Jon Palmer
In the world(s) of gaming, analytics isn’t just about tracking KPIs. It’s about understanding entire economies, player behaviors, and strategic decisions. Whether you're running a single title or managing a portfolio across multiple studios, the stakes are high and the challenges are nuanced.
This is why it’s crucial for gaming data teams to have the right analytics platform: you need to be fast and flexible enough to monitor real-time operations, but you also need governed, consistent metrics to understand performance across games. Personally, I felt this struggle while running data teams at King and Scopely – I always felt we had to compromise something. That is, until I saw how Omni's just-in-time data modelling approach fundamentally addresses this challenge and I realised this would be a complete game changer.
In this post, I’ll talk through the unique challenges of gaming analytics, some lessons I learned along the way, and how to leverage modern data modeling practices to blend governance with speed.
As an overview of these concepts, here's a demo from my teammate Conner. He walks through defining a central definition for "Installation Date," then uses this field for a quick, ad-hoc data exploration.
The hidden challenges of gaming analytics #
First, let’s address what makes gaming analytics so challenging; it sits at the intersection of two often conflicting needs:
Centralized reporting requires strong governance: In multi-studio publishers, each title is operated like a mini-company. But executives need to roll performance up into portfolio-level dashboards to compare ROI, engagement, or monetization across titles. Metrics like retention or LTV must be defined consistently so roadmap reviews don't devolve into debates over which numbers are accurate.
Individual studios require flexibility & speed: Meanwhile, teams managing an individual game need the freedom to run fast. They’re analyzing how promotion for a new skin performed in a live event, or modeling how the “sink and source” mechanics of their in-game economy are impacting soft currency (e.g., booster) inflation. Having to wait weeks for centralized teams to add new metrics or rebuild models simply doesn’t work for the pace they need to operate.
This is why gaming teams have historically chosen between two types of tools. Some choose tools focused on strong governance and standardized metrics across games. Others choose tools optimized for ad-hoc reporting. Many end up choosing multiple and have to manage the chaos between 😅
None of these scenarios puts you on the best path to win. Gaming teams don’t need more nuances to manage; they deserve a few power-ups — not trade-offs — in their stack.
Challenge 1 #
Centralized reporting & the importance of the AARM Framework #
To get a cross-studio comparative view, most companies choose a few key metrics for measuring each studio or game. Of those, one of the most common structures for portfolio reviews is the AARM framework, which covers:
Acquisition: Top-of-funnel metrics, like installs & CPI
Activation: Players’ progression through the tutorial & early sessions, which is often a strong predictor of retention
Retention: Day 1, Day 2, Day 7, and Day 30 are the canonical milestones. Does your curve look like a healthy reverse hockey stick?
Monetization: Everything from conversions (players paying for the first time) to lifetime value
When definitions of these metrics vary from game to game, it’s nearly impossible to compare performance. Does “Day 1” refer to the day the user first downloaded the app, or is there a “Day 0”? Does opening the app count toward retention, or does the user have to take another action?
That’s where a centralized semantic model helps, so core metrics like retention and LTV can be defined once, reused everywhere, and audited for consistency. In the demo above, you can see how easy this is in Omni: Conner defines “Installation date,” promotes it to the shared model, then re-uses it for a different analysis.
Challenge 2 #
Studio flexibility to understand game economies, live ops, & more in real-time #
But, once you zoom into the studio level, the reporting needs to shift.
Game teams want to understand how players are interacting with their world:
Are players progressing too quickly or churning at specific levels?
Are boosters being hoarded or are we being overly generous in our seeding? Is our economy balanced?
How did a three-day event impact engagement or spending — did it result in a hangover?
To answer these questions, studios need the freedom to explore and build new metrics on the fly without waiting for the central BI team to update a dbt model.
Most data modeling tools are too rigid, and teams end up exporting or turning to something else. In Omni, this is where our just-in-time data modeling philosophy – the ability to model as you go – really shines. In the demo above, you can see this when Conner measures the impact of a new update: he models a few new metrics to quickly compare a benchmark cohort and a test cohort.
Game teams can explore the data deeply, create new metrics or dimensions on the fly, and if something proves valuable, promote it to the central model – making it instantly available across the org.
New level unlocked: Governed and fast #
In any game, you need the right tools and strategy to win. The data teams behind the games also need the right stack to set them up for success.
With Omni, you don’t have to choose between centralized governance for business leaders and agile, ad-hoc reporting for studio analysts. A semantic layer makes definitions explicit, transparent, and scalable, while workbooks enable fast, real-time insights.
If you’re looking for a new BI solution, I’d love to chat. Please feel free to reach out at jonathan.palmer at omni.co.