Avoiding Noise in Gaming DataCollection and Analytics
Like many industries, the gaming industry relies heavily on data-driven metrics to identify potential anomalies and patterns in player behaviors and game economy. These metrics are crucial as they affect performance and revenue. Although data collection is a critical component in gaming analysis, it must be strategic and streamlined. Following the “less is more” way of thinking, analytics solutions will generate more accurate reports and allow gaming developers to pinpoint mechanics that could be affecting gamer satisfaction and revenue.
Why Collect Gaming Data?
There is more to game development than code, strategies, algorithms, and database
optimization. A big part of successful gaming development involves data collection from users themselves. Most gaming developers know that they need data collection, but don’t know where to start. Several data analytics and visualization tools exist, but starting with these tools often leaves the developer confused and unable to decipher the output. Collecting the wrong data has an even worse effect. Inaccurate data is responsible for misinterpreted output and incorrect analysis.
The value of user-generate player data is undeniably one of the biggest components in
successful games. It gives developers a data-driven outlook on the way users interact with gaming elements. Data can be used to optimize performance, make gameplay changes, detect anomalies and fraud, and find areas of the game that might need refinement.
To give you a real-world example, King Digital Entertainment -- developers of the popular game Candy Crush -- supports over 1.5 billion players a day. In 2013, data collected in-game alerted developers to a concerning gameplayer drop when they reached level 65. After analysis and review, King Digital determined that level 65 no longer challenged players but frustrated them to the point of leaving. Data collection and analytics told developers to refine game mechanics to better support player satisfaction at level 65.
Rushing into Data Collection
Rushing into data collection carries with it risks and potential “paralysis of analysis.” Developers can’t just collect everything, as this is both unnecessary and resource-intensive (e.g., waste of disk storage), but it also has compliance issues. Collecting personal data should be done with care, as now the developer must ensure that they follow privacy laws and implement the right data protection on personally identifiable information (PII).
Some data can be misinterpreted if collected without any other information. For instance, you could collect data on the time spent in game, but the time spent in the game can be affected by issues outside of the developer’s control either from dropped connections, failed user devices, or bandwidth limitations. A more impactful way to collect data would be to gain insights to the user’s location, device resources, and if a bug is causing a crash.
Collecting everything without any strategies makes more work for everyone in the analytic pipeline from developers and quality assurance people (QA) to data teams responsible for turning it into actionable advice. It also leads to bugs in analytics, inaccurate interpretations, and potential revenue-impacting mistakes in gameplay and mechanics.
Streamlining Data Collection
Instead of rushing into data collection without a strategy, the better way of implementing gaming analytics is to first ask a question. Developers should ask a question, and then a strategy can be created around the data that can be used to answer it. Collection can then be strategically implemented to find an answer. These data points are called Key Performance Indicators or KPIs.
Using a simple example, suppose that data shows most players choose a specific character as their avatar. Other data shows that players rarely choose the alternative character. This information could be used to determine if gameplay is unbalanced based on character and weapon choice and the number of times most players die compared to when they use another character.
The process of data collection follows some basic steps:
● Find the right telemetry tools
● Determine objectives and gaming attributes to track
● Define gaming variables
● Begin collecting data
● Review reports and visualized analytics
Data collection is a continuous process, but it must always be validated and normalized. This process is tedious and many organizations don’t have the resources and staff to effectively do it. By streamlining the data collection process and creating a strategy around it, normalized data becomes powerful information that can give developers answers to their pointed questions.
Focusing on Key KPIs
There are dozens of KPIs to target and collect. Some of them are basic and easily interpreted, but when you work with gaming development, mechanics and the algorithms surrounding them, KPIs could encompass several data points. To determine the right ones, always remember to ask a question first. KPIs should then answer these questions. There are dozens of KPIs that can be introduced into gaming analytics, but we can focus on basics to give an example of streamlining data collection to avoid analysis paralysis.
Using a simple example, suppose that developers support both Android and iOS. To focus on upgrades and support, developers need to know which operating system is more popular with users. Device resources could be beneficial with other metrics, but a simple collection of device operating systems is all that’s needed to determine the most popular OS platforms.
Another common KPI answers the question of where game players are located.
Developers might collect additional player demographic information, but it’s generally unnecessary for analytics and does not help determine physical location of players. Language could be useful for this KPI, as it can tell developers where to target globalization and translations, which can be used to drill down into more specific analysis. However, having full player demographics and personal information would not be necessary.
KPIs fall into four main categories:
● Discovery and acquisition
● Player experience and engagement
● Retention and churn
The data points that make up these metrics should be determined using answers to key
questions that help developers navigate changes to their game in a meaningful way. As you determine your own analytics and data collection strategies, remember that less is more. The more streamlined the process, the easier it is to normalize and structure for accurate answers to critical development questions. The end result is higher revenue, better player retention, and a more popular game.
Finally, to avoid analysis paralysis, here are some tips:
● Focus your data on the question and keep it simple
● Be realistic and understand the human element
● Understand uncertainty and be open to changing data collection requirements
● Factor in context when interpreting data
● Identify objectives before defining data collection
● Eliminate unnecessary clutter from unstructured data