Scaling in-game surveys to millions with taxonomy

What we implemented for variability and visibility is a game changer. It eliminates the logic we used to hack together in the frontend.

That was the immediate reaction from one of our enterprise customers after rolling out their new survey system with Hygraph.

This customer manages surveys across games and platforms that reach a 100-million user base. At any given time, around two million players are being surveyed, and the ambition is to scale that number even higher.

Until recently, they relied on a homebrew CMS that was being phased out. Adding new question types was slow, scaling to more users was impossible, and too much logic had to be manually coded in the frontend.

With Hygraph, and especially with the use of taxonomies and variants, the team was able to completely rethink how surveys are delivered. The result is a flexible, scalable, and efficient system that can handle millions of players seamlessly.

#Why taxonomy was a game-changer

The project scope was daunting: thousands of survey questions, multiple game modes, and an audience of millions across platforms. Under the old CMS, every variation had to be hard-wired, duplicated, and maintained. It simply didn’t scale.

Taxonomy changed that. By restructuring data models, we gave the team a clean way to organize modes, attributes, and platforms. Instead of repeating endless configurations, they could now define relationships once and apply them everywhere. What previously required heavy manual work became as simple as selecting an option from a taxonomy tree.

The customer described it as a “game changer,” not just because it simplified their setup, but because it opened the door to entirely new insights.

#Structuring game modes and attributes

One of the first challenges was managing the complexity of game modes and attributes. Surveys often needed to target players based on whether they were playing solo or in squads, in standard or alternate modes. Previously, each of these variations had to be recreated repeatedly.

With taxonomies, we started from a root content group and broke it into manageable pieces — Solo, Squad, Mode Alpha, and more. This structure lets the team trigger surveys for specific modes without duplication. It also gave them the flexibility to add new variations at any time without touching the frontend logic.

#Targeting by platform

Platforms were another major obstacle. The old system couldn’t group or compare platforms efficiently, which made it nearly impossible to see how Android compared to iOS, or how mobile stacked up against console.

Now, taxonomies group platforms hierarchically. Android and iOS sit neatly under “Mobile,” while PlayStation and Switch are tracked alongside. This makes it effortless to break down audiences by device, understand how many players are coming from each, and target surveys accordingly. The customer was impressed by how intuitive it felt: no more repetitive setup, no more guesswork.

#Handling features with availability dates

Features like weapons or accessories often rotate in and out of availability. The old CMS had no practical way to manage this without constant manual intervention.

By combining taxonomies with availability dates, the customer can now schedule when a feature is “rotated out” and when it’s reactivated. Surveys automatically update based on these parameters, ensuring players only see relevant questions. This approach doesn’t just reduce manual effort; it also adds a level of precision that wasn’t possible before.

#Building smarter surveys with conditions and variants

Once taxonomies were in place, the team layered in conditions and variants to control survey delivery. A survey could be shown to mobile players but hidden from web users, or targeted only at players in a specific mode. Conditions like these can be combined in endless ways, giving the team full flexibility.

On top of that, surveys can take different forms — multiple choice, free text, ratings, rankings, even QR code surveys. Variants let the same survey adapt to different contexts: one version for mobile, another for console, and another for web. This not only saves time, but it also ensures the player experience feels native across devices.

#The feedback we have received

The response has been overwhelmingly positive. In their words:

• “This is a game changer.”

• “Taxonomy made it simple and easy to target exactly what we need.”

• “It’s very intuitive — creating new survey types is straightforward.”

Instead of being stuck with hard-coded surveys, the team can now generate insights faster, reach more players, and continue to expand their survey program without being limited by outdated infrastructure.

#What’s next?

When running projects at this scale: millions of users, thousands of survey variations, and the need for constant iteration, taxonomy becomes essential. It provides the visibility and flexibility needed to manage complex operations, while eliminating the manual overhead that holds teams back.

As one of our colleagues put it: “I’m not sure how we would have done it before — having taxonomies made everything so much simpler. The timing was perfect for this use case.”

This demo shows that with the right structure, surveys don’t just collect feedback, but they become a scalable, dynamic part of content operations. And with taxonomies and variants at the core, organizations can move faster, operate smarter, and deliver at a scale that once felt out of reach.