Splits are a powerful way to add granularity to your Streams, enabling more precise customer segmentation. By incorporating decision points such as an "Attribute Split" or "Response Split," you can segment customers within a treatment based on their behavior, creating a more tailored experience.

There are five types of splits you can create:
 

Response Splits

Use the Response Split to segment treatments based on customer activity. This allows you to adapt your approach depending on how customers interact with your campaigns or key metrics.

A response can be based on:

• KPI: Whether the customer met the Stream's Key Performance Indicator (KPI). For example, if the customer generated an increase in the Total Net Sales Amount.
• Engagement Activity: Whether the customer engaged with a campaign, such as clicking on a link or interacting with promotional content.

Please Note: If you want to use Response Splits for third-party channels, you must report the campaign metrics to Optimove via API. Read here to learn how. 

Attribute Splits

Personalize each treatment based on a customer’s activity or behavior. Select any attribute from the dropdown, and tailor your approach by setting multiple values for a single attribute.

You can also use multiple attributes in a single split. For example, you can split a treatment between high-value VIP customers and those at risk of churn to personalize your messaging. 

Please Note: The Attribute Split is prioritized by order. If customers qualify for multiple values, they will only receive the treatment associated with the first value. Customers who do not qualify for any value will be categorized under 'else'. 

Trigger Split

Segment treatments in real-time using the Trigger Split. This allows you to respond instantly to customer activity. For example, you can send tailored campaigns based on whether a customer added an item to their cart or performed a similar action.

Web & App Activity Split

The Web & App Activity Split is used to segment customers based on their digital engagement within your website or mobile app. It helps you tailor customer journeys by checking if a user has performed a specific action within a defined lookback period. Customers are then routed down different paths based on whether they met the activity criteria.

You can configure the split based on the following:

• Type of history: Decide whether to segment customers with or without the selected activity.
• Activity Type: Select from tracked activities such as Page Visits, App Opens, or custom events like 'Bonus Claimed'.
• Condition: Use a filter to narrow the criteria, such as by Device Type, Page URL, or Screen Name.
• Timeframe: Define the lookback window in days (e.g., during the previous 7 days).

This functionality enables dynamic personalization based on behavior. For example, you could use this split to:

• Segment customers who have recently visited the website into a specific journey.
• Differentiate campaigns for customers visiting from a desktop versus a mobile device.

Wait for Event Split

The Wait for Event Split monitors whether customers perform a selected action after receiving a triggered campaign within a defined timeframe. According to their behavior, customers are then routed down a Yes path (action performed) or No path (action not performed).

You can configure the split by:

• Event & attributes: Choose the action to monitor (e.g., deposit, purchase) and apply optional filters.
• Wait time: Set the timeframe (5 minutes–7 days).

Example use cases

• Deposit Reminder: If a deposit is made within 12h → Bonus Campaign; if not → SMS reminder
• Abandoned Cart: If a customer made a purchase after receiving an abandoned cart campaign→ exit the stream; if not → Send an email with a discount
• Reactivation campaign - Webpage pop-up for churn customer: If a purchase was made within 3 hours of the campaign → Send a second discount via phone notification, if not →  Send an email with a discount
   

Rules and limitations

• Only simple events can be selected. Uncompleted sequence events or repeated events are not supported.
• Streams must recur daily and be never-ending
• Must follow a triggered campaign and only be followed by a triggered campaign or exit point.
• If a Stream is stopped manually, all campaigns end immediately, and customers cannot complete the Stream