Creating a Segmentation Query

To find groups of similar users in your project, you'll need to set up a segmentation query that searches for shared attributes or event histories among your users. This article describes how to do that.

# In this article

# Required permissions

# Steps

A segmentation query includes one or many criteria groups, each of which can contain one or many search criteria. You can choose whether a user must satisfy all, any, or none of the criteria groups — and, within a single criteria group, you can choose whether a user must satisfy all, any, or none of the included search criteria.

Let's say you're a marketer at Fiterable, and you're in charge of increasing attendance for the fitness classes your brand hosts in Paris, France. You need to create a list of users who have never signed up for a fitness class and who live in Paris. Let's walk through how to create a segmentation query that finds these users.

# Step 1: Set up criteria group(s)

The first step is to decide whether you need a single criteria group or multiple. When you open the Segmentation page to begin creating a new query, the query builder automatically includes one criteria group. To add additional criteria groups, click + Group.

In our example, we want users who don't have one characteristic and who do have another, so we'll need to set up at least two criteria groups.

# Step 1.1: Select your logical operator(s)

Next, you'll need to specify which and how many criteria your users need to meet to qualify for your segment. To do this, you'll select one of three search operators at the beginning of your query and at the beginning of each criteria group. There are three logical operators to choose from:

• All (default) – Users who satisfy all of your specified criteria will be included in the segment.

• Any – Users who match at least one of your specified criteria will be included in the segment.

• None – Users who don't match any of your specified criteria will be included in the segment.

The search operator at the top of the entire segmentation query determines which criteria groups a user must satisfy in order to be included in the segment (all of the criteria groups, any of the criteria groups, or none of the criteria groups). Similarly, the operator at the top of each criteria group determines which search criteria in that group a user must satisfy in order to be included in the segment (all of the group's search criteria, any of the group's search criteria, or none of the group's search criteria).

There are three possible configurations:

• All (default selection) = Must Have, And
• Any = Must Have, Or
• None = Can't Have, Nor

In our example, we'll need to select None for the first criteria group (containing the characteristics we don't want users to have) and All for the second criteria group (containing the characteristics we do want users to have). Since we want to find users for whom both of these conditions are true, we'll need to select All for the top-level operator.

# Step 1.2: Specify your desired attributes

Specify the attributes that make users eligible for the segment. You can add multiple requirements for user or event attributes within a group of search criteria.

Let's say Fiterable has a custom event called classSignup that indicates when users sign up for fitness classes and a user data field called zipCode that indicates where users live. To segment by this information, we'll select Custom Event > classSignup in the first criteria group and User Profile Field > zipCode in the second criteria group.

For a full list of the attributes you can select in a segmentation query, see Segmentation Reference.

# Step 1.3: Choose your comparator(s)

Next, choose a comparator for each user or event data field that's part of your search criteria. The comparators that appear in each dropdown depend of the data type of the user or event data fields you selected in the previous step.

In our example, for the first criteria group, we're trying to find users who have no instances of the classSignup custom event. Therefore, we should select the Is Set comparator. Since we selected the None operator for this group, this tells Iterable to exclude users who have ever signed up for a class from the segment.

For the second criteria group, we want to find users who live in any of the Paris ZIP codes. Here, we should select Is One Of. This comparator lets you specify a list of possible values that make users eligible for the segment. Since we selected the All operator, this tells Iterable to include users whose zipCode value is one of the Paris ZIP codes in the segment.

For a full list of the comparators Iterable supports, see Segmentation Reference.

# Step 1.4: Specify your desired value(s)

Once you've chosen your comparators, select or enter the desired value(s) for the fields you selected.

In our example, we don't need to specify a value for the first criteria group (the Is Set comparator doesn't take a value). For the second criteria group, we'll need to enter the 21 possible ZIP codes for Paris. We can do this by individually typing them into the Select a value... box or copying and pasting a list of the values (separated by commas).

When you type a value into the Select a value... box, the options that are returned depend on the number of possible values for your selected field:

• Results for fields with fewer than 1000 possible values are returned using Contains matching.
• Results for fields with more than 1000 possible values are returned using Starts with matching.

For example, if the field address.city has 2000 possible values and you want to set the value to "Harrisburg", you must search by the first few characters ("Har"). If you search by "burg", "Harrisburg"—and any other values ending in "burg"—are not returned.

# Step 1.5 (Optional): Specify a number of event instances

If you select an event as one of your criteria, you can specify a particular number of instances of the event to narrow your search and identify groups of users for whom a particular event has occurred a specific number of times. For example, you might want to target customers who have made two purchases or received three emails.

1. Select an option for how many instances of the event a user should have:

   • Ocurred any # of times (Default) – Any number of instances of the event
   • Exactly – A specific number of instances (for example, 3)
   • Between – A range of instances (for example, 3-5)

   

2. If desired, enter an exact number or range of instances you want to search for.

   

# Step 2: Run your search query

Click View Results to run the query and see all of the users who fit your specified criteria.

You can click any email address or user ID in the results to view a specific user's profile.

# Step 3: Save your results

To learn how to save or export the results of a segmentation query, see Adding Users and Creating Lists.

# Summarizing a Segmentation query with AI

When you create or edit a query on the Audience > Segmentation page, you can click the AI icon at the top of the query builder to generate a summary of the query.

This can help you understand queries created by other team members, verify that your criteria target the intended audience, and draft descriptions for the lists you create from segmentation results.

When you view or edit an existing dynamic list on the Audience > Lists page, you can also use the AI icon to generate or update the list's description.

# Example use case

# Determine whether a recurring campaign was sent to a specific user

To check whether or not Iterable ever sent a specific recurring campaign to a particular user, set up a query like this:

It's not possible to use Segmentation to determine whether or not a user has ever opened or clicked any email from a recurring campaign. However, you can determine whether or not they've opened or clicked a specific instance of that recurring campaign.

# Tips and best practices

# Using relative dates

It's possible to search for users using relative dates — time frames that differ for each user based on triggering event data (like the amount of time that has passed since they signed up to receive marketing messages from your brand). For tips on using relative dates in segmentation queries, see Relative Dates.

# Case sensitivity and formatting

Searches are case- and formatting-sensitive (a search for a value of multicity will yield different results than a search for multiCity or multi-city). For the most accurate segmentation results, be sure to check your search values for correct formatting, capitalization, spelling, and typos.

# Reducing query complexity

To help maintain system performance, overly complex queries can't be saved. In the rare circumstance that you receive a warning, don't panic! You can make your query more efficient by grouping conditions where possible.

For example, this query contains 10 conditions to find all ZIP codes in a certain area:

user.zipCode=80020 OR user.zipCode=80021 ... OR user.zipcode=80029

Instead of specifying 10 separate conditions like this or using a regular expression (RegEx) to run a search on all possible variants of the 8002* ZIP codes, you could use the Is One Of comparator to specify all of the ZIP codes you're interested in within a single query condition. Using the Is One Of / Is Not One Of comparators results in a simpler, more performant segmentation query.

If you see a warning and you're unable to reduce your query's complexity, reach out to technical support for help.