In the latest trailer for Wreck-it Ralph 2, Ralph plays a game on a little girl’s tablet that lets him feed pancakes to a bunny and milkshakes to a cat. When his friend Vanellope shows up with a cart full of pancakes, Ralph starts feeding the bunny pancakes like crazy, fattening him up to comically massive proportions. The little girl reacts in shock, when she sees her simple game take a horrific turn (just watch the trailer).

Transforming the Mobile Gaming Industry with AI-powered Analytics
Ralph Breaks the Internet: Wreck-It Ralph 2 (source IMBD)

This scene from the trailer is pretty hilarious. Yet, it doesn’t just depict a little girl’s online game terror, but an ongoing nightmare for any mobile gaming company – user reactions when the game goes off course.

In our previous two posts in this series, we discussed the Big Data Challenges that Gaming Businesses Need to Tackle followed by a look at the Challenges for online gaming to Stay Competitive and Profitable. Now we’ll explore how an AI-powered Analytics solution can help identify business issues (and opportunities) before complex data operations impact a user’s gaming experience.

Data is Critical for Gaming

Mobile gaming is one of the few platforms where there is constant contact with other people. It gives people a shared experience that is hard to replicate beyond a virtual space.  And games can go anywhere, with mobile gaming. That means the point of contact is greater than ever.

Debuting in 2009, Rovio’s Angry Birds quickly became a smash hit. Over the next four years, the game had 14 different versions and got 3 billion downloads worldwide. Clash of Clans, Candy Crush and Temple Run continued the agile, data-driven approach to game development, parsing user data and releasing new versions weekly, as well as new spin offs. In 2017, revenue from those titles hit or exceeded $40 billion.

Big Data and gaming pretty much goes hand in hand. Gaming has become a major contributor to big data. The fact that games happen in a virtual world, every aspect of the game can be measured. From a revenue point of view, operating in a virtual world also presents new sets of money making opportunities. Because in certain games, the player is constantly engaged with the game, offering tremendous opportunities to advertise and allow users to buy things in this virtual space.

What Big Data Can do for Gaming

For monitoring general performance in gaming, it is important to measure key KPIs like; active users, where they came from, number of new sign ups, DAU (Daily Active Users) or MAU (Monthly Active Users), and Customer Lifetime Value (CLV). Applying big data, event-based analytics, can reveal opportunities by showing more about what users are doing on an app, like why they bought what they did, or what is the optimal path for conversion (online purchase).

Analyzing millions of hours of player data can provide insight into which elements of the game are popular. Companies can increase user engagement through analytics, revealing when players abandon a game. This data can be used to find bottlenecks and issues within gameplay.

Data analytics tools can make sense of player data, session data and game data for every game session in order to keep raise engagement and ensure performance:

    • Analytics after changes: What was the effect on monetisation and retention of this change, enhancement, or new feature?
    • User acquisition: Where should advertising budget be spent considering geo/platform/ad channel/game for maximum uptake?
  • Churn: Where are users getting out of the game? What can be done?

Mobile Gaming Needs Innovative Solutions for best Handling Big Data

Big data plays a central role in online gaming operations. Generating $40.6 billion in global revenue last year on mobile devices alone, the gaming industry is set to boom in the next five years. One of the biggest reasons for the proliferation of online and social gaming is big data. The capabilities of understanding how to use big data for monetization and optimization are coming into the mainstream.

Particularly prone to network-related metrics, such as ping and lag rates — exacerbated during peak gaming times, online game need to support a complex architecture while being ready for growing demand. Analytics can enable gaming companies to use data from servers and networks to understand exactly when, and how, their infrastructure is being stretched.

A Smarter Way To Approach Big Data

Big data by itself is of little value. To be useful, it has to be operated on by various analytical methods. By combining big data with machine learning, gaming can improve their business intelligence.

The gaming world is a complex place with millions of interactions taking place at any time, many of which affect the others in subtle or not so subtle ways. Sophisticated machine learning algorithms which not only instantly and accurately identify anomalies, but are also smart enough to adjust to long term shifts and other changes in metrics. In gaming, nothing is ever static: features are constantly added and removed, ads and promotions from advertisers begin and end, competitors attempt to pull over users with new offerings and social media marketing campaigns, and so on.

Analysis of Time Series Data Powers Insights

For tracking any given value over time, time series data sets are very useful. Since this is a near-universal need in many industries, it’s no surprise that one of today’s fastest growing (and data-intensive) industries, gaming, is looking to use time series data sets. Outliers can occur in any data set, from abnormally bright pixels in an image to an isolated spike in time series data. Time-series analytics, which measure a series of player action over time, can take player behavioral analysis a step further.

Mobile gaming companies still need to constantly monitor top-level metric KPIs and metrics such as MAUs, DAUs, and LTV for different segments. Speaking volumes about a gaming business, time series data is also the basic input for automated anomaly detection. The time series data itself is a record, not a projection, containing information which can allow data analysts to make educated guesses about what can reasonably be expected in the future. It is the task of time series anomaly detection to use those expectations to find actionable signals in the data, because those signals often take the form of anomalies (i.e. unexpected deviations in the data).

Translate Gaming Data into Inishts-driven actions with AI-powered Analytics

For gaming companies, supporting a growing user base can also mean the possibility of experiencing more crashes and hiccups. Worse, collecting terabytes of data a day means that it could sometimes take days before engineers ever become aware of issues. Due to changes of Daily Active Users (DAU), like when usage spikes on weekends, detecting crashes with absolute thresholds is practically impossible, often resulting in piles of false positive notifications that leave engineering teams immune to alerts.

AI-powered gaming analytics, using anomaly detection,  can quickly discover issues that can lead to costly business incidents in real time. With AI-powered gaming analytics, teams can fine tune alerts and remove false positives. Automatically learning from the data,  AI-powered gaming analytics can provide unique insights, enabling teams to quickly review critical metrics by breaking down crashes according to specific areas, like app, platform, and country. Engineering teams can deploy the appropriate fix, long before any impact to gameplay. Without this, companies may have to rely on crash reports, customer support case details or, even worse, wait for a long term DAU comparison, just to detect problems.

Anodot’s real-time, large-scale AI-powered gaming analytics solution is fully automated and can detect any drop in app performance, seamlessly correlating this data with other relevant metrics to present the full story for any business incident in a way that can be clearly recognized and acted upon.

With a top-quality, reliable alert mechanism, Anodot’s AI-powered gaming analytics detects anomalies for key business and technical KPIs, reducing the amount of time that would be spent having to review KPIs. This makes engineering and data analytics teams more productive and their processes more efficient and effective, relying on real time and accurate alerts. 

Who knows, maybe in the next Wreck-it Ralph, the main villain putting an end to online gaming glitches will be an AI-powered Analytics character.

Topics: AnalyticsBig Data
Get the latest news and insights on the next wave of analytics
Get a Taste of Anodot
footer background