Segmentation and Its Types in Data Cloud

Data Cloud enables businesses to handle, organize, and utilize massive datasets efficiently. One critical function of the Data Cloud is segmentation, which helps businesses tailor their strategies by grouping data based on specific attributes. This blog explores segmentation, activation, their types, and practical applications, particularly for Data Cloud Architects, Engineers, and Analysts. Whether you're just starting or an expert, you’ll walk away with actionable knowledge.

What is Segmentation and Activation in Data Cloud?

Segmentation in Data Cloud refers to dividing a dataset into groups or subsets based on specific characteristics or patterns. These segments help organizations target specific customer groups, optimize marketing campaigns, and improve decision-making processes.

Activation, on the other hand, is the process of taking these well-defined segments and using them to perform actions—such as targeted ads, tailored services, or personalized recommendations.

Think of segmentation as carving the raw data into meaningful sections and activation as using those sections in ways that yield tangible results.

Example Dataset in Action:

Imagine your dataset consists of customer details such as:

• First Name, Last Name, Email
• Gender
• IP Address, Physical Address
• Phone Number
• Mobile Handset & Model
• Mobile OS
• Credit Card Number

Say you want to launch a campaign targeting users who own Android devices and reside in urban locations. Here's how it would work:

Step 1 - Segmentation: Use segmentation to isolate users based on their mobile OS (Android) and geography (urban addresses).

Step 2 - Activation: Activate the segment by running location-specific ads tailored for Android users or notifying them about an exclusive promo for Android apps.

This is a simple example demonstrating how segmentation and activation empower businesses to leverage their data efficiently.

Types of Segmentation in Data Cloud

There are three core types of segmentation in Data Cloud:

• Standard Segment
• Waterfall Segment
• Real-time Segment
• Nested Segment
• Einstein Lookalike Segment

Each serves unique functions and is suited to particular use cases. Here's a breakdown:

Standard Segment

What is it? A Standard Segment allows you to create static groups of data based on predefined conditions. Once created, these segments don’t update dynamically.

Example: A retail company could create a Standard Segment for customers who’ve spent more than $1,000 in the past year.

Use Case: Send these customers a “Thank You” gift or premium membership offer.

Subtypes of Standard Segment:

• Standard Publish: Ensures standard processing speed for moderate data volumes.
• Rapid Publish: Optimized for faster performance when segments need “near-instant” outputs. Exclusively available within Marketing Cloud environments for real-time activation.

Waterfall Segment

Traditional segments are based on attributes like identity, interest, and behavioral factors. Therefore, it’s possible that an individual customer can exist in more than one segment. A waterfall segment creates a hierarchical structure to prioritize campaigns and avoid oversaturating key audiences.

For example, you can have four different discounted offers where you don’t want a customer to receive more than one offer. In the absence of using waterfall segments, it would be possible for one customer to receive multiple offers as long as they meet the segment criteria for each of the segments. You can solve this problem with waterfall segmentation, which is easily accomplished by using the Salesforce drag-and-drop user interface.

To have a waterfall segment, you’ll need to start by creating the individual segments that will be included in your waterfall segment. Next, from the Segments tab, create a waterfall segment and then drag in those individual segments to define the priority. Be sure to activate the individual segments after saving your waterfall segment.

To create a waterfall segment, use the Salesforce drag-and-drop interface to define the priority of individual segments.

Real-time Segment

What is it? Real-time Segmentation enables dynamic grouping of data that updates as new information flows in, providing up-to-the-minute insights and actions.

Example: A food delivery platform can group users browsing vegan food options in real-time and send immediate recommendations or discounts.

Dynamic Advantage: Real-time Segments ensure actions are always timely and relevant.

Nested Segments

Nested segments make it easier to manage and reuse parts of your customer groups when creating marketing lists. Think of it like this:

• Base Segment: A starting group of people with common traits (for example, customers who bought something last year).
• Target Segment: A smaller, more specific group that is built from the base segment (like customers who bought something last year AND clicked on your recent email).

By using nested segments, you save time by reusing the base group whenever you need it, instead of starting from scratch. This keeps things organized and helps ensure consistency.

Nested segments also help reduce the work the system has to do. For example, if your base group doesn’t need to change often, the system won’t have to repeatedly process the same customer data, saving time and resources. This makes your marketing campaigns faster and uses fewer system credits, all without losing accuracy.

Example

A retail store wants to target customers in the following hierarchy:

• Customers who have purchased in the past 3 months.
• Of these, those with a loyalty membership.
• Among them, those who have made recent purchases online.

This approach allows for precise targeting and personalized messaging based on past behavior and current engagement.

Einstein Lookalike Segment

Einstein lookalike segments help you find a new audience similar to an existing group you’ve already identified (called a seed segment). If you have a group of customers that responds well to your promotions, Einstein can analyze that group and find other people who have similar traits or behaviors.

Here’s how it works:

• Start with a Seed Segment: You need to have an existing group (like "loyal customers who spent over $500 last month"). Einstein uses this segment as the starting point to find lookalike audiences.
• Set Up the Lookalike Segment: Go to the segment you want to use as the seed. Click the “Create Einstein Lookalike Segment” button, and hit "Run" to start the process. This takes less than 24 hours to complete.
• Customize the Lookalike Segment: Once processing is done, you can decide the size of your new audience, name the segment, and add details like a description and where it should be shared (like which advertising platform). You’ll also set how often this audience gets updated, usually every 7 days.
• Manage the Segment: If you’re not using the lookalike segment, it’s a good idea to pause it (using the “Stop Running” option). This saves your system credits. If you don’t need the audience anymore, you can delete it.

Important Notes:

• You can’t delete a seed segment until you stop and delete all lookalike segments tied to it. For example, if several lookalike segments are built from one group, you'll need to stop and remove each before deleting the original seed segment.
• You don’t have to delete lookalike segments if they’re just not active, but it’s good housekeeping to do so when you no longer need them.

Comparison of Segmentation Types

Here’s a summary of segmentation types:

Segmentation Type	Definition	Key Characteristics	Use Case
Standard Segments	Segments created using static data and specific criteria.	Pre-defined, static, and requires manual updates.	Suitable for defining broad audience groups for recurring campaigns.
Waterfall Segments	An ordered segmentation process filtering data sequentially.	Prioritizes certain conditions over others, filtering step by step.	Useful for narrowing down specific groups by progressively applying more detailed condition.
Real-time Segments	Segments updated dynamically based on real-time data inputs.	Dynamic, constantly changing, and reflects users' immediate behavior.	Ideal for personalized marketing or customer interaction based on live behavior and preferences.
Nested Segments	Segments built upon existing segments for more refined targeting.	Flexible and allows complex audience hierarchies.	Effective for targeting subcategories within overarching audience segments.
Einstein Lookalike Segment	Segments generated using AI to identify customers similar to a defined group.	AI-driven, predictive, and self-optimizing.	Best for targeting prospects who resemble high-value existing customers, boosting conversion rates.

The Segment Builder Canvas Explained

The Segment Builder Canvas is a Data Cloud tool designed to simplify segmentation creation. It serves as the workspace where you choose populations, define attributes, and filter data effectively. Below are some highlights:

Populations

Population in Salesforce Data Cloud refers to the base set of individuals or entities (like customers, leads, or accounts) that segmentation starts with. It’s like the starting point or foundation for creating more specific groups based on shared characteristics, behaviors, or actions.

Example: Start with a Customer Population that includes every user in your system.

Segment Attributes

Attributes in segmentation are details about people in your audience that help you create smaller, focused groups from your larger audience (called the population). You use these attributes to narrow down your audience by applying rules or filters.

Types of Attributes:

• Direct Attributes: These are one-to-one details linked to each person in your segment. For example, "first name" or "age" is unique to each individual. Every person has only one data point for these attributes (like one age or one name).
• Related Attributes: These can include multiple data points for one person. For example, a "purchase history" attribute might list all the things someone bought. A single person could have many purchases linked to them.

How They’re Used

Organizing Your Audience: Attributes are like building blocks that help you sort people based on their details. For instance, you could pick "location," "age," or "purchase history" as filters to find your target audience.

Example of Direct vs. Related Attributes:

• Direct Attribute: If you're looking at a customer, a direct attribute might be their "email address"—there’s only one.
• Related Attribute: A related attribute could be "orders placed"—they might have made many purchases.

Key Canvas Features

Container Path

Container paths in segmentation help you decide where the data for your audience is coming from. Think of them as a way to pick the exact source of information that fits your campaign’s goal.

A container path shows the data source linked to the attributes you’re using to build a segment. If you have multiple data sources, you can choose one from a menu. For example:

• If your campaign is focused on retail purchases, you’d pick the container path linked to retail sales.

How to Use It

When setting up your segment, you’ll see a default container path, but you can switch to a different one from the drop-down menu if needed.

Only one container path can be selected for each attribute, so make sure it aligns with your campaign’s purpose.

Key Points:

• Once you save your segment and complete the container setup, you can’t change the container path for that segment. If needed, you can delete it and start over to select a different path.
• There’s a section for you to add more related attributes after everything is set up. These extra attributes help you refine your target audience even more.
• Example: Creating paths to link a user’s device type with their browsing behavior.

Container Aggregator

Performs aggregation calculations like averages or sums based on the selected dataset.

Example: Average purchases made by users within a specific age group.

Container Filter Attributes

Apply specific filters to narrow down datasets further.

Example: Segmenting customers with more than three product returns in the last quarter.