Table of Contents
- Two States, One Theatrical Domain
- Trigger Events Beat Calendar Locks
- Four Arrivals Before Any Wireframe
- Cinematic Shell Versus Campaign System
- Registers, Stall Points, QA Hours
- Night Circuit in Three Phases
Two States, One Theatrical Domain
Theatrical briefs arrive with two campaign states already named: an initial teaser, then a full immersive-world launch.
Drawing on Exopolis dual-state launch files, the teaser window still opens 12 to 16 weeks before premiere. That span is long enough for a disposable URL to go stale and short enough that a domain migration at launch would punish returning traffic. The first architecture decision is to keep a single primary domain and let the experience thicken in place.
Controlled expansion is the real brief. Returning visitors should find more rooms. Late arrivals from a trailer drop should still land on a coherent first screen. The site grows by revealing modules while the frame stays put.
Scope here is structural planning for theatrical campaign microsites. Visual direction, framework selection, and backend code sit outside this blueprint.
The dual-state model assumes that single primary domain. Social-first hubs with disposable landing pages need a different structural model, and that limit should be named before anyone sketches a homepage.
Trigger Events Beat Calendar Locks
Transitioning from teaser to immersive launch means toggling visibility on 15 to 25 distinct content modules, including character profiles, expanded galleries, and interactive story nodes. That range is a planning envelope. It tells the team how many doors the state logic must own.
Early phase matrices tried to pin those doors to calendar dates. Trailer drops slipped 48 to 72 hours on late-stage editorial approvals, and the dates became fiction. The matrix now runs on sequential trigger events. Someone with clearance toggles the state when the media is actually cleared.
Name four phases even when a given campaign uses fewer: teaser, reveal, immersive launch, and post-release maintenance. Count and timing follow the title, not a template.
For each phase, fill five fields:
- available assets
- primary audience question
- desired action
- visible routes
- material that must stay hidden
Durable content crosses states. Synopsis and credits belong there. Time-sensitive modules rotate in and out: countdowns, premieres, ticket prompts, launch announcements.
Trigger After Approval
Bind the immersive switch to cleared trailers, signed copy, and a producer who can flip the state. A calendar cell cannot clear legal.
Four Arrivals Before Any Wireframe
I map four arrivals on paper before any art direction meeting. Screens come after journeys, or the homepage tries to do every job and does none of them well.
First-time visitors from a teaser placement arrive with almost no context. Entry condition: a paid or owned teaser unit. Orientation cue: title treatment plus one atmospheric loop. Primary action: a single save or notify control. Optional path: a muted credits peek at most. Exit: the same URL they will reuse.
Returning fans hit a saved URL. Orientation should match the live state, with persistent navigation once routes have opened. Primary action: continue from the last visible chapter. Optional path: gallery. Return origin: the homepage, kept stable on purpose.
Trailer-driven traffic wants the player. Cue it. Watch is the primary action. Characters can wait until after the cut. Release information is the clean exit.
Search traffic hunting dates and tickets needs practical release information in the first viewport. Outbound to ticketing is the action. Synopsis is optional. They leave when the errand is done.
Trailer, story, characters, gallery, credits, and ticket information need persistent destinations when the campaign warrants them. The homepage orients. It does not warehouse the whole campaign.
Hold Locked Routes
Deep links shared in the teaser phase need redirect logic that holds for 8 to 10 weeks. Traffic to locked routes defaults to the teaser homepage until the immersive state is triggered.
Cinematic Shell Versus Campaign System
One approach rebuilds the visual wrapper when the world launches. Another keeps a cinematic shell and swaps campaign logic underneath. Keep the shell. Change the system.
Break the microsite into four layers: stable route structure, campaign-state logic, reusable content modules, and the motion or audio presentation layer. Routes exist early. Visibility is a flag. Navigation can survive the jump from a restrained teaser to an expanded world because the destinations were never improvised at launch.
Define modules once, then show or hide them by state: hero media, trailer player, story copy, character profiles, galleries, credits, release information, legal notices, and calls to action.
The cinematic presentation layer needs a baseline asset payload of 4 to 6 megabytes for initial atmospheric video loops, optimized to load within around 2.5 to 3.5 seconds on standard broadband before fallback stills display. That payload lives in the shell. Campaign copy does not.
Exopolis, working as a digital creative agency at the intersection of interactive storytelling and motion-led campaigns, treats the split the way sequential drops once worked in Xbox Kinect Fun Labs: the container stays, the surface area grows.
Registers, Stall Points, QA Hours
The final QA handoff needs a dedicated 72 to 96-hour window strictly for testing state transitions across target devices. That window exists so the teaser-to-launch switch does not break persistent navigation or mute audio controls. Treat it as a locked calendar block, not leftover afternoon time.
Convert the architecture into an asset register covering trailers, stills, copy, logos, billing blocks, audio, captions, legal language, release information, and approved calls to action. Then mark the dependencies that stall launch: final media delivery, content approval, route activation, metadata, performance checks, and device testing.
When an acquiring agency such as McGarrah Jessee inherits a live campaign, the register is the document that travels. A SunnyD-style digital agency of record relationship makes the same point from the client side: billing blocks and legal language sit on the critical path whether the product is a beverage or a feature.
Measure trailer starts, completed transitions, section visits, gallery engagement, outbound release actions, and repeat visits. Leave the targets to the campaign team. This blueprint does not invent them.
Handoff The Register
If captions or an approved call to action arrive inside the QA window, the immersive toggle waits. Launch day is a dependency graph.
Night Circuit in Three Phases
Copy this against a clearly fictional thriller, Night Circuit, so the sequence can be reused without being mistaken for a documented release.
Step 1. Establish stable routes on day one: homepage, story, characters, gallery, and release information. For Night Circuit that is /, /story, /characters, /gallery, and /tickets. Most of those routes stay unpublished. Unpublished paths redirect to the teaser home. The route table exists before any public media does.
Step 2. Launch the teaser state with one atmospheric media loop, the title treatment, a release cue ("Winter 2026"), and one primary action: Notify Me. The loop is the 4 to 6 megabyte shell payload, with a still fallback if the load misses the 2.5 to 3.5 second window. Unavailable deep links, including a fan-shared /characters/lena, return to the teaser. No chapter nav. No secondary audio. No tickets.
Step 3. When the trailer clears editorial, a producer toggles immersive launch. The system activates 6 to 8 new chapter-level navigation nodes and a secondary audio track, swapping primary presentation payloads without a hard browser refresh. Character modules, the expanded gallery, and story nodes come up among the 15 to 25 visibility toggles. /tickets becomes the primary action. Persistent nav never restyles itself into a new information architecture; it only unlocks destinations that were already named in Step 1.
The payoff of Step 1 shows up in Step 3. That fan-shared /characters/lena, bounced to the teaser home for weeks, resolves to Lena's profile the second the producer flips the state — same URL, no redirect edit, no rebuild, no refresh.
