Who Controls Discovery?

Same Build, Different Launch Day

In a corner of a shared office in Seomyeon, Busan, two people are making a game. One codes, one does art. They export builds in Godot, layer AI-generated images as sprites, and match color contrast and mood words exactly as 2025 award winners did. Looking at the build file alone, it's indistinguishable from what a two-person studio in Austin, Texas would produce. Indeed, indie game production quality has leveled off these days. Shaders, particles, UI polish, sound—the gaps once determined by studio size have been filled by asset stores and generative models.

But the moment they press the launch button, the two games fall into entirely different worlds. The Austin studio's game has been accumulating wishlists for six months before release. Steam's algorithm reads that wishlist curve and allocates visibility on the 'Coming Soon' and 'New Releases' pages on launch day. The Busan studio's game appears on the same pages on the same day, but with an empty wishlist, the algorithm gives it almost no exposure. Same build, same price, same genre—yet first-day traffic splits by an order of magnitude.

Here's where the familiar diagnosis appears: Korean indies don't understand marketing, their English is weak, they lack global sensibility. A narrative that attributes everything to individual founder deficiencies. I believe this diagnosis is wrong. The two people in Busan already built with the world in view. The real problem is the structural position: they enter as outsiders into the layer that makes their game visible to the world—the discovery layer.

The Bottleneck Is Discovery, Not Production

Breaking down the game industry value chain into segments reveals which parts have commoditized and which still collect tolls. The segments from planning through production to polish have rapidly modularized over the past decade. Engines became free (Godot, Unity personal licenses), art prices dropped with CC0 asset packs and generative models, and build and distribution happen with one click. These segments no longer generate significant rents—because anyone can enter.

The rent has shifted to the next segment: discovery. Steam's recommendation algorithm, wishlist graphs, curator networks, the first 24 hours of exposure controlled by influencers, preview curation by foreign media outlets. This segment hasn't modularized. Instead, it has condensed. As dozens of games pour out daily, discovery became a scarce resource, and the algorithm distributing that scarce resource looks at a single input: data on interest already accumulated before launch.

This raises the key question: where does that data come from? Wishlists are built by people who watched pre-launch trailers. So where do they watch trailers? English-language Discord servers, genre subreddits on Reddit, verified curators' recommendation lists, English-speaking streamers who received pre-launch builds. All these channels are closed within English-speaking gamer communities. A Korean founder's ability to make good games and their ability to accumulate trust capital within these channels are different kinds of resources. The former is controlled by individuals, but the latter is supplied by the ecosystem.

The Austin studio essentially inherits this data for free. Other indie developers in the same city, former coworkers, streamers met at local game jams, a publisher run by a school alumnus. The entire weak-tie network built while growing up in English-speaking communities converts into pre-launch wishlists. The two people in Busan lack that network—not because their game is inferior, but because they're outsiders to the ecosystem that produces discovery inputs.

Global Ecosystems Cultivate Discovery as Capital

The decisive point where American and European indie ecosystems diverge from Korea's is that they cultivated discovery not as individual luck but as capital and institution. The very job category of indie publisher is discovery agency work. Companies like Devolver Digital, Annapurna Interactive, and Raw Fury don't make games. What they sell is discovery: verified curator networks, foreign journalist contacts, streamer pools, relationships with festival judges. Indie developers give up a portion of revenue to borrow this discovery capital. Not extraction, but one form of creating and selling discovery.

Festivals and awards layer atop this structure. IGF, BIC, various indie showcases aren't mere award ceremonies—they're discovery funnels. Win there and curators see it, publishers make contact, and that trust flows into wishlists that Steam's algorithm reads. The path where a student work first gains recognition at BIC and connects directly to publishers is exactly this. American and European founders grow up knowing where these funnel entrances are. Korean founders only learn that funnels exist after launch. Though facing the same funnel, one side meets it at the entrance, the other at the exit.

Here someone will strongly object: talk of discovery is just excuses, and truly good games rise on their own. Didn't Balatro, a solo developer's unknown game, conquer the world through word-of-mouth alone? Valid point. And precisely at that point, the structure reveals itself. Balatro's developer maintained anonymity but circulated builds within English-language Discord and Reddit, and English-speaking curators who discovered it lit the first spark. The game's quality exceeding the threshold is true. But the discovery layer that first witnessed and amplified that quality was the English-speaking community. Even if a game of equal quality emerges from Korea, if it can't reach that initial witness pool, it never reaches the threshold at all. The proposition that good games break through is true—but only for games that have already entered the discovery layer. Games that can't even enter aren't granted a threshold to break through.

Four Bottlenecks: Capital, Talent, Regulation, and Customer Access

The structural position of standing outside the discovery layer splits into four concrete bottlenecks. First is capital. Korean game investment is optimized for reading mobile in-app purchase revenue curves. Evaluators who assess discovery signals like pre-launch wishlists, foreign media ratings, and curator reactions as assets are rare. So indies making PC and console packaged games have nowhere domestic to borrow discovery capital, and fill that gap with personal savings and labor.

Second is talent—precisely, not the absence of talent but the disconnection of talent networks. Korea overflows with excellent developers. What's scarce are people routinely connected to English-speaking curators, streamers, and foreign journalists. These connections can't be quickly purchased through learning—they're accumulated trust over time. Third is friction in payment, exchange rates, and taxation, more fearsome than regulation. Administrative costs like global store settlements, VAT, and foreign remittances burden a two-person studio equivalent to one full-time person. American indies start with this infrastructure as a taken-for-granted default.

The fourth and deepest bottleneck is customer access. The discovery layer being closed within English-speaking communities means the cost of meeting first customers is structurally more expensive for Korean founders. When gathering the first 1,000 wishlists with the same game, the Austin studio reaches them nearly free through networks, while the Busan studio pays costs converted to advertising or local agent fees. This asymmetry doesn't end with one game. The discovery failure of the first game erodes the wishlist foundation for the second, and since algorithms look at cumulative data, gaps widen compound. American indies accumulate discovery capital with each release; Korean indies restart from zero each time. The real bomb is here: a structure where not cash but discovery data fails to accumulate.

The Ecosystem Must Keep Pace with Founders

So the condition for Korean indies to go global doesn't lie in whipping individuals harder. It lies in building a structure where the ecosystem shares the cost of entering the discovery layer. Specifically, three things. One, capital that cultivates discovery agency as a job category. Korea needs indie publishers and evaluators who assess and bet on pre-launch discovery signals as assets, not revenue. Two, shared connection infrastructure that pushes Korean games into English-speaking curator and streamer networks and festival funnels—pooling the inefficiency of individual studios each breaking through from zero via cooperatives or collective publishing. Three, standardizing payment, tax, and remittance friction so two-person studios don't grind a full-time person into administration.

From the coordinates of Busan, this isn't abstract theory. The local indie scene already has a discovery funnel entrance called BIC. The problem is that entrance remains a once-annual event, and the standing pipeline connecting winners to English-speaking curators and publishers afterward is weak. The next task is connecting the festival not as a point but as a line—as standing discovery infrastructure.

If you're a reporter covering the Korean game beat, you can change one question starting tomorrow. When covering a new title from a Korean studio, instead of asking 'Is this game well made?', ask 'Who, in which community, built up this game's wishlists?' Trace where discovery inputs were produced, and the answer to why well-made Korean games remain invisible is all there. Founders already build with the world in view. Now, so they're visible to the world, the ecosystem must keep pace with them.