Identifying High Value Customers with Kumo

One of the most important questions growth and GTM teams ask their data teams is – who are my most profitable customers and how can I get them to increase their spend?

Identifying users who have a high potential to spend more in the future allows you to better segment your customer base and prioritize your limited resources (whether it’s ads, outbound outreach, marketing campaigns, or direct sales) to capture the best ROI. You can read more about the benefits to using predictive ML to identify future spend and lifetime value in our LTV blog.

In our last walk-through, we showed how you can identify users likely to churn or decrease significant activity on your platform. In this walk-through, we’ll show you how you can use Kumo to predict future purchases of specific users and ultimately identify their lifetime value over a particular time period.

Building the Graph

Your enterprise data naturally is represented by a graph – the entities include users, transactions, clicks, views, and attribute data. This graph includes the patterns and interactions between the entities that represent users, purchases, actions, and more. 

Kumo lets you quickly connect your data sources to build the graph in a matter of seconds. This graph is built in a way that’s optimized for GNN learning.

To analyze future spend, you can include tables that are related to purchases/transactions, user activity (views, clicks, ratings), as well as any user profile information. Online marketplaces and ecommerce platforms can include data on the products or items sold. The more data the better!

You only need to build your graph once – once you have it, there are a near infinite number of use cases you can tackle, and you can start asking many different questions about your data. 

In this example, we’ve added tables to analyze the ecommerce platform’s customers and their transactions. Link your tables based on Primary / Foreign Key relationships and you’re done!

You can scale to dozens of tables, terabytes in size, with tens of billions of rows collectively. With a large graph created, Kumo will enable you to learn complex patterns and interactions from the different schemas – the more tables you connect, the greater the depth of the learning.

Now, you can reuse this same graph to build any number of predictions across a broad range of use cases. Kumo.ai will maintain and update the graph as new data comes in or as tables are updated.

From Business Problem to ML Pipeline in A Few Lines of Simple Code

With your graph built, you can now start generating predictions! Kumo offers a flexible and customizable language called the Predictive Query, which specifies the machine learning problem you’re trying to solve. 

Kumo uses the query kick off an end-to-end ML pipeline (no feature engineering or pipelines needed!) where the Kumo AutoML process runs under the hood, identifying the best architecture, model, and parameters for your task and graph. You can run a near-infinite number of predictions for different types of problems in a matter of hours, each one deploying a new, unique GNN built for the specific task.

Below is an example of a query that predicts the 30-day lifetime value for every user by predicting the total revenue they’ll generate:

PREDICT SUM(Sales.Price, 0, 30)
FOR EACH Users.ID

Now, if you want to focus just on the 3-month active users, you can modify the query based on users who have opened a session in the last 90 days:

PREDICT SUM(Sales.Price, 0, 30)
FOR EACH Users.ID WHERE
        COUNT(Sessions, -90, 0) > 0

Another benefit is the ability to perform what-if analysis or scenario planning to see how specific actions can change the outcome. Below, we make the same predictions but assume that we send these users a coupon over the next week:

PREDICT SUM(Sales.Price, 0, 30)
FOR EACH Users.ID WHERE
        COUNT(Sessions, -90, 0) > 0
ASSUMING COUNT(Coupons, 0, 7) > 0

Once the query is trained, you can review the evaluation metrics to validate model performance across relevant metrics.Assuming the results meet the bar, it’s time to scale out predictions across your user base to understand potential sales across each segment.

Using these predictions directly in your decision-making process, you can prioritize your high value users in retention, marketing, or specific outreach campaigns based on their projected purchases for any given action, or you can make long-term revenue projections to forecast your overall sales for any given segment.

Stay tuned for more walkthroughs on making the most of your enterprise data!

Request a demo today!