The global games market is projected to reach $90.7 billion by 2020. But gaming companies have a huge problem to overcome: their metrics. They often struggle to collect, interpret and visualize user interaction data, which is imperative for understanding, learning from and predicting behavioral trends.

However, it’s not for lack of trying. A majority of online game companies are cobbling data together that is stored in many different places trying to achieve a unified behavioral view. Unfortunately, this approach only shows bits and pieces of the customer journey — very few companies actually understand the entire customer journey.

Game companies know that behavioral insights are key to their competitive advantage. If they can make their game more engaging, entertaining, or more addictive than the next guy’s, that’s their edge. That’s why behavioral analytics is gaining traction in the gaming space. And they’re changing the gaming industry in a number of ways:

1) Assisting small game developer teams to do more, with less

Game development environments need to be flexible to adapt to changes in user behavior. Most teams don’t have the luxury of having an entire business intelligence and data warehousing team on staff to reconfigure analytics solutions for each release.

A relevant example which comes to mind is social-gaming company LuckyFish. Their entire company, from the c-suite to finance, marketing, and the development team, have turned to analytics, and use it as a daily barometer to help them understand the success of their games, make decisions and better target and retain customers. The retention team analyzes both backwards-looking and forward-looking behavioral data, which indicates user segments at risk of churn. The data helps shed light on what problems are causing users to churn helping the retention team respond with targeted campaigns. Then, they use the same dashboards to understand if retention activity is effective or not.

2) Leveraging AI to help understand player journey from conversion to churn

When your data is fragmented across multiple platforms, getting the answers you need requires wrangling data manually. Many teams lack the expertise to dig deep into the data and gain meaningful insights. Often, this requires the skills of a university-trained data scientist. To overcome this limitation, companies are now beginning to turn to artificial intelligence and machine learning.

In order to prevent user churn, one must understand the actual cause. By using AI, not just click-stream data or purchase data, there’s much more depth to the data. This creates the ability to analyze churn patterns in order to better predict and forecast user turnover. With machine learning in play, product managers are better equipped to predict, diagnose, and course correct because often, slight nuances in data reveal what may be a very strategic and material trend.

Examples of this may be: grouping users by their behavior so you can take differing, and appropriate actions for each, identifying cohorts of users who are likely to convert to paying customers, or conversely, identifying users who are at risk of churning before it actually happens.

Previously, this was a manual, tedious and error-prone effort, but leveraging AI and automation is enabling companies to become a lot “smarter” and more effectively engage with their customers.

3) Revealing questions surrounding player’s behavior

The challenge with traditional analytics is that each platform gives you a different set of data. There are many point solutions to help you understand web and mobile analytics. However, there lies a vast array of what is known as “unstructured data”, which is often overlooked and under-analyzed.  This may include messages from customer chat, customer success information and much more. Natural language processing (NLP), a machine learning technique, can be used on chat data to infer the sentiment of a particular user within a conversation. Running NLP analysis on chats can generate reports that paint a much more nuanced picture of customer satisfaction.

Another way behavioral analysis can be used to drive business value is to predict and prevent churn. If you analyze a list of 100,000 users by profile attributes such as gender, age, activity level within the application, behavioral patterns and historical churn data, you can generate predictions that rank their likelihood of churning top to bottom. From there, you can cluster your users by their behavioral attributes to identify which groups of users are represented in the ‘highly likely’ to churn cohort.

The analysis of this raw event data generated provides marketers, product managers and executives with countless new insights into the behavior of their players. This, in turn, allows them to conduct far more effective marketing campaigns and dramatically reduce player churn using this combined set of previously unavailable data.

4) Improving overall path to conversion strategy

While there are some pretty intuitive ways of tracking your game’s success, it can be tough to convert those insights into growth and retention strategies.

Given all of the metrics that are touted as the next best thing at helping you improve your marketing or your monetization, it can be pretty easy to get lost in all of the noise. However, there are a few metrics that when measured and tracked allow game marketers to build a data-driven conversion strategy.

One of the best metrics to keep track of the development of your customer referral strategy is called k-factor. Your game’s k-factor is the amount of invites sent by each customer of your application multiplied by the conversion of each invite.

The k-factor can be calculated as:

This metric neatly summarizes the effectiveness of your referral growth strategy, making it a quantifiable metric that can then be compared at different times in your game’s development in order to measure the effectiveness of your approach.

As an offshoot of the k-factor metric, game marketers should measure the number of invites sent compared to daily active users (DAU). Instead of measuring your referral conversion rate, however, this metric gives a more comprehensive picture of how well your referral program is retaining the users that have downloaded and played your app.

For games with a built-in currency system, they should also consider source, sink and flow metrics. The source metric indicates the amount of currency a user has earned as they progress through the game. The sink metric indicates the stages at which a user needs to spend the currency to move forward or compete with other players. The flow metric is a measure of both sources and sinks: it is the total balance of currency that a player has earned and spent over a period of time. This metric allows you to see how you can nudge a player toward in-app conversions or purchases.

If you know how users spend their in-game currency, you will be able to richly understand user engagement and the thought process of the people playing your game through source, sink and flow data.

Looking forward, the next phase of gaming software advancements will rely heavily on the integration of behavioral analytics to understand gamers’ tendencies and preferences. Empowered with that kind of information, games can be calibrated in real time to hold user attention, increase purchases, and grow the bottom line.

Dan Schoenbaum is CEO of Cooladata, a predictive and behavioral analytics platform, is using cohort analysis to bring advanced analytics solutions to fast-growing digital companies that rely heavily on user behavior insights.

Evan Kaeding is a Data Strategist at Cooladata. Leveraging a background in finance and investments, Evan has led teams to use data to make investment decisions and improve operational efficiencies.