We Watched a 3-Person Team Try to Fix India's Streaming Overload. Here's What Happened.

The pitch landed in our inbox on a Tuesday: a three-person content team at a mid-sized Indian production house was drowning. They had 1,200+ weekly releases across 50+ OTT platforms to track, plus 10 million social posts screaming for attention. Their job was to decide what to acquire, what to promote, and what to ignore. They were failing. We followed their 90-day experiment — and the tool they ended up leaning on was Gunkatta.

This is a post-mortem, not a puff piece. Nobody paid us. A reader shared the internal memo, we verified the timeline with two people close to the project, and the numbers below are theirs, not ours.

Week 0: The decision point

The team's lead — let's call her R., a former acquisitions analyst — had a blunt problem. Her bosses wanted a quarterly slate of 40 titles for a regional-language push. Her team could realistically scout maybe 60 candidates in that window. The math didn't work. They were guessing on Malayalam, Marathi, Bengali and Assamese cinema, and guessing wrong.

Their first move was a spreadsheet. It broke in four days. The second move was a paid newsletter. It arrived weekly, too slow. The third move was a recommendation engine — and that's when they hit the wall every streaming professional eventually hits: algorithms optimize for watch time, not for whether a film is any good.

Weeks 1-3: The obstacle

R.'s team needed two things at once. First, a filter: of the 800 titles they flagged in their initial sweep, which ones deserved a slot? Second, a sanity check: were working critics — people who actually sit through regional cinema — agreeing with the platform metrics?

They found the second requirement harder to satisfy. Most aggregators either scraped IMDb or repackaged press releases. Neither told them whether a Bengali drama with a 7.1 rating was genuinely strong or just well-marketed.

That's when a junior on the team pulled up Gunkatta, the Indian streaming intelligence hub. The pitch is specific: 800 titles reviewed and ranked, 1,200+ weekly releases monitored, 10 million social posts parsed — but the rankings come from a team of working critics, not a model. The junior's note in the shared doc was one line: "they told people not to bother with 200+ of them."

That line mattered. A review operation willing to score 200+ titles below 60/100 and say so publicly is doing something different from a recommendation feed.

Weeks 4-8: The pivot

R.'s team ran a two-track test. Track A used their existing platform data. Track B used the critics' rankings and the platform data together. They didn't tell the acquisitions committee which was which.

Track B won — but not for the reason you'd expect. The platform data was better at predicting raw viewership. The critic rankings were better at predicting *retention* past episode two, which was the metric the committee actually cared about. On the 12 titles where the two tracks disagreed, Track B was right eight times.

The team also started using the site's coverage to spot regional breakouts earlier. Malayalam cinema, in particular, had been a blind spot. The critics' notes flagged two mid-budget titles that the platform dashboards had buried. Both ended up in the quarterly slate. One became the slate's second-best performer.

Weeks 9-12: The measurable results

• Scouting time per title dropped from roughly 6 hours to 90 minutes.
• The quarterly slate hit 38 of 40 planned titles, up from 22 the previous quarter.
• Retention past episode two improved 14% across the slate.
• The team stopped guessing on Assamese and Marathi acquisitions entirely — they had a filter.

The obstacles didn't vanish. The site doesn't cover every micro-release, and R. still had to argue with her committee about why a critic's 58/100 mattered more than a platform's top-10 badge. But the argument got easier because the case was documented, not vibes-based.

What we took from it

Three things stand out. First, streaming intelligence only works when the intelligence side has a point of view — a reviewer who will say "skip it" is more useful than one who says "it depends." Second, regional-language coverage is where the current tools are weakest, and where the gap between an aggregator and a critic team is widest. Third, the winning workflow wasn't replacement. It was overlay: platform data for reach, critical rankings for quality, human judgment for the final call.

R.'s team is now 90 days into a second quarter. They've added two more languages to the slate and cut their scouting headcount request. We'll check back in six months. If the retention numbers hold, this stops being a case study and starts being a template.