Gamescom 2026 proved the games industry's fragmentation problem is getting worse

Gamescom 2026 broke records for exhibitors: 1,743 companies from 66 countries, up 11% on last year. The show floor sold out for the first time in the event's history. Total attendance landed at 368,000 visitors from 131 countries, a 6% rise on 2025 (though still short of the 373,000 who showed up in 2019, so not every record fell).

Read as a trend line, this article isn't just to say "wow, the show got bigger." It's the games industry adding more platforms, more regions, more storefronts and more ways to distribute the same game, every single year, and doing it faster than any single data team can keep up with. Put plainly, game catalog fragmentation is scaling with the industry.

We spent Gamescom week in Cologne, in and around the halls, talking to the platform, catalog, ad tech, measurement and AI teams actually building on top of that growth. Different job titles, different stands, different agendas. But strip away the vocabulary and most of the conversations were describing the same operational problem from a different department.

The game catalog fragmentation problem wearing different names

Ask a catalog team what's hard and they'll say metadata normalization. Ask an ad tech team the same question and they'll say targeting granularity. Someone on a connected-TV team will say onboarding friction while someone building an AI product will talk about the context layer.

Those are four different ways to talk about the same problem: a game's metadata, pulled from more than one source, stops agreeing with itself. The platform ends up holding several versions of the same game catalog entry and has to decide which one to trust.

Catalog and discovery teams.

The connected-TV and gaming-hub platforms we spoke with are aggregating titles from multiple external partners into a single browse and search experience, and the friction shows up exactly where you'd expect: two providers list what is obviously the same game under different names, different genres, or different availability windows, and somebody has to reconcile the metadata by hand before the platform can trust it enough to recommend it. The teams furthest along on this aren't asking for more raw data. They're asking for one normalized source of truth to sit underneath the game catalog they already have, because every new provider they add multiplies the reconciliation work rather than adding to it in a straight line.

Ad tech and audience teams.

A separate set of conversations, mostly on the advertising and audience side, kept circling a narrower version of the same complaint: genre and merchant-level tags aren't just too blunt, there's no shared standard behind them across the market. A sci-fi platformer classified one way on one ad marketplace and a different way on another isn't obviously the same audience to target, even though it's the same game. "Action" or "RPG" doesn't tell a media buyer what a player actually does in a game, which mechanics they respond to, or whether a title tagged that way on one platform means the same thing as a title tagged that way on another. Broad classification is a taxonomy problem wearing a media-planning costume, and it's the same shallow-metadata issue the catalog platforms above are fighting: without a shared standard behind the tags, the same audience can't be bought twice.

Measurement and ACR teams.

Connected-TV and automatic content recognition providers described a more technical flavor of the same gap: their detection models catch gameplay reliably in cutscenes and loading screens, and lose confidence everywhere else because the underlying game data thins out mid-session. Coverage that drops off isn't a rounding error. It's the same fragmented, incomplete metadata problem showing up as a measurement accuracy issue instead of a catalog issue.

AI and search teams.

The newest version of the conversation, from teams building search features, recommendation agents and LLM-backed assistants: generic models are confident about games they've seen a lot of text about, and much less reliable about everything else. What they need is to make games machine-readable for search, recommendations and AI-assisted workflows. Because there's no structured data layer underneath telling the model what a game actually is, how it relates to other titles, or which of five near-identical catalog entries a query is really asking about. Unstructured wikis and scraped text get an agent partway to an answer on a given game. They don't get it to a defensible one.

Four teams, on four different parts of the platform stack. One root cause: nobody at Gamescom this year had a metadata supply chain that scaled as fast as their distribution footprint did.

Why scale can make it harder for your games to reach the right players

You might be thinking: "doesn't a bigger show, with more exhibitors and more countries, just mean a bigger addressable market? Isn't fragmentation the price of growth, and a fine one to pay?"

It would be, if the cost were linear. It isn't.

Every additional catalog partner, region, or storefront a platform adds doesn't get bolted onto the existing system, it gets reconciled against every system already in place. A platform integrating a third data provider isn't doing three times the integration work it did with one. It's doing something closer to three times the work, squared, because each new source has to be checked against every source that came before it, not just added to the pile.

That's the part a record-breaking exhibitor floor makes visible in a way a quieter year wouldn't. 1,743 companies from 66 countries is 1,743 different ideas of what counts as a genre, an availability window, or a canonical title, and every one of those disagreements is another data point of fragmentation. Growth at that pace doesn't dilute the fragmentation problem. It compounds it, game by game, catalog by catalog.

That compounding effect shows up for players too, as discovery fatigue: bouncing from platform to platform because none of them seem to get what they're actually in the mood to play. Search and recommendations only feel personal when they're built on mood, mechanics and motivation instead of genre, novelty or popularity. A platform that can't speak a player's language can't build filters, curate collections collections or launch ads that keep that player around, so the player goes looking somewhere else.

Structured game metadata closes the gap

None of the teams we spoke with wanted another dashboard. Evaluators want to see structured metadata depth and curated catalog relevance measured against their own data, not a bigger number of total titles. Teams with existing data science capability want the underlying, structured data layer they can combine with what they already have, not a walled interface sitting on top of it.

That's the same conclusion Gameopedia reached building the Unified Game Catalog & Discovery layer: the fix isn't more data, it's four specific capabilities sitting underneath whatever game catalog a platform already has.

• Canonical identity resolution: one trusted ID per game, so a title isn't five different catalog records because it shipped on five different storefronts under five slightly different names. In practice, that's the difference between a platform showing the same game twice under two names, or once, correctly, with every edition and release folded into one entry.

• Availability mapping: a single, current answer to where and how a game can actually be played, reconciled across cloud, subscription and storefront sources instead of contradicting itself.

• Taxonomy normalization: genres, mechanics and tags mapped to one standard, so "action" means the same thing whether the metadata came from your feed or a partner's.

• Regionalization: ratings, languages, regional titles and regional release dates handled as a first-class part of the catalog's metadata, not a patch applied after the fact, so a release date isn't fragmented into five conflicting versions depending on which region's feed you're reading.

It's the same structured data layer Discord uses today to strengthen how it connects players to the games they love. The pattern holds whether the buyer is a platform trying to stop recommending the wrong edition of a game, an ad platform trying to target by mechanic instead of genre, or a product team trying to give an AI agent enough structured metadata to stop guessing.

The takeaway from Cologne

The headline out of Gamescom 2026 was scale: more visitors, more countries, more exhibitors, a sold-out floor for the first time. The more useful story, if you build on top of game metadata for a living, was underneath it: the industry keeps growing and generating more game catalog fragmentation with each new source it adds, faster than most internal data teams can absorb.

That's not a reason to slow down the integrations, but to stop treating catalog reconciliation as a one-off cleanup project and start treating structured metadata as infrastructure. This layer has to hold regardless of how many new partners, regions or storefronts get added to the platform next year.

If your team spent time after Gamescom doing that reconciliation by hand, or explaining to leadership why the game catalog "mostly" agrees with itself, that's worth a conversation with our team. It's the problem we built Gameopedia's Metadata to solve, and after a week in Cologne, it's clearer than ever that the problem isn't going anywhere on its own.