default answers. Every question that a standard product would have settled in advance has to be answered by somebody, and if it is not answered it gets assumed.

This guide lists what has to exist before production starts. It is deliberately ordered by consequence rather than by chronology, because the expensive omissions are not the obvious ones.

The single sentence everything else depends on

Before any document is requested, one sentence has to be written and agreed: what should a visitor understand or feel after this that they did not before.

It sounds soft and it is the hardest working line in the project. It decides the sequence, the level of detail, the length, whether interactivity is needed and what counts as finished.

Without it, every subsequent decision is made on taste, and taste disagreements between stakeholders have no resolution mechanism. With it, most arguments resolve by asking whether the proposed addition serves that sentence.

If the group cannot agree on the sentence, that disagreement is the real project risk, and discovering it in week one is a gift rather than a delay.

The audience, described specifically

A brief that says buyers, investors and stakeholders has described nobody.

What production needs is concrete: how many at a time, whether they are accompanied, how long they have, what they already know, and what they are worried about.

A supervised pair of prospective buyers with twenty minutes is a completely different design problem from forty people at an evening session with fifteen minutes and a scepticism about the project.

Both are legitimate. Designing without choosing produces something that serves neither.

The subject, and what exists of it

This is the item that determines cost more than any other and it has to be established honestly.

Does the building exist. Does part of it exist. Is there a 3D model, and has anybody opened it recently and confirmed what state it is in. Are there drawings, and are they current. Is there a survey.

The common and expensive situation is a project that believes it has a model because the architect has one. A coordination model built for construction is not a display model, and converting it is production work rather than a file export.

Establishing this before quotes are requested turns an unpredictable price range into a comparable one.

The room, measured rather than described

If the experience lives in a physical space, the space has to be surveyed before a format is proposed, not after.

The five numbers that decide feasibility are ceiling height, clear floor area, distance from a projector position to a surface, available power, and how much daylight enters and when.

Ambient light is the one most often waved away and the one most likely to ruin the result. A projected experience in a room with uncontrolled daylight either needs blackout treatment or substantially more expensive equipment, and both are budget items.

Column positions and ceiling structure decide mounting and sightlines. In an older or converted building neither can be assumed from a plan.

Who owns the space, and what they permit

Somebody controls the room and their rules arrive whether or not anybody asked.

Can anything be fixed to the structure. Is there a limit on power draw. Who has to approve the installation. What happens to it at the end of the lease, and who removes it.

For a sales centre inside a building under construction, the contractor programme governs access, and an installation scheduled outside that programme will simply not happen when planned.

These are not technical questions and they stop projects more often than technical ones do.

The decision about supervision

Guided or self service has to be settled at the start because it changes the design rather than the settings.

Guided experiences can carry complexity, adapt to the visitor and recover from confusion, because a person is present. They depend on that person being available, trained and consistent.

Self service experiences must be obvious in seconds, tolerate misuse, return to a clean state without help and work for somebody who reads nothing.

Deciding late produces the worst of both: too complex to leave alone, too shallow to justify staffing.

Content sources, itemised

Every element that appears in the experience has to come from somewhere, and listing them early prevents the most common schedule failure.

Architectural geometry, materials and finishes, landscape, surrounding context, interior furnishing, branding, photography of anything existing, plan and unit data, pricing or availability if it is shown, and any text.

Each item needs an owner and a date. The ones that reliably arrive late are finishes, unit data and anything requiring legal review.

Anything showing price or availability also needs a rule about how it stays current, because an experience displaying a sold unit is worse than one displaying nothing.

Approvals, and who can say no at the end

Identify early who has the authority to reject the finished thing, because they will exercise it.

Design, marketing, sales and often an owner or investor each have a view, and their criteria differ. Design cares about accuracy, marketing about impression, sales about usefulness, ownership about cost and risk.

Getting all of them into the sequence review, on paper, before production starts is far cheaper than getting them into a review of a finished installation.

If somebody with veto power cannot be brought into the early review, that is a risk to name explicitly rather than hope about.

The technical facts about the network

Unglamorous and responsible for a large share of failures on opening day.

Is there network access in the room. Is it reliable. Does the experience need it to run, or only to update. What happens when it drops during a presentation.

The safe default for anything in a sales environment is that the experience runs locally and uses the network only for updates and analytics. Experiences that stream content are hostage to a connection nobody in the room controls.

For browser based delivery this reverses, and the question becomes what mobile signal is like in that space, which has to be measured rather than assumed.

The operator, named

Somebody has to run it, and that person needs to exist before the build rather than after.

They need to know how to start it, how to reset it, what to say while it loads, what to do when it fails, and who to call. That is a training deliverable and a one page sheet, both of which belong in the scope.

The strongest predictor of whether an immersive experience gets used is whether a person confidently offers it. An uncertain operator quietly stops offering it within a fortnight.

The support arrangement for after

Anything intended to run beyond a single event needs a written answer to what happens when it breaks.

Who responds, in what time, at what cost, and what the site can do without them. Spare hardware, if any, and where it is.

This is routinely deferred and it is the difference between an experience that runs for two years and one that is dark by month four with nobody authorised to fix it.

Ownership of the content, in writing

The content outlives the experience, the vendor and usually the hardware, so ownership decides what the project has at the end.

Owning the source content means the model, materials and assets, not only a compiled experience running inside a vendor system. With it, the project can produce renderings, a browser version, a later phase and a replacement experience.

Without it, every future output starts from nothing, and the leverage to negotiate that clause disappears the moment the first invoice is paid.

Accessibility, decided at the start rather than retrofitted

Accessibility in immersive work is treated as a compliance item added late, and treating it that way reliably produces something that satisfies nobody and costs more than designing for it would have.

The practical requirements are concrete. Can somebody seated reach and see the interface. Does the experience work for a visitor who cannot hear it, which in a noisy sales gallery is most of them. Does it work for somebody who reads slowly, or in a second language, or who has limited fine motor control on a touch surface.

Most of those improve the experience for everybody. Large legible type, high contrast, generous touch targets and an argument that survives with the sound off are simply better design, and they cost nothing when specified early.

What is expensive is discovering after installation that the only interface is a small touch screen mounted at standing height, and that a meaningful share of visitors cannot use it.

The legal and factual review nobody schedules

Property experiences make claims, and claims attract review, so somebody has to be assigned to check them before the experience is public.

Views shown from a specific unit, distances to landmarks, finishes described as included, availability, pricing, and anything implying an amenity that is not contractually committed all carry exposure.

Views are the recurring problem. An experience that shows a spectacular outlook from a specific floor is making a representation about something a neighbouring development can remove, and being able to demonstrate what was shown and when is worth the effort of recording it.

The review takes days when planned and weeks when it collides with a launch, so it belongs on the schedule as a task with an owner rather than as a final formality.

What to do when the design is not finished

The uncomfortable common case is that the experience is needed before the design is settled, and refusing to start is usually not an available option.

The workable approach is to build what is stable and defer what is not. Massing, structure, orientation and layout tend to settle early. Finishes, furniture, landscape detail and amenity fit out change late.

Structuring the content so the volatile parts are separable, rather than baked into a single fixed output, is a decision made at the beginning that determines whether a late change is an adjustment or a rebuild.

It also requires an agreement about how many revisions the volatile parts get, because without one they get revised until the launch date arrives.

A realistic timeline, and what runs in parallel

Specification first, and it takes longer than anybody plans because it requires agreement rather than work.

Content production is the longest single task and can begin as soon as geometry sources are confirmed. Material decisions are needed partway through, which provides some slack but not much.

Room preparation, power and any construction coordination run in parallel and have external dependencies, which means they set the real completion date more often than the content does.

Testing and staging happen only when something exists to test, and they are the tasks compressed when everything else slips, which is exactly backwards because they are what the visitor encounters.

What happens when items are missing

No agreed objective: production proceeds on the loudest opinion and the review is a negotiation rather than an approval.

No usable model: the largest cost appears mid project, after the budget has been approved for something else.

No room survey: a format is proposed that the space cannot hold, and the discovery happens during installation.

No named operator: the experience is delivered to nobody, and it stops being offered within weeks.

No support arrangement: it works until it does not, and then it stays broken because nobody owns it.

Every one of these costs more to fix than the original item cost to obtain, which is the entire argument for settling them first.

Frequently asked questions

What is the first thing to decide?

One sentence describing what a visitor should understand or feel afterwards that they did not before. It decides sequence, detail, length and what counts as finished, and it is the only mechanism that resolves taste disagreements between stakeholders.

We have an architect model. Is that enough?

Usually not. A coordination model is built for construction and carries detail no viewer sees, so converting it into something a real time experience can display is production work rather than an export. Confirm what state it is in before requesting quotes.

What has to be measured in the room?

Ceiling height, clear floor area, projector throw distance, available power and how much daylight enters and when. Ambient light is the item most often waved away and the most likely to ruin the result.

Should the experience depend on the network?

For anything in a sales environment, no. Run it locally and use the network for updates and analytics only. Experiences that stream content during a presentation are hostage to a connection nobody in the room controls.

What should be agreed about ownership?

That the project owns the source content, meaning the model, materials and assets, rather than only a compiled experience inside a vendor system. Negotiating leverage for that clause disappears once the first invoice is paid.