User Segmentation Systems: Why are they Important?

What you'll learn

Values of User Segmentation

Random Hash Segmentation

Targeted Segmentation

Combining Segmentation Systems

User Segmentation Systems in mobile apps are a way to separate your users into buckets and provide a unique experience to that user based upon that bucket.

Values of User Segmentation

The principle value of a user segmentation system is that you can customize and experiment with features, design and approaches.  By analyzing the user feedback and results of different presented approaches, you can better determine which version of a feature is best for your unique user base.

For example, let's assume you have a set of user progression tasks.  When it comes to designing these progression tasks, you may be asking yourself whether it is better to have more tasks that are individually smaller, or fewer tasks that are individually larger?  To test this theory, you can create two different experiences of the same feature.  Give experience A (fewer but larger tasks) to 50% of your userbase while experience B (more but smaller tasks) to the other half.  You can then analyze feature engagement results to determine which approach is better.  

Let’s spend a few ways to look at how we can segment and classify our users.

Random Hash Segmentation

The most basic way to classify users is to separate them based upon a random hash.  You can accomplish this by taking a piece of data, let’s say their user id or their Ip address and performing some sort of hash calculation against it.  The output of the hash calculation determines the bucket they are in.

Let’s assume that we have users with a numeric user ID.  I want to randomly hash them into 20 different buckets.  I will simply perform a modulo calculation on them that looks something like:

[user_id] % 20 = [bucket]

The benefits of this segmentation approach is that it offers random based assignments.  This approach can be critical when we want to do testing, experimentation or if we want to slowly ramp up a feature releasing it to one random bucket at a time.

Targeted Segmentation

A targeted segmentation approach is very useful when you want to assign the segmentation groups based upon a criteria.  Let’s assume that we want to divide our users with a slightly different background in the application based upon their IP Addresses location.  This will be an example of a targeted based segmentation system.

Targeted segmentation systems are very useful when you are wanting to alter the various features of your app to better tailor them to your users needs.  This is very common in games where game designers may actually change user flows, thresholds or even payment packages based upon the users previous game trends.

This approach is also very helpful when you need ot change gameplay based upon country or intellectual based property restrictions.  For example, if you are running a casino style game, youmay need to alter certain gameplay components for certain countries.

Combining Segmentation Systems

There is no rule that states that you have to use A or B approach.  In fact many game studios empower a combination of the two.  Perhaps they release certain game play features based upon targeted segmentation criteria.  Then within those targeted buckets they may have a couple different versions of the experience that they experiment with to see which one is the most successful.