Leveling Up Gaming Analytics: How Supercell Evolved Player Experiences With Snowplow and Databricks

Summary

• Unified player analytics: Snowplow and Databricks enable Supercell to unify player behavior data across web, mobile, and in-game platforms.
• Real-time insights: Streaming event data into Delta Lake powers dynamic game balancing, personalized recommendations, and engagement strategies.
• AI-driven engagement: Churn prediction and personalization models leverage unified data to improve player experience and lifetime value.

Leveling Up Gaming Analytics: How Supercell Evolved Player Experiences With Snowplow and Databricks

Talk by Alex Dean, CEO and Co-Founder, Snowplow
In the competitive gaming industry, understanding player behavior is key to delivering engaging experiences. Supercell, creators of Clash of Clans and Brawl Stars, faced challenges with fragmented data and limited visibility into user journeys. To address this, they partnered with Snowplow and Databricks to build a scalable, privacy-compliant data platform for real-time insights. By leveraging Snowplow’s behavioral data collection and Databricks’ Lakehouse architecture, Supercell achieved: Cross-platform data unification: A unified view of player actions across web, mobile and in-game Real-time analytics: Streaming event data into Delta Lake for dynamic game balancing and engagement Scalable infrastructure: Supporting terabytes of data during launches and live events AI & ML use cases: Churn prediction and personalized in-game recommendations this video explores Supercell’s data journey and AI-driven player engagement strategies.
Here's more to explore:
Databricks named a leader in the 2024 Gartner® Magic Quadrant™for Cloud DBMS: https://www.databricks.com/resources/analyst-paper/databricks-named-leader-by-gartner
An open, unified approach to your data, BI and AI workloads: https://www.databricks.com/product/databricks-sql
See all the product announcements from Data + AI Summit: https://www.databricks.com/events/dataaisummit-2025-announcements

FAQs

Why do gaming companies need unified player data?

Understanding player behavior across all touchpoints—web, mobile, in-game—is foundational to engagement and monetization. Fragmented data visibility limits the ability to personalize experiences or predict churn, ultimately reducing player lifetime value.

What is Snowplow customer data infrastructure?

Snowplow is a data collection and processing pipeline that validates, enriches, and streams player behavior data into Databricks. It provides governance aligned with Unity Catalog, enabling scalable, compliant player data infrastructure.

How does real-time streaming improve game balancing?

Streaming event data into Delta Lake enables real-time analysis of player behavior, allowing developers to dynamically adjust game difficulty, rewards, and mechanics. This keeps gameplay engaging and responsive to player patterns.

What ML use cases benefit from unified gaming data?

Churn prediction identifies at-risk players for retention campaigns. Personalization models recommend in-game content and offers tailored to individual preferences. Engagement scoring prioritizes players for special events and promotions.

Full transcript

Obligatory forward-looking statement. Uh, and the thing I'm going to be talking to you about uh, for the next 20 minutes is around gaming analytics uh, with a shared Snowplow and data bricks customer Supercell. So, uh, yeah, looking forward to digging in. So. Uh, I'm one of the two co-founders of Snowplow. My other co-founder is in the audience. Uh also CEO and from oh it's quite loud from the attire. You might be thinking never played a computer game in his life.
I I uh I'm a recovering gamer and uh yeah I help help my kids uh with the with their platformers when they get stuck. Um the point I really want to land with you all today is that knowing your player, understanding your player is foundational in the emerging kind of games ecosystem. It's foundational to sustainable growth. It's foundationable to monetization in the games industry. And that games industry that we're all in is changing.
And we're going to talk about that a little bit. So case study that we'll talk about um today is around Supercell and how we've been powering crossplatform analytics for Supercell. We've been working um with our friends in the Nordics since uh early 2023 and um yeah, it's been a incredibly rewarding partnership and looking forward to sharing a little bit more about that in a moment. So, quick quick show of hands. Who in the audience has heard of or interacted with Snowplow before?
Okay. Okay. Cool. Cool. Cool. Some hands going up. So, um I will for you guys you will know this. Um, for the rest of you, I want to spend a couple of moments explaining what Snowplow is. So, Snowplow is customer data infrastructure for data bricks. And when we talk about customers in this industry, we're going to be talking about players. So, think of it as player data infrastructure. And it's a pipeline and it has multiple steps.
It's a single directional pipeline and it starts with data collection. So, we have a bunch of different SDKs. We've been building SDKs um for the digital estate uh since 20 2012 and that includes all the different digital platforms that you might build a video game in. We then do data processing. So we validate the data, we enrich it to make it more interesting for analytics. Uh and we have governance as well that plays nicely with Unity catalog.
Um and then we stream the data into the lakehouse into data bricks. Um, and you know, if you if you're familiar with the medallion architecture, you can think of Snowflower as really working that data, generating the bronze data from the game, processing it into silver, and then processing it into um, gold tables that are ready for game analytics. Um, what you can do with Snowplow is a variety of different use cases.
A lot of people come to Snowplow to solve analytics, measurement, attribution use cases, but there's some really exciting new operational use cases that Snowplow's being leveraged for. And a lot of our road map at the moment is around connecting the snowplow data in for um agentic experiences. So very dynamic operational agentic experiences. The kinds of things that you guys saw at the keynote this morning around like you know building your own agents with um agent bricks.
Um, and so here's a lovely testimonial from uh the the community tech lead at Supercell. So, Snowplowout on data bricks has been crucial for Supercell. It gives us deep insights into how players engage with our games and our events. It's empowered our teams to build better experiences. Um, especially for this game, Brawl Stars, while driving more collaboration and getting us closer to the community. Uh, it's been a game changer for understanding player feedback and elevating our product decisions.
Um, so yeah, that's a it's a lovely testimonial. I wanted to take a moment and talk about the significant evolution we've seen um in the gamer experience in the video games industry um going back quite a long way and you know I won't spend too much time on this this slide happy to chat in detail about this afterwards if anyone's interested but the point I really wanted to make is that the games industry has really evolved from you know console games into more mobile into casual into much more complex dynamic like service-based enir environments.
Um, you know, a lot of a lot of um game studios in the last few years have gone really heavily into live services and crossplatform. So, kind of moving away from that kind of single purchase experience for games into live service into, you know, uh, seasonal games into crossplatform and there's been a huge amount of change in that industry. Um, in many ways games companies have had to become a lot more like um, software as a service companies in terms of this concept of longunning games.
Where's the industry going next? We don't know. None of us know. Um, but there are some really interesting trends we think around uh what we're calling ambient gaming. So, this idea of potentially much more agentic experiences in the games. So, almost moving from a world where the games are sort of, you know, preassembled uh built by coders in a studio and then shipped or you know, live updated every every month into a much more dynamic um uh mode where potentially agents are going to be starting to build those games.
So, it's going to get very interesting. Not sure what this next chapter will be called, but um but excited to see what happens. The way the industry is changing, the player journey is just getting far far more complex. Um and this is a kind of an indicative we windy journey um from, you know, modeled off some different studio games we've we've seen and we've helped do analytics for. And the key point here is really that just like the journey is much more multi-threaded.
It's much longer running. It's more crossdevice. Um you know there are there are gamers when they're still anonymous. There are gamers when they're in your store. There are gamers when like they've you know signed up on different devices. And so the point really is that this is a very complex environment to work with. Um and Snowfl really helps to manage this complexity. Um, and this is the point kind of on the on on the the earlier slide that I wanted to make is like the shift in studio strategy.
So, this shift from the one-time purchases to continuous engagement to this idea that we need to be constantly observing um the gameplay, constantly observing how our players are operating and using that to dynamically influence the game. You know, in the old world um you know, we worked with a studio a few years ago that like um put Snowflower in during the build phase. So, you know, they were building a AAA title. Uh, they put Snowplow in and they were like, "We've got Snowplow in like whilst we're building the game, like whilst we're in sort of build mode, and they were like, well, we're not going to need it in when it goes into production."
That world doesn't exist anymore. And the world now, like, you need to be constantly iterating the game um post-production, post launch. What we see when we talk to game studios um kind of pre- snowplow if you will is we see a lot of different data silos. So we see a real fragmentation um we see a lot of different package tools that they've adopted. You know different departments inside the studio have adopted different tools over the years.
Um and it leads to like the player data being very fragmented. So, you know, we see player data fragmented across digital analytics tools like Google Analytics. Uh, product analytics tools like Firebase. Um, a lot of studios will use things like CDPs like Optimove to drive engagement and run campaigns. Um, and we see a lot of homebrew. So, we see a lot of studios that, you know, studios are full of developers. They thought, well, how hard can it be to build a real-time, you know, um, event collection pipeline?
And so we see a lot of homebrew pipelines that uh people are growing growing tired of. All of this together creates a lot of issues. So it creates issues around data quality. It creates issues around latency, cost, a lot of data governance problems as well because you're you know it's very hard to tie all this up. It's very hard to like govern this data centrally. So it becomes a real sort of hairball um that we uh we come in and we often unpick and that hairball prevents games from accelerating growth.
So, you know, as a game studio in 2025, you've got to know your player. You know, you've got to be thinking about player metrics. You got to be, you know, trying to reduce churn. You've got to be looking at the endto-end journey. You've got to be resolving identity uh for when the player hasn't yet signed up or subscribed to your your title. Um, and then you got to grow your player base. So, you know, you need to be understanding what are the different behaviors you're seeing, what are the cohorts and clusters of different players that you're seeing in that game.
Um, and you got to be engaging with the community. So, you know, you got to be engaging with the people who aren't yet in your title, who, you know, maybe are playing the previous uh release of your game and and working to bring them in. Uh player monetization, it's not a topic we always want to talk about all the time, but it's important, especially in the kind of live service world. Um, and you know, we've got interesting challenges.
We've got fraud detection, we've got like promotions, we want to run, personalization. Um, we've got a purchase funnel that we need to optimize and make sure people aren't dropping off. Um, so all of these are like, you know, tier one challenges for studios today. And that kind of hairball of package tools that we saw on the previous slide, they're not helping to to solve these problems to look at the player journey in a holistic way.
So um, this is kind of a technical architecture on a on a deeper level than the earlier one. Um, I won't go through this uh in its entirety. I think we talked already a little bit about SDKs, but the point I really wanted to land here is around the snowplow architecture and philosophy. So, an awful lot of tools and kind of analytics or personalization, whatever, they're kind of a walled garden. They have their own storage.
Um, you know, they're kind of like a SAS style tool. That is not snowplow. Snowplow is a real-time streaming pipeline to put all your game data through uh all your game event data through and it runs natively inside your own AWS or is your Google cloud account. We can host it too but it runs uh for most of our customers natively. It's bring your own cloud model and it's bring your own lakehouse. So the data goes straight into your own data bricks environment.
That's where you you build, you know, we help you build the gold level tables. And that's really really important because it means you've got all the snowplow, you know, player behavior data flowing through straight into your lakehouse and there you can join it with all the other data that you're um you know, managing. So, you know, for example, if you're you've got a matchmaking server system that's throwing off its own telemetry, you're putting that in, you can build that join.
You can build a player 360 in your own lakehouse off of the snowplow data. And you know this concept we could call it uh bring your own lakehouse we could call it lakehouse native. It allows studios to take much more ownership um of their player data of their player journey. Um it's much more futurep proof. You're not locked into a specific package vendor. Um and you know I I call out a few different um points here. So you know we a lot of our um customers in in gaming and real money gaming they care a lot about identity stitch. they have their own views on how they want to, you know, identify who these players are.
So, that's much more powerful than lakehouse native. Um, real-time streaming is really important. So, you're you're going to have some really low latency use cases that you're going to want to uh run inside the lakehouse and the snow pipeline is streaming. So, we get the data into data bricks super fast. Um, and it's also like very well governed as well. So, you know, you have full transparency, full control over the data processing.
It's not a black box. You can understand how it's processing the event data. um and all of that is quite different from uh the package solutions world. So I wanted to talk just for a moment about um Supercell and uh I think we'll have time for questions after this. So very happy to make this interactive and I think I think got a mic at the back. Um but yeah, let's talk briefly about Supercell. So uh we started working with them a couple of years ago. um they had a very specific game um that they wanted us to help with.
They also wanted us to help with the online store and we helped them in a few different dimensions. So um we brought in that very rich, reliable, high quality data straight into the lakehouse. Um we did it at high volume, we did it at low latency, uh we passed all the different volume spikes that they had and that enabled them to start doing crossplatform user journey analytics. And there's a really important point here, which is um Supercell didn't just want to understand player behavior in a specific game.
They wanted to understand what was happening upstream of a a gamer discovering that title, signing up for that title. They wanted to understand how the gamer was interacting with their website, with their online store, and similar. And that's enabled um Supercell to run much more dynamic, spontaneous um in-game events and create a lot more community engagement because they fundamentally just have a much clearer view of who these gamers are and how they're behaving like prior to um uh prior to signing up to a title.
And uh lovely lovely quote from. So our previous digital analytics tool was very limited for us. We couldn't fetch the data easily. We couldn't send custom identifiers and there were GDPR concerns. So there were kind of data governance compliance concerns. Most importantly, we couldn't integrate it with our other data sources to get a full picture of player behavior for marketing websites to in-game actions. And so it's a you know lovely case study from Supercell.
We're of course working with other uh studios um in Europe and US. And you know, I think key point here that I wanted to leave you guys with is like if you're going all in on data bricks on the lakehouse as your kind of central data repository for building the games, marketing the games, promoting the games, and life serving the games. Then you've got to be bringing the player behavior, the in-game behavior into that environment.
You've got to be bringing the marketing touches as well when the gamers are interacting with your websites, your micro sites, etc. you got to be brewing all that into the lakehouse and um yeah, snowplow is a is a great way of getting that done. Okay. Yeah. So, um we've got time for questions and before we get to questions, just want to say um do come and see us at our booth. So, our booth is um by the uh helicopter, the helicopter downstairs.
So, we'd love to see you there. And um I love talking about this topic. Um that's my my LinkedIn and yeah very happy to talk about that further. Do do connect to me and uh we can chat further. Um but yeah I would love to answer any questions that you guys have.

Learn more about the Databricks Data and AI platform.