Maximizing Customer Retention: How Segment CDP Can Help

On this page

Getting Started

To complete this recipe in its entirety, you will need Segment with a license for Engage Premier. Your instance of Engage Premier should be configured and working. As a note, if you leverage a different platform for activation, you could simply create the audiences through the steps below and send that data to the activation platform of your choice.

As another prerequisite, you must have already configured your event tracking. You will need to complete this step to fully implement the platform. For this recipe we have assumed that you've instrumented your site or app already and that data is flowing into Segment. Order Completed is the main event used in this recipe but any event that indicates a customer’s level of engagement may be useful when building out comprehensive retention journeys. For retention campaigns it is helpful to have data from transactional systems or products in order to determine how engaged a customer is with your brand. Other sources like email or CRM can also help understand a customer's level of engagement. For this example, we will focus on transactional data. With a Source created, it’s time to start putting together events that allow Segment to understand all of the engagement with the site. The events created for this recipe should be based on the V2 Ecommerce spec.  

Note: In building this for a live production instance, the spec should be followed as closely as possible and should include all of the events available.

We will also assume that we have the appropriate tracking calls (page, identify, track, screen) configured for your website.

In summary:

• Licenses to Segment and Engage Premier

• Tracking required includes:

  • Identify calls 

  • Tracking calls for: common Ecommerce events and Order Completed in particular

• Website (including shopping cart) is required for tracking

Step 1: Configure a new Space for your Customer Retention campaign

First, if you don’t have a space available for this, you will want to configure a new space that we’ll call “Retain Slipping Users” or “Customer Retention,’ which uses the website and/or app source feed we set up prior to getting started.  If you have a space already in place that you want to leverage, please skip to Step 2 to ensure it is configured correctly.

This can be accomplished by navigating to the left side of the tool, and selecting “Profiles.”

From there you can see “Spaces” and any existing spaces like you see below:

When selecting “Create space,” the user is prompted with the following screen where you can provide a name for your space.

Note: the first space you create will default to a “Dev Space.” 

Step 2: Set your Identity Rules.

Within our new Retain Slipping Users Space, navigate to Profiles and then Profile Explorer. Here we need to select “Set Identity Rules” and go through this configuration selecting the appropriate options based on your data source and desired identity management. Some of the common “Default Identifiers” include user_id, email, and anonymous_id. There are several others available and custom ones can be added (including phone), so this should be well planned to ensure your identity resolution matches business needs. 

Step 3: Connect Sources to your Space.

Next, click “Connect Sources.” Select the website source you created for the prerequisite section or whatever source you plan to implement with a Retaining Slipping Users campaign.  

Note that when you add a source, you’ll be given the option to replay data. Replaying data allows you to populate profiles with the last month of data automatically. In our instance, we will be selecting this to populate historical data about our customer base, but optionally you could choose not to leverage the historical data and only have data for the campaign from the point of activation.

Step 4: Create Computed Traits.

RFM (Recency, Frequency, Monetary Value) analysis is a commonly used customer segmentation technique that can be applied to identify customers who are at risk of churning. By analyzing a customer's recency, or the last time they interacted with your business, their frequency of interactions, and the monetary value of those interactions, businesses can gain insights into the customer's level of engagement and loyalty. Customers who have not interacted with your business in a while or who have lower frequency and monetary value of interactions are more likely to churn. The data required for RFM analysis typically comes from transactional data like purchase history from an eCommerce site or a retailers POS system but can also include data regarding a customer product usage, like the last time a customer used your app, or the number of articles a customer has read. However, it's important to note that other methods, such as app usage or social media engagement, may be more relevant depending on the industry your company operates in. By using tools like Segment and Twilio Engage, businesses can collect and analyze customer data to determine which method is best suited for identifying churn risk and take proactive steps to retain slipping users.

For this example, we will use Segment’s Computed Trait feature to build the necessary user traits to determine a customer’s Last Order Date (Recency), Total Number of Orders (Frequency), and Lifetime Value (Monetary Value). By using these customer traits, we can then build audiences of customers that are likely to churn and journeys to encourage or incentivize  them to re-engage with the brand.  

We will do this by navigating to the Engage section on the left side of the application, and then select “Audiences” followed by selecting the “Computed Traits” tab. Once there, you should see you do not yet have any Computed Traits,and we’ll get one started by selecting “Create computed trait.”

To build our Last Order Date trait, we will select the Last option and build a trait that equals the timestamp of the last time an Order Completed event was tracked for a given customer.

To build our Total Number of Orders trait, we will select the Event Counter option and build a trait that counts the number of times an Order Completed event was tracked for a given customer.

To build our Lifetime Value trait, we will select the Aggregation option and build a trait that sums the revenue of each Order Completed event that was tracked for a given customer.

Step 5: Create your first Audience.

Now that we have data flowing into your Retain Slipping Users space, the next step is to create an audience within Engage Premier that allows us to determine when someone creates a cart or places something in their cart, but does not complete a purchase within a set timeframe. We’ll use 4 hours for this example, which would allow an individual enough time to come back to their cart that day if they were simply interrupted or switched devices during their purchase.

We will do this by navigating to the Engage section on the left side of the application, and then select “Audiences.” Once there, you should see you do not yet have any audiences, and we’ll get one started by selecting “Create.” 

Here you can add the various conditions you expect to use to build the audience for your campaign.The conditions may vary based on your website, event tracking, and business objectives. Our audience will include any users that 1) have not made a purchase in the last month, 2) have ordered at least 10 times, and 3) have spent at least $1,000 dollars with our company total in their lifetime as a customer.

Step 6: Preview and save your Audience.

Once you have the desired logic in place, you can then select the “Preview” button on the right side of the window to preview the audience that this would include. If you don’t have any individuals that qualify for the audience yet, it may come back with no results. If this occurs, you may need to rethink your inclusionary and exclusionary criteria in order to increase the likelihood that customers will qualify or QA your source data to ensure the events you are using are actively collecting data. After previewing the audience, if it matches your expected results, select “Save audience.” The next screen will allow you to send the audience to a destination (if you are not leveraging Engage Premier, you likely want to send your target audience to the desired activation platform), but we are going to skip this step and simply select ‘Next.’

The following screen allows you to set the name of your audience and determine if you want to use historical data, or only new data going forward. In our case, we will leave the default state of gathering historical data and click “Done.” This will automatically trigger the building of the audience.

Step 7: Create your content.

As the audience goes through the build process, navigate into Engage on the left side of the tool again, and select “Content.” From here you should be able to select either the “Email” or “SMS” tab, and create a template message for use in a campaign.

These templates can include incentives to come back and re-engage with the brand (free shipping, coupons, etc.), and product recommendations for similar items to those they previously purchased. Your content will vary greatly depending on the industry you operate in. For example, a service company may include content regarding new offerings whereas a SaaS company may offer opportunities to participate in beta testing a new feature. 

Further information on creating templates can be found here:

Email: https://segment.com/docs/engage/content/email/template/

SMS: https://segment.com/docs/engage/content/sms/template/

Step 8: Create your Journey.

Once your templates are ready, navigate to the “Journey” menu item on the left. Here, you can either explore existing journeys, or create a new journey depending on whether you have been through this before. In creating a journey, you will need to give it a name, and set some basic configuration, including if users can re-enter, and when they should be ejected from the journey.

For this example, we will eject users 4 days after entering when all messaging is completed, and they can re enter up to every 2 weeks.

It's time to build out your journey. You will need to have an entry step that retrieves everyone from the audience previously built, followed by an email activity for that initial message to your previously engaged customers.

After the initial message, you can wait a couple of days and then confirm they haven’t purchased yet, and if appropriate, send a second message.

When completed, your journey flow may look something like this. You can change it as you do more testing. For example, based on the test results, you may add a 3rd email message, adjust messaging, include text messages or push notifications, or potentially send audiences to ad platforms when appropriate.

When you’re happy with your journey configuration, select “Save as draft,” and then you’ll be able to publish your journey, turning it into a live campaign. 

Wrapping up

Here’s what we’ve done in this growth recipe:

• Created a Space within Engage

• Created computed traits and an audience of users who are a churn risk

• Built a journey that sends users in the created audience emails encouraging them to re-engage with the brand