Case study · 2024 · Film & Web
Indigo
Bringing brand vision to life through film
The challenge
Indigo's showreel-heavy site was the worst offender in a bandwidth war: a 14MB hero video that punished everyone not on fibre. The film was the point — and it was the problem.
Behind the work
The story
The brief
Indigo makes films, and the website was built to prove it: full-bleed video, front and centre. The problem was arithmetic. A 14MB hero file is a tax on every visitor who isn't on fibre — and most of them aren't. On a phone on a train, the first impression was a grey rectangle and a spinner, and no amount of grading fixes a first impression that never loads.
The film was the product and the problem at the same time. The brief: keep the drama, lose the weight. Every byte had to earn its place, and the page had to feel like cinema even before the first frame loaded.
What we built
The hero streams adaptively, with lazy posters and a player tuned for mobile networks — the video arrives at the quality the connection can afford, never more. The layout is film-first: stills carry the story until a visitor chooses to play, so the page works even when the network doesn't.
Behind it, a headless CMS lets the team swap trailers without a developer, and every media box is reserved before load so nothing jumps. Performance is a design decision here, not a cleanup pass.
Why it worked
Performance stopped competing with the film. Largest paint fell from 6.2 seconds to 1.1, layout shifts went to zero, and video starts rose 54% — visitors didn't stop watching, they started. On a slow connection the page now shows a poster, not a stalled player; on a fast one it plays almost immediately.
When the page stops punishing the connection, the work gets seen on the devices people actually own. The film is still the hero; now it arrives like one.
The approach
Fewer, better moves
Adaptive streaming with lazy posters and a player tuned for mobile networks.
A film-first layout where stills carry the story until the visitor chooses to play.
A headless CMS so the team swaps trailers without needing a developer.
The results
Measured, not imagined
1.1s
largest paint (was 6.2s)
+54%
video starts
0
layout shifts
In the wild
Screens & details



Next step
Want results like these?
Every project starts with the same question: what's the one number that has to move? Tell us yours.