Once you’ve connected an external data source using the Data Connector – Sources feature in NotifyVisitors, the next critical step is to create data models. These models serve as the backbone of your composable CDP experience by transforming raw database tables into structured, meaningful datasets that can be used for analytics, segmentation, and more.

This article walks you through everything you need to know to successfully create and configure data models in the Models section of NotifyVisitors.

Understanding what a model is

A model in NotifyVisitors acts as a bridge between your raw data and the platform's features. It defines how a particular table from your external data source, maps to platform-specific attributes. By creating a model, you’re essentially classifying and organizing your imported data into either a User model or an Event model.

User models represent person-level data such as user IDs, emails, names, or subscription status, while Event models represent action-level data such as logins, transactions, or clicks. Once a model is created, it can be used to build segments, creating possibilities to run personalized omnichannel campaigns within NotifyVisitors.

Steps to create a new model

To begin creating a model, navigate to the Settings section of the NotifyVisitors dashboard, then click on Data Connector and open the Models tab. Once there, click on the Add New Model button to start the guided setup flow.

Step 1: Select data source

In the first step, you’ll be prompted to select a data source from the list of sources you’ve previously connected. Once you make your selection, click the button labeled Continue to preview source to move to the next step.

Step 2: Preview source

On the preview screen, NotifyVisitors will fetch and display a list of tables available within your selected source. You can select the table you wish to base your model on. Before proceeding, it’s often helpful to preview the table contents to ensure you’re using the correct dataset. Simply select your table and scroll down to click on Preview source table, which will show you a snapshot of the table's structure and data. Once you’ve verified the table, you can proceed to the next step.

Step 3: Enter model details

Now comes the stage where you define the model details. Begin by entering a model name that helps you identify the purpose or structure of the table at a glance. You can also add a short optional description, which can be helpful later on when reviewing your list of models, especially if you’re working with multiple datasets. This description allows you to quickly understand what type of data the model holds without having to open and inspect it.

Next, choose the table type, User or Event, based on the nature of the data in the table. If your data describes individual users and their characteristics, choose User. If it captures user actions or interactions (such as product purchases or app logins), choose Event. After that, specify the Primary Key, which is the unique identifier in your table, such as user ID, email address, or invoice number.

For event based models, you also need to define a timestamp. 

The dropdown will show all your table columns having data type as date, you need to select the one which defines when the events are performed. For instance, in the below table, the event_time column shows when the events are performed.

Therefore, in Time Stamp dropdown we should select the same field.

Once your model type, primary key, and timestamp is set, you’ll need to configure how identifiers are mapped. For this, you have two options:

Option 1: Same model

If you're working only with a single table and want to map columns like user_id, email, or mobile directly to NotifyVisitors attributes, choose Same Model Mapping. In this case, mapping at least one primary attribute is mandatory to enable identification.

Option 2: Reference model

Alternatively, if your model is intended to work alongside another model that holds the original version of identifiers, choose Reference Model Mapping. This allows you to map the primary key of your current model to a shared identifier from another model created using the same data source. 

Once you select the reference model option, choose your reference model from the models dropdown and select the column name of your reference model which should be mapped with the primary key of your current model.

Step 4: Map attributes

In the previous step, we’ve only mapped the primary identifier; now, here we’ll map all the columns of your table to NotifyVisitors attributes. This involves selecting a data type for each attribute/column, such as string, integer, boolean, date, or timestamp. By default the data type and label sections are auto-filled based on your table column, however you can customize them as needed.

You also have control over which columns/attributes should appear in NotifyVisitors. By default, all attributes are included, but if there are specific columns you do not want to bring in, then you can simply toggle their status off. We recommend not enabling columns with PII data.

Once you’ve completed the attribute mapping, click on Save Model, and then click Finish to finalize the process. Your model will now appear in the Models list under Data Connector.

Managing and reviewing models

All created models are displayed in a consolidated list under the Models tab. Each row in the list shows key details such as the model name, the source it was built from, the last time it was updated, and an action menu. This menu provides options to edit the model or delete it permanently if it's no longer required.

Models that are no longer needed or those that were created for temporary testing can be safely deleted from this section. For models that need adjustments, such as changing model name, description, enabling an attribute, changing label names, etc., you can use the edit option to revise the model configuration without starting from scratch. Editing the model won’t effect the segments or campaigns running through this model.

Why models matter in a composable CDP

Models are central to unlocking the full value of your external data within NotifyVisitors. By mapping structured data into meaningful user and event records, you enable the platform to create advanced audience segmentation.

The model-building process is designed to be flexible, allowing you to maintain control over your data structure while still benefiting from NotifyVisitors’ campaign and analytics capabilities. Whether you’re creating a model for user demographics, event tracking, transactional logs, or behavioral flags, the result is a rich, actionable dataset that reflects your actual business context.

Frequently asked questions

What’s next?

Once your models are in place, you’re ready to take the next steps. You can start building Query-Based Segments using attributes from your models, enabling you to create highly targeted audiences. You can also use the Analyze Models section under Analytics to uncover trends, identify behavioral patterns, and make data-driven decisions, all from within the NotifyVisitors platform.

Models bring your raw data to life. With just a few configuration steps, you convert static tables into dynamic, actionable insights that power your marketing strategy end-to-end.