The Discovery Bottleneck

While Sanabi Sells Well

Korean-made titles like Sanabi, published by Neowiz, have begun appearing on Steam's global charts. Devsisters has expanded its indie lineup by combining in-house production with external discoveries since Cookie Run. Critical reception has been positive. Viewed this far, it's a good scene for the Korean game industry.

But when you look at this scene a little longer, anxious corners emerge. It's not that well-made games sold well, but rather that among well-made games, the ones lucky enough to be caught by Steam's algorithm and curation sold. Dozens of games with the same level of completion are pushed outside the Discovery Queue in their first week and disappear without a trace. Behind the one success we applaud, there are ninety-nine that went undiscovered.

This is not a matter of luck. It's a matter of which nodes Korea has and which it doesn't.

Luck Cannot Explain the Next Title

The most common perspective reads the Korean indie surge as the coincidence of individual hit titles. It's a comfortable but wrong reading. Because if it were coincidence, there would be no way to prepare the next title.

The reason globally viable indies are emerging from Korea now is that planning and art, engines, and live operation know-how have accumulated in people's bodies and studios' practices through the 1990s package era, the 2000s online game era, and the 2010s mobile era. Developers who passed through Nexon and NC become independent and form small teams, and the output those teams create exceeds the average quality of the global market. The accumulation of production capability has already happened. At this point, Korea has finished catching up.

The problem is a different node. The power to create has been localized, but the power to display those games before the eyes of hundreds of millions of gamers worldwide still belongs to others. Steam, a single gateway, holds the discovery bottleneck, and Korean studios are merely applicants standing in line before that gate.

Why Steam Made Distribution Free

Here, people who see structure change the question. Why did Steam make it so anyone can upload games? It wasn't goodwill.

Valve deliberately modularized and commoditized the distribution node. Anyone can upload a game for just $100. Distribution became near-free infrastructure, and as a result, tens of thousands of games pour out every year. When distribution was unleashed, the real scarce resource moved up one level. Discovery. The more games overflow, the more the value of visible positions skyrockets. Instead of abandoning the toll on distribution, Valve seized a 30 percent commission and effective power of life and death at a higher node: the discovery algorithm and recommendation slots.

This is the core of value chain reorganization. When one node becomes commoditized, power and profit move to the still-scarce nodes above or below it. The moment Korean studios rejoiced that global launches were now possible for everyone, the real bottleneck had already moved up to the discovery domain. Launches became easier and discovery became harder. What we localized was the node that had become easiest.

Advanced Countries Accumulated Discovery as an Industry

In the era of catch-up, Korea won with the proposition that if you make it well, it sells. Semiconductors and automobiles too—if you made better products cheaper, the market bought them on its own. Production capability was competitiveness itself.

In the frontier era, this proposition breaks down. Making well is just an entry ticket, and being discovered is a separate capability. Moreover, discovery is not bought with a single marketing campaign but a system built over time. Data and field sense about what text increases wishlists, what first three seconds of a trailer lowers refund rates, what timing of discounts boosts algorithmic exposure. This only accumulates through trial and error.

What American and European indie publishers have is precisely this accumulation. They failed hundreds of games and grasped discovery patterns as data. Korea accumulated 30 years of production-side trial and error, but accumulated almost no global discovery trial and error. That Devsisters and Neowiz entered publishing should therefore be seen not as simple business diversification, but as an instinctive movement to fill an empty node.

Here we must face an honest counterargument. Isn't discovery not an industrial capability but just an ordinary problem where good games spread through word of mouth? It's half right. Overwhelming masterpieces break through bottlenecks. But an industry runs not on one masterpiece but on a hundred decent and good titles. The ability to break through with one stroke of heavenly luck and the ability to stably enable discovery of a hundred titles are completely different resources. Industrial policy must take responsibility for the latter.

What Korea Must Accumulate

So what Korea must build is not the next hit title but discovery infrastructure.

First is making global publishing data a shared asset. The structure where individual studios each collide with Steam's algorithm and repeat the same mistakes is wasteful. A node is needed where data like wishlist conversion rates, trailer retention, and regional price elasticity accumulate and are shared among studios. Next is a workforce that makes discovery a profession. There are enough developers who create. What's lacking are publishing specialists who design global launch strategies with data. Last is diversification that reduces platform dependence. As long as everything is bet before Steam's single veto power, the toll flows out forever.

Look at Busan. G-Star is held every year, but that's an exhibition venue, not discovery infrastructure. Exhibitions end in four days and data doesn't remain. If Busan becomes a hub that year-round gathers and shares indie studios' global publishing data, and thus creates a structure where one studio's failure becomes the next studio's asset, the coordinates shift from an exhibition city to a city that accumulates discovery capability.

If you're a reporter writing the game beat tomorrow, engrave this next question: Did this Korean game sell because it was well-made, or who held which node of discovery and collected the toll? Where did that toll flow out to?

In the era of catch-up, countries that found answers quickly won. Making well was enough. In the frontier era, countries that create questions first win. Korea solved the question of making. Now it must create the question of discovery before anyone else.