Optimove performs predicted future value calculations based on a combination of (1) the likelihood that customers from one micro-segment will move to a different micro-segment, and (2) the value represented by each migration. Let’s look at each of these items separately.

Once you’ve got the model’s results, here’s how you can use them to build your target groups right in the UI.

Customer Migrations from One Micro-segment to Another

As you probably already know, much of Optimove’s customer modeling is based on the use of customer micro-segments. Micro-segments are small groups of customers sharing very similar behavioral attributes. Optimove is constantly tracking which customers belong to which micro-segments and how they migrate among micro-segments over time. Based on the history of customer migrations within the database, the software can predict the likelihood of customers in one micro-segment moving to a different micro-segment in the next period.

|| Note that the typical customer database contains hundreds of different micro-segments. Customers frequently migrate among these micro-segments based on their recent activities and behaviors.

The Historical Value of Micro-segment Migrations

Optimove also tracks the historical value derived from an average customer migrating from every micro-segment to every other.

|| Note that “value” usually reflects revenues, although other metrics are sometimes preferred (e.g., net gaming, profit, number of transactions).

Calculating Future Value

Now that we understand the two components of the calculation, we can describe how Optimove predicts the future value of customers in each micro-segment. Let’s illustrate this with an example.

The following is a graphical representation of eight micro-segments that Optimove is tracking for a particular customer database. The arrows represent ways that customers can move from one micro-segment to another. For this example, let’s take a look at how customers migrate from micro-segment 1 to three others: 2, 3, and 6. The heavy arrows show these migrations, along with the percentage likelihood of a customer making each migration and the value of each migration.

In this example, 40% of the customers in micro-segment #1 are expected to migrate to micro-segment #2 in the current period. Based on prior customers who made this particular migration (#1→#2), we expect each customer to generate $320 of value while making this migration during the current period.

Likewise, we expect 50% of the customers to migrate #1→#6 while generating an average of only $80 (because #6 is a lower-value micro-segment).

Finally, we expect the remaining 10% of customers to migrate #1→#3, while generating no value during the current period (#3 is a micro-segment from which customers frequently move to a “churn” micro-segment) and generate no further value at all.

The software performs these calculations for every micro-segment in every period.

Into the Future

Future Value.png

In the example above, we only talked about one period into the future. To calculate the total future value for a particular micro-segment, Optimove repeats these calculations into the future for all the possible micro-segment-to-micro-segment migration routes that all customers who are currently in that micro-segment could take in the future. Optimove’s algorithms use these results to arrive at a predicted future value for customers in each micro-segment.

It is important to notice that:

Customer future value and customer lifetime value are two very different metrics. Customer future value is a predicted figure representing the future revenues (or other metric) that the customer is expected to generate from today until the future value calculation horizon. Customer lifetime value is the sum of all revenues actually generated by a customer until today, plus the predicted future value.
The distance into the future that Optimove calculates future value depends on a number of factors, such as industry, typical customer lifetime, and the Optimove user’s needs. You can see the period in use on the Dashboard page, at the bottom of the Predicted Future Value Trend chart.