Almost every study of the Indonesian market is a study of Jakarta. Ours started that way too — 55.9% of our interview base came from Tier 1. That imbalance is the finding, not a footnote to it.

In 2025 we ran a twelve-week AI-enabled research program for Warner Bros.’ Southeast Asia division, testing a 2025 theatrical slate against local audiences. The problem the studio brought us was blunt: local films were dominating the box office, the release calendar left almost no slack, and there was limited local bandwidth to go find out why.

Why the tiers are not interchangeable

A market split by city tier is not a split by income. It is a split by:

• Which films are physically playing, and for how many weeks

• Which language a trailer arrives in, and whether it is subtitled or dubbed

• Who decides what a group watches — friends, family, or a partner

• What a ticket costs relative to everything else that evening

  
  

Treat those as one audience and every recommendation you write is a Jakarta recommendation with a national label stuck on it.

What a condensed timeline does to research

• You default to the cities where fieldwork is easy

• You interview in English because it is faster to analyse

• You take the sample you can get and call the skew a limitation

• You ship a finding that is true of 56% of your base and false outside it

How we ran it instead

1. Aligned the whole program to the studio’s release calendar, not our own

2. Ran 200+ qualitative interviews in Indonesian through Sayle, our own AI interviewer

3. Kept the tier split visible in every cut of the data — Tier 1 at 55.9%, Tier 2 at 34.7%, Tier 3 at 9.4%

4. Wrote recommendations by region and demographic instead of one national answer

The Indonesian-language interviewing mattered more than anything else on that list. Nuance does not survive translation, and the reasons people give for skipping a film are almost entirely nuance.

What we handed over

• Bilingual executive summaries, so the regional and global teams read the same thing

• Data visualisations built for people who were not in the interviews

• Region- and demographic-specific recommendations

• Localized go-to-market tactics, not a global playbook with the names swapped

Bilingual was not a courtesy. A finding only the regional office can read never reaches the people setting the slate.

The part most studies skip

We reported our own skew. The base was 45.8% aged 18–23, 55.1% female, 58.5% students, and heavily Tier 1. Which means:

• Every Tier 1 read is well supported and can be trusted

• Every Tier 3 read is directional and should be treated as a hypothesis

• The gap between them is where the local-film advantage actually lives

• Closing that gap is the next study, not a rounding error in this one

A client can act on that. Nobody can act on a national average that hides which half it came from.

Final thought

If local films are winning, they are winning hardest in the cities that show up as 9.4% of everyone’s sample. That is not a coincidence — it is why the gap keeps holding. Research the places that are hard to reach, or keep being surprised by them.