What We Measure Next

Cricket has two ways to look at a batsman (yes, I prefer batsman over the horrible-sounding ‘batter’). There's the average, which tells you how many runs they scored. And there's the strike rate, which tells you whether those runs changed the game.

Indian streaming has, for the last decade, mostly cared about its average. The next decade is going to be about its strike rate. And the substitution is bigger than it sounds.

For the first 10 years since its advent, Indian streaming has been a distribution game. Smartphones got cheaper. Data got cheaper. A new audience came online every quarter, and the job was to get them watching. We measured everything in views and watch-time, and we were right to. Every additional minute on platform was a minute pulled from television, UGC platforms, gaming, or the famous sleep. The number going up meant the strategy was working.

That game is mostly over.

The average Indian smartphone user now spends several hours a day consuming video across multiple platforms. There is little untouched attention left to win. Most platforms that crossed the threshold of utility already have an audience. The question now is whether they come back tomorrow, and the day after, and whether they value those hours enough to defend them when something newer shows up.

This is where Goodhart's Law quietly catches us. When a measure becomes a target, it stops being a measure.

Watch-time worked beautifully for a decade because it was a proxy for engagement. Then we learned how to optimize it. Auto-play the next episode. Auto-play the next show. Surface a trailer between episodes. Trigger a notification about something happening right now.

All of these can increase watch-time. Whether they increase long-term habit is a more complicated question. The dashboard may say we're winning. The retention cohort, three quarters later, may tell a different story.

Watch-time isn't a bad metric. It built this industry. It will remain important for years. The shift isn't about abandoning it. It's about putting other measures alongside it, so we stop optimizing around the one number that has started to tell us less than it once did.

The reason this matters goes beyond analytics.

Entertainment businesses are unusual. We don't manufacture products. We manufacture affection.

The value of a great show isn't just the hour someone spends watching it today. It's whether they remember it, talk about it, recommend it, return for the next season, and choose one platform over another when time is scarce.

The measurement layer should reflect that reality.

Many mature consumer platforms eventually discover that total consumption is an imperfect proxy for long-term value. Several have evolved toward measuring the quality of engagement rather than simply the quantity of it. The exact metric differs by company, but the underlying question remains the same:

Are users merely consuming, or are they building a habit?

Indian streaming may have an opportunity to make that shift faster than most. Almost every major streaming service in India operates a hybrid model, with advertising and subscriptions sitting inside the same product. That's unusual globally.

The hybrid matters because it aligns incentives. Subscription businesses naturally care about retention, habit formation, and long-term value. Advertising businesses care about attention. When both sit inside the same product, one side can help discipline the excesses of the other.

The result is a chance to build a richer measurement framework than either model creates alone.

What might that framework look like?

Not a replacement for watch-time. A layer above it.

Completion by intent. Did users finish what they came to watch?

Return cadence. Do they come back organically, or only after being prompted?

Recommendation acceptance. Do they trust what the platform suggests next?

Ad tolerance. Do they remain engaged through the commercial experience, or does the experience create abandonment?

None of these metrics are exotic. Most analytics teams already track some version of them. The question isn't whether they exist. It's whether they sit in a forgotten corner of a dashboard, or whether they sit at the centre of how product, content, marketing, advertising, and commissioning decisions get made.

Without this measurement layer, the next few years of AI-accelerated production, recommendation, personalization, and advertising will feel like navigating by dashboard light alone. With one, you can let everything else speed up without losing track of where you're actually trying to go.

When these measures move closer to the centre, interesting things begin to happen.

Recommendation systems become more accountable for bad suggestions.

Advertising becomes more focused on attention that actually matters.

Content investment decisions become less about generating volume and more about generating lasting audience relationships.

The math isn't that difficult at all. The measures already exist inside most data warehouses. The hard part is cultural.

Watch-time is wired into the way the industry talks about itself. Ad sales pitches quote it. Programming reviews start there. Executive scorecards revolve around it. Analyst conversations gravitate toward it. Moving beyond it requires leadership teams willing to look slower on the industry's favourite number for a few quarters, with the conviction that something more durable is compounding underneath.

Most won't make that trade. The ones that do will likely define what Indian streaming looks like over the next five years.

Watch-time, in the end, is the average. The next chapter of this industry is going to be about the strike rate.

The teams that learn to read both, and know when the second matters more than the first, are the ones that will keep batting through the next innings.