To truly understand how data is reshaping the operational landscape, we sat down with Daniel Tidström, Senior Partner at Codento and Advisor in Data & AI. Recognized as one of the top analytics professionals in the Nordics, Daniel brings extensive experience leading data and insights strategy within the iGaming space—having worked closely with heavyweights like Kambi and Svenska Spel—as well as driving product insights for global tech brands like Spotify and Epidemic Sound.
In our conversation, Tidström pulled back the curtain on the real value of data control, the mechanics of successful personalization, and how organizational frameworks are evolving to handle the next era of AI innovation.
1. Buy vs. Build: Tighter Control Over the Data Foundation
As the global market matures, many operators are re-evaluating their dependency on outsourced platforms. The consensus is clear: data has become far too critical to be left in the hands of third parties.
- Owning the Foundation: Across all platforms, data is now the primary fuel for daily operations. To maintain strict control over this asset, a growing number of companies are choosing to build their own internal data foundation layer rather than buying generic solutions. We see this shift even among Tier-1 operators migrating sportsbooks to gain better alignment.
- The Regulatory Edge: Building internal data systems isn't just about analytics; it’s a massive compliance advantage. For instance, a major success factor for some B2B giants has been its powerful, in-house regulatory engine, which handles shifting regional regulations systematically.
- The Compliance Hurdle: As companies scale out their data collection to fuel AI, GDPR and compliance quickly become major bottlenecks for the unprepared. True speed of innovation in an AI-driven world requires a seamless marriage of technical architecture and compliance knowledge—a rare hybrid skillset that very few legacy companies have managed to successfully bridge.
2. The Personalization Reality: Metrics, Retention, and Sustainable Play
While marketing teams love to pitch the idea of a "fully personalized lobby," the data science reality requires a much more calculated approach.
Composable CDPs & Real-Time Action
To achieve true personalization without data fragmentation, the trend is moving toward Composable Customer Data Platforms (CDPs). By collecting every touchpoint into a single hub—spanning marketing, risk, compliance, and product usage—operators can orchestrate a true 360-degree view of the player. This enables features like real-time push notifications that make sense, ensuring a player doesn't get a prompt to buy a lottery ticket two hours after they have already bought one.
Frameworks & testing
How do you prove personalization actually works? Tidström emphasizes that the best way to measure its effect is through strict A/B testing or multivariate testing tethered to clear success metrics like conversions or engagement metrics.
One of the historic debates in sportsbook design is whether to give users a highly personalized view or keep it consistent. The answer lies in data segmentation: using robust metric frameworks to dynamically offer a simplified experience for an average player, while serving a complex, in-depth betting view to an advanced user. Once this is done one needs to apply strict A/B testing to really see if the personalized view does have the intended impact.
Shifting the Focus to Sustainable LTV
When money was cheap, the market was flooded with new operators utilizing third-party technology, resulting in massive spending on customer acquisition. But you cannot infinitely buy players from your competitors. The strategy must shift to long-term retention and Player Lifetime Value (LTV).
"Using data to build a product that fosters sustainable, long-term play—even if it means foregoing short-term profits—is the ultimate retention tool. It is incredibly complex to compute over time, but vital for responsible gaming."
3. The future of the data organization
When an organization scales, the traditional "monolith" model—a centralized data warehouse team responsible for ingest, modeling, and output— tends to fail. It is difficult to become truly data-driven if every department is waiting in the same queue.
- Going Horizontal: Forward-thinking organizations are actively shifting toward distributed responsibility through initiatives like Data Mesh. Instead of a vertical pipeline, data architecture is being distributed horizontally across the organization, directly aligning the data pipeline with specific business outcomes and localized decision-making.
- Full-Stack Teams, Specialized Roles: Mirroring frameworks popularized by Spotify and Epidemic Sound, team compositions are becoming highly multidisciplinary. We are seeing AI engineers, platform engineers, and analytics engineers embedded into the same autonomous team. The team is full-stack, but individual roles remain deeply specialized.
- Context is King for AI: As the industry moves toward automated AI agents, old data platform truths must be challenged. AI cannot contribute to a decision-making process if it is only fed structured, tabular data. Future systems must focus heavily on contextual data—giving AI agents nuanced behavioral background they need to act directly and intelligently on the platform.
Closing Thoughts
The main takeaway from our discussion with Daniel Tidström is that data is no longer a post-back review tool—it is the operational infrastructure itself. Whether you are aiming to deploy autonomous AI agents, optimize player retention, or navigate tightening regulations, success hinges on a decentralized data architecture that prioritizes business context and tight in-house control.
Is your data trapped in a centralized monolith, or are you building the horizontal foundations required to fuel true AI innovation? Let’s start a conversation!
