They’re dubbed duanju, or microdramas – and they’ve become big business.
Author
- Johnny Chan
Senior Lecturer, Faculty of Business and Economics, University of Auckland, Waipapa Taumata Rau
In China, the smartphone-friendly shows reportedly generated US$7 billion in 2024 , overtaking the country’s domestic box office for the first time. Their audience numbers in the hundreds of millions, while the format is also growing rapidly in other markets, including the US .
Here’s how they work. Episodes typically last about a minute and play vertically on a smartphone, while a season can stretch across dozens of instalments.
Many compress the heightened plots of soap operas – romance, betrayal, revenge and improbable wealth – into bursts of drama where something is almost always about to happen.
Early episodes are often free, with cliffhangers used to keep viewers watching before they’re eventually asked to pay to see what comes next.
Generative artificial intelligence is now accelerating their production, with Chinese platforms reportedly releasing hundreds of AI-made microdramas each day .
So, what is it that makes this form of entertainment so enticing? In my recent research , I modelled how viewers, algorithms, creators and platforms interact – and how that system can make it increasingly difficult for people to disengage.
What keeps us from tuning out
We might not think much about why we stop watching or reading something, but traditional media give us plenty of opportunities to do so.
Some come naturally from us. Repeated experiences tend to lose their effect as novelty wears off – a process psychologists call hedonic adaptation . Others are built into the medium itself. A television programme ends, a book chapter finishes or a newspaper runs out of pages.
But microdramas have the potential to weaken both these stopping points – one reason their rapid growth has raised concerns about unintended binge-watching .
Catching another one-minute episode feels like a small commitment, while ending an episode at a moment of suspense encourages viewers to continue. Someone may never decide to spend half an hour watching but instead make a series of smaller decisions to watch for one more minute.
Microdrama platforms can also learn from viewers as they watch. Recommendation systems track signals such as whether someone finishes an episode, replays or shares it, or moves straight to the next one.
Older media had relatively large units of measurement – an episode, article or song. On short-video platforms, even a swipe can help determine what is recommended next.
Creators receive feedback too. Episodes and scenes that hold attention can get greater distribution, encouraging producers to repeat techniques that work. More engagement also brings in more revenue, which platforms can spend on better recommendations and new content.
The model developed in my research shows how these processes can reinforce one another: viewing generates data, recommendations become more targeted, successful techniques are repeated and greater engagement funds further investment.
It’s within this cycle that generative AI could have a significant impact.
New characters, settings and story ideas can be produced much faster, allowing platforms to keep offering variations on familiar formulas before viewers lose interest in them.
This is still a hypothesis rather than something my research has demonstrated experimentally. Real-world viewing data is needed to test how strongly these effects operate.
Can the cycle be broken?
For those of us more accustomed to conventional TV shows such as Slow Horses or The White Lotus, microdramas may still be unfamiliar territory. But they are rapidly moving into the mainstream.
A 2025 US survey found about 45% of consumers were familiar with microdramas or micro-series, while major streaming and social-media platforms are increasingly experimenting with short-form vertical video.
What happens if their growth brings concerns about excessive viewing? Screen-time limits, age restrictions and warning labels are already used to curb digital overuse. But these largely put the responsibility on viewers.
Platforms themselves could instead provide clearer opportunities to stop, such as a pause before the next episode begins rather than loading it immediately.
Payment mechanisms could also be made clearer. Viewers buying virtual coins to unlock episodes – as is already common on microdrama platforms – could be shown the real-world cost of each episode and asked to confirm the purchase.
As well, greater transparency around recommendation systems, along with access for independent researchers and regulators, could help reveal how platforms encourage continued viewing.
A more fundamental approach would be to reconsider how heavily recommendation systems reward measures such as viewing time and how often users return.
There is still limited evidence about how well any of these approaches would work for microdramas. But their growing popularity signals a shift in digital content, which can now be produced, tested and recommended to viewers faster than ever before.
