Maximizing The Value Of Player Data

There’s one key ingredient that’s the most valuable part of any game for developers, and it might not be the most obvious element or the first thing that you think of. It’s not the source code, database, live ops strategy, or anything other than the players themselves. Or more accurately, the data they generate.

What’s more, maximizing the value of that data might be the most valuable thing that you can do as a marketer or game developer. And there are plenty of ways to accomplish that, starting with understanding and organizing the data itself. In fact, there are five groups of activities or sets of actions that you’ll take during your data collection or journey through the realm of gaming telemetry:

1. Understanding Telemetry As It Applies To Your Game
2. Knowing Your Objects And Attributes
3. Defining And Name Your Features Or Variables
4. Collecting Your Data And Communicate Your Results
5. Gleaning Real World Insights And Taking Action

Let’s start at the beginning.

Understand That Data Collection Is Always Ongoing

Every player in your game is constantly giving up useful data every time they log on. Player data is not just demographics, money spent, and contact info, it’s a collection of user behavior. Organized and analyzed properly, this gives you an incredibly valuable insight into the minds of your users, allowing you to deliver a better gameplay experience for them and more profits for you. Everyone wins!

This is all tracked from multiple sources and cross referenced in order to create a collection (or should we say wealth) of information called telemetric data or telemetry.

Telemetry In The Context Of Gaming

But first, let’s define telemetry and what it means for gaming companies. Defined as an automated communications process of recording and receiving data, telemetry is utilized in wide range of industries. In the gaming industry, telemetry is a fundamental part of game analytics. In fact, it’s the core of the game analytics realm in a lot of ways.

What’s more, the definition of telemetry can be simplified even more as the ability to collect multiple variables and data sets over time.

Gaming telemetry or recording and receiving data on user behavior covers a lot of activities – in fact, everything from the naming conventions and structure of the data itself to the operations performed on it could be covered under that umbrella term. However, it’s often also used to describe the raw data itself as it exists in analytics software or reports. And of course, the goal of all of this is improving performance and solving problems in game development, marketing, and research.

So that’s all well and good, but what does it mean for you? Getting your game’s telemetry and related processes in order is the foundation of maximizing the value of your data.

Telemetry can be simplified even more as the ability to collect multiple variables and data sets over time.

Clarifying What Telemetry Means For Your Game
At its heart, telemetry tracks attributes and objects, or in the context of gaming, processes and the people who performed them.

In some cases, it’s fairly clear which is the object and which is the attribute, however in others it’s murkier.  For example, if a certain update to your game results in a significant amount of player abandonment at a specific point, the object is the problematic spot of the update and the player abandonment is the attribute. On the other hand, if a player spends certain amount of time completing a certain mission in a game, the player is the attribute and the mission and the time required to complete it are the objects.

Understanding the difference between objects and attributes as well as how they are connected is the foundation of your game’s telemetry – and the beginning of extracting the most insight, knowledge, and value from your player data.

Develop Naming Conventions

Defining what certain data elements and behaviors are called – creating naming conventions – is one of the first steps you should take when setting up tracking or telemetry for your game(s), This is perhaps the most overlooked part but also one of the most important parts of game analytics. After all, if you don’t know what you’re tracking and you can’t organize it in a cohesive fashion, you’re going to have a tough time figuring out what to do with all that raw data.

Granted, different fields and even different companies or analytics programs use different naming conventions; although there are some relatively standardized terms in the gaming realm and other fields. That said, the specific terms being used aren’t as important as making sure that your team and any other stakeholders have a clear understanding of any unique naming conventions.

Understand Data Operationalization (And Deploy It Wisely)

Once you have your objects versus attributes and your other naming conventions sorted, you need to operationalize them, or determine a standardized way of expressing and measuring the attributes. For instance, are you measuring lengths of time in minutes or seconds? Amount of money spent or the number of items purchased? Or both? Player duration in hours played or the amount of days they’ve had the game installed? Or consecutive days with significant playing time?

Are you measuring lengths of time in minutes or seconds?

Operationalizing your attribute data means turning it into trackable variables or features; note the terms “variable” or “feature” are used in different industries to describe very similar concepts. For instance, in the game development field the word “variable” is often used to describe an attribute, while in other data science or computer science fields, “feature” is used.

Regardless of whether you call it a variable or a feature, each data set has a domain, or a defined collection of all possible values that it can be populated with, whether those are binary 0’s or 1’s or the number of hours in a day. In fact, in many ways defining the domain is the most important aspect of operationalizing and giving naming conventions to your data.

At the end of the day, using the term “feature” versus “variable” is not particularly important, as long as you set clear naming conventions for your game’s analytics from the start. In fact, the context of game development, the definition of telemetry can be simplified even more as the ability to collect multiple variables and data sets over time.

Make Your Data Shareable

The best laid plans of mice and men are useless when the can’t come together again, at least when it comes to player data. That’s why your telemetrics (encompassing all the objects, attributes and their subsequent domains, etc.) should be stored in various database formats that are easily interpretable – and easy to report upon. This would be your collection of game metrics or your game analytics.

Making your game metrics shareable is a key part of maximizing their value; after all, what good is all this information if it can’t be easily shared and expressed with easy-to-understand reporting. And that might mean turning your telemetrics or game variables into commonly understood metrics like time metrics, involvement metrics, revenue metrics, or event metrics. Translating your telemetrics into game KPIs is generally the final step in the process, at least when it comes to reporting and sharing these reports.

Realize Your Limitations

But let’s go back to collecting the valuable data in the first place. When you think about extracting the most of value out your data, you probably think of data mining in general. While there are certainly practical issues with data mining like transparency, privacy issues, security, and data cleaning, performance and sampling might be the most important when it comes to maximizing the value of your data.

Consider the fact that most data analysis techniques were not designed to deal with massive (think terabytes upon terabytes) of data that modern games generate. Plus there is the sheer number of variables or features generated by all the data, which can be referred to as its dimensionality. And that can be a massive amount of information to deal with.

 Most data analysis techniques were not designed to deal with massive amount of data that modern games generate.

Sampling is one of the obvious solutions, of course, or at least the one that is often used in the game development field. Naturally, that also holds some challenges, namely ensuring you capture a dataset that accurately represents your population as a whole, as well as its key features and variables. This is commonly used in many industries that have to collect and perform statistical analyses of large populations or data sets (think censuses or polling), and it can work surprisingly well for the vast data sets generated by today’s complex games.

There’s also the option of parallel programming, where you divide the datasets in subsections in order to track and analyze them, and merge the results later. But that’s a subject for a future article.

Be Aware Of Unexpected Psychological Bonuses From Telemetric Data

While the ability to track, measure, and better understand your user behavior through data – thereby maximizing its value – it certainly the number one reason to develop a solid analytics program and understand telemetry, it’s not the only one. There are some other less obvious boosts or softer benefits that come from fully maximizing your data’s value. For instance…

Like Using Data Telemetry To Get Yourself Out Of The Micro Optimization Trap

Something that many game developers do is get stuck on minute details or performing dozens if not hundreds of tiny micro-optimizations. And while perfecting each micro-format isn’t necessarily the worst thing ever, it can certainly hinder progress in general.

Which is why looking at the bigger picture and seeing what portions of game play really matter – and therefore it would be the smartest to expend your perfectionist energy. And isn’t helping you direct your focus where it might be the most effective perhaps the most significant way to maximize the value of your data?

Or Using Your Learnings To Avoid A Dreaded Procrastination Loop

Procrastination is the enemy for many of us, developers most definitely included. And that’s tied into the temptation to get caught up in the details; e.g. you have to make just more one thing perfect before you launch an update or make any other significant changes.

Analytics and looking at the broader data trends, however, can force you out of the feedback loop of endless perfecting and procrastinating. Which might be one of their not-so-obvious capabilities. By looking at trends in your analytics data and seeing where you are actually losing out on performance (and therefore profit, most likely), you might be jolted into action and out of procrastination mode. There’s nothing like the cold harsh truth of numbers that don’t lie; but of your need to set up your analytics with appropriate naming conventions and operators first – so don’t procrastinate on that!

Know The Value Of Starting Early

Usually the analytics and game metrics programs are applied relatively late in the process; in fact, they are often applied to too late to be effectively “built in” to the game. But of course, in order to truly maximize the value of your data, the tools, processes, naming conventions, and operationalism that need to be planned out before you get too far along in the development – data collection or telemetry and analytics are not something to be tacked on at the end, especially if you want it to make a difference in the live ops programming for your game.

In addition, setting up proper telemetry and analytics allows to track the lifetime value of your players along with behavioral patterns over time, which can give you crucial insight into how your players’ minds work as well as giving you more direction as to which types of player personas you should concentrate developing towards.

So How Do You Make All This Applicable In The Real World?

Once you collect all the data, you need to figure out how to apply it to the real world – which is the goal of all these efforts to begin with! After all, you have to be creating value to begin with in order to maximize that value in any way, shape, or form.

The various ways that developers use their telemetric data to generate real world results tend to fall under three main umbrellas:

Data And Developing New Game Features

Gone are the days that once a game was shipped, it was done. Now there are regular updates, new patches, characters, loot, and other features that make games a consistently new experience – a successful game requires a constant stream of new features and activities in order to keep things fresh and interesting for the type of highly engaged valuable player.

Being able to properly understand the current state of a game’s players and their activity requires appropriate and accurate telemetric data; inaccurate or misrepresentative data leads to poor versions of games and weak updates, which is certainly undesirable for everyone involved!

This is perhaps the most valuable for game developers working on live ops or making updates to existing games since they can use telemetric data to make optimizations in real time, but building on knowledge from other games and associated data can be valuable when you’re starting a new game as well. That said, the live ops team should be closely connected to the games analytics when making updates and building new features; in fact, the predictive or prescient nature live ops means that maximizing the value of the game’s data is of utmost importance.

Data And Performing Platform Analyses

The data derived from games can also be used to provide key learnings about the platform (e.g. the device or console) that it’s being played on, e.g. engagement, technical and graphical bandwidth, and other user behaviors that can benefit gaming engineers and their equivalents in related fields.

Knowing how users use the actual devices and hardware to play popular games along with console software and operating systems helps all the engineers involved in the gaming world in some way, shape, or form make things better. While this is a bit less obvious in the way that it helps you maximize the value of your data, building a better gaming ecosystem is better for the industry as a whole, and that has a halo or butterfly effect on your game(s).

Data And Attributing Purchases Or Advertising Channels

Perhaps the original reason for collecting all this data is to figure out who should get paid and how much – and who should get what amount of credit for the action that led to the sale! So naturally being able to maximize attribution and track buying behaviors is a key goal of any game analytics program.

Telemetric data can not only tell when and where a player made a purchase, of course, but also how they got to that point and their behavior prior to that purchase or ad click; accordingly, analyzing that behavior can help you create more of it. This is probably the most explicit use of telemetric data in regard to maximizing its value, since you are literally using the data to track and attribute earnings and learn how to create more of them.

Here Are Five Steps To Take To Maximize The Value Of Your Data

At the end of the end the day, all of this data collection and study can be grouped into five sets of behaviors or activities:

1. Understanding Telemetry And Applying It To Your Game
2. Defining Your Objects And Attributes
3. Understanding And Naming Your Variables
4. Collecting Your Data And Communicate Your Results
5. Gleaning Real World Insights And Taking Action

While each of one of these sets of behaviors contains numerous potential activities or tasks, together they make up a combination of steps that will help you make the very most of the data you’ve collected from your game’s players – and help you build a game that’s more fun for your players, more profitable for you, and more effective in general.