Segmentation by User Events

An event is any event on a website/in a mobile app that has a date, type, and a set of mandatory and optional parameters.

Examples of such events are:

• orders;
• abandoned browses;
• abandoned carts;
• new promotions;
• price drop.

Set up segmentation by events so that custom events and their parameters are recorded in contact cards and can be used to create dynamic segments.

Event Parameters for Contact Mapping

When using the Generate event v3 method, a new contact is created, provided segmentation is enabled for the event.

If a contact identifier is found in the event parameters, the system associates it with this contact; if not, it creates a new one. The source for creating such a contact will be recorded as Event in the contact card.

To save the event assigned to the contact, the event parameter must contain the contact identifier, by which the contact can be found in the eSputnik. Also, you need to know which contact field is used as an identifier.

By default, the system connects events with contacts when passing the following parameters:

Suppose an event has one of the parameters listed in the left column. In that case, eSputnik will automatically associate it with the corresponding contact field in the system (right column) when it finds such an identifier.

Here is the event example automatically assigned to the contact:

{
  "eventTypeKey": "loginToMyWebSite",
  "keyValue": "bkeUc9sCpWoKLCOwmcZFKQ--",
  "params": {
    "contactId": "12312345612"
  }
}

The event key is "contactId" : "12312345612". eSputnik will assign the event to this contact if such contact exists.

Mapping Events with Contacts by External IDs >

Checking the event parameters

Select the event type which you will use for segmentation, and check its parameters. Let's look at the OrderDeliveredRefferal example.

1. Go to Automation → Event history.

2. In the Event type drop-down menu, select the OrderDeliveredRefferal or search it.

3. Click the event name to see the event's parameters and check them with the parameters from the table above.

In the example, email_client parameter doesn't fit any value from the table. In this case, additional configuration for the email_client parameter is required.

Setting event parameters

Set a matching rule to bind event parameters to a contact ID if you want to use a different event parameter name or a different contact field.

1. Go to Settings → Events.

2. Activate Use custom parameter for event assignment switcher, select the contact field, and enter the event parameter name.

In the example, the search for a match will go by Email additional field and email_client event parameter.

3. Click Save.

Using Events for Segmentation

1. Go to Automation → Event types and activate the switcher for the required event type.

2. In a pop-up window, click the Use event for segmentation button.

Adding a Dynamic Segment for Segmentation by Events

1. Go to Contacts → Segments. Click the Add segment button.

2. Choose the Dynamic segment type and click Create.

3. Fill in the general properties:
   • Segment name: it will be displayed in the general segments’ list.
   • Segment purpose: for example, for regular promo campaigns.
   • Tags: to find and filter the segments in the general list.

Click the Next button.

4. Click the Add condition button. A new set of fields will appear in the list of conditions after the system receives event. Select the Custom event → Other fields.

• With basic segmentation, you can segment events only by the number and type;
• With advanced segmentation, you can segment events by the event parameters.

For example, we need to find everyone who ordered the Super Device brand in London today for more than $ 1,500.

To do this, first, create a card with all the events corresponding to making an order (you can specify other parameters of the number of events: for this, in the Event count condition, specify equals, greater than, less than or the range between).

📘

Note

The event count in a condition is scoped to the selected event period (for example, "5 times today" or "5 times during the last 30 days") — it's not a running lifetime total across all time. To segment by a true all-time count (for example, "5+ orders ever"), export events to BigQuery, aggregate the count there, and feed it back as an additional field on the contact.

📘

Note

Numeric event-parameter comparisons in card conditions can fail to match values that look equal, such as 1.2 vs 1.20, if one is stored as a floating-point number and the other as a decimal. If you see this kind of mismatch, make sure the parameter is passed consistently as a decimal value.

Card Conditions

We can select the event parameters required for additional filtering by clicking on the three dots in the card.

Let's choose the following options:

• brand → Super Brand.
• totalCost → more than $ 1500
• town → London

The segment card will look like this:

Event period

To change the period, click on the word today and select another option from the drop-down list:

Example: streak → 4 → previous complete weeks.

The segment will include only those contacts who were active every week, with at least one order in each of the four previous complete weeks.

Click Done to save the segment and use it to send promo campaigns or build workflows.

To create segments based on events that reflect the customer lifecycle stage, use the following conditions:

• First event – to identify the user's initial interaction (e.g., subscription, unsubscription, purchase, etc.);
• Last event – to understand when they last engaged in the same event.

These conditions help you:

• Identify new users (e.g., the first event occurred within the last week, month, or a specific period);
• Find inactive users (e.g., the last event happened a month ago or hasn't occurred within the selected timeframe);
• Analyze returning users (e.g., the first event was a year ago, but a new one occurred last week).

This allows you to clearly distinguish between new and returning users. For example, if a contact was previously subscribed, then unsubscribed, and re-subscribed a few months later, they should not be considered new.