Leading Effective Data Teams in Gaming

Interview with Mehmet Turan

From understanding player behavior to refining game mechanics, the insights derived from data analytics can make or break a game’s performance. However, harnessing the power of data requires more than just collecting numbers; it demands a skilled and cohesive team capable of navigating vast datasets and prioritizing the most impactful work among endless requests and backlog items.

We recently had a chance to sit down with Mehmet Turan, a data professional and leader in the video game industry, and asked him to share some of his insights on how to maximize the value your data team can deliver.

GGWP: Mehmet, you’ve had an extensive career at major companies like Electronic Arts and 2K Games. Could you share some of the main challenges you’ve faced in data and analytics roles?

Mehmet: Absolutely. Over the last 15 years, I’ve encountered several challenges, but three stand out: talent acquisition and resources, data infrastructure limitations, and project prioritization. These challenges recur, regardless of the company’s size or stage.

GGWP: Starting with talent acquisition, how significant has this challenge been?

Mehmet: It’s been a major hurdle. Video game companies often compete with tech giants for data talent, which has historically put us at a disadvantage. Although it’s improving as the industry recognizes the value of data, securing top-notch resources and building exceptional data teams remain tough.

GGWP: What about data infrastructure? How does that impact your work?

Mehmet: Often, the existing infrastructure isn’t optimized for data science needs. It serves multiple functions, from product development to finance, which can lead to operational challenges. This lack of focus on data science needs means our data teams are frequently hampered by inadequate support from decoupled data engineering teams.

GGWP: With such an array of tasks, how do you manage prioritization within your teams?

Mehmet: It’s daunting. At 2K, I once led a team where we had no data engineering support and everyone demanded priority for their projects. It forced us to critically assess each project’s value, measurability, and the autonomy it would grant us. These factors have to guide our prioritization to ensure we’re not only effective but also demonstrating the value of our work.

GGWP: Speaking of demonstrating value, can you elaborate on how you approach this?

Mehmet: Sure, the impact on profit and loss, alongside measurability, is critical. However, data science models don’t operate in a vacuum—it’s challenging to isolate their impact from other factors, even with A/B testing. That’s why we need a robust framework for implementation and evaluation.

GGWP: What role does autonomy play in these projects?

Mehmet: It’s crucial. High dependency on other teams or external data sources can lead to delays and unmet expectations. The autonomy over data collection, analysis, and model operationalization is essential to control outcomes and effectively implement solutions.

GGWP: What strategies have you found effective in fostering more collaborative efforts?

Mehmet: Clear, transparent and methodological communication is the key. Outlining the reasons for prioritizing a project, the benefit it will bring to the company, and specifying what exactly is needed from partner teams will open the doors for an effective collaboration among different functions. Additionally, aligning goals and objectives with your partners at the start of each year will make these partnerships much easier.

GGWP: Could you give us an insight into how different data science projects stack up against these criteria?

Mehmet: Certainly. For instance, anti-fraud and anti-toxicity models often emerge as top priorities due to their direct impact on monetization and relative ease of quantification. They also allow for considerable autonomy in model development and application, making them highly valuable.

GGWP: Could you walk us through a specific case study where your team’s data project had a significant impact? Perhaps the anti-fraud models you mentioned earlier?

Mehmet: During my tenure with a leading sports video game franchise, I encountered a challenge pertaining to the detection of virtual currency farmers and the transaction of these currencies and items within the game. Initially, the studio employed a very manual approach, acknowledging its limitations but prioritizing simplicity and control. However, it became evident that this method was not scalable.

Recognizing an opportunity for improvement, I proposed and led the development and implementation of a semi-automated solution leveraging machine learning models. We broke down the problem into distinct components: discovery, labeling, detection, and subsequent action-taking upon detection.

Collaboration was key as we shared responsibilities for executing each component effectively. Additionally, we conducted rigorous testing to demonstrate the efficacy of our models. The studio was receptive to the solution and its implementation proved to be a significant success, enhancing the overall outcome of the operation.

GGWP: What were the outcomes of this project?

Mehmet: The adoption of our solution resulted in multifaceted benefits. Not only did it enable more precise, fast, timely and extensive detection of players engaging in toxic and fraudulent behaviors, but it also freed up a substantial amount of time for team members previously dedicated to manual processes.

Furthermore, the success of our implementation fostered trust with the studio, opening the door to more future projects. This enhanced collaboration paved the way for more innovative solutions moving forward.

GGWP: Lastly, how do you communicate the importance of these projects to executive leadership?

Mehmet: Clear, methodical communication about project priorities and their potential benefits is key. For example, when we highlighted the losses from illegal activities in one of the major sports franchises and presented our solution, I was able to significantly expand my team. It’s about showing not just the immediate benefits but also how these projects position us for future success.

GGWP: Mehmet, thank you for sharing these insights.

About Mehmet Turan

Mehmet Turan –

Mehmet is a data professional and leader in the video game industry with more than 13 years of experience. With a PhD in Computer Information Systems, he has held diverse roles in leading gaming companies such as EA and 2K Games, where he established and managed data functions from inception. As the VP of Data and Analytics, he led large teams in data science, engineering and analytics. Currently, he serves as the Head of Data and Analytics at Horizon Blockchain Games, driving insights in this innovative space.