A segment is a group of respondents defined by a combination of conditions, such as demographics, answers to questions, countries, or waves. Segments allow you to go beyond overall averages and analyse how different groups think, behave, or change over time.

This article explains how to create segments in the results dashboard. For guidance on which segments to analyse and why, see Using segmentation to interpret survey results.

Where segments are created

Segments are created from the Filter modal or Split panel in the results dashboard. When you open the Filter modal, you'll see a section called Segments on the left-hand side

if segments already exist for the survey, they’ll be listed here
if not, you’ll see the option to Create a new segment

Segments can then be used as filters or as splits in charts and crosstabs.

Building a segment

Step 1: Name your segment

Start by giving your segment a clear, descriptive name.

This could be:

a predefined segment name used across your organisation, or
a short description such as Gen Z – non TikTok users

You can rename the segment later if needed.

Step 2: Add conditions

Next, define the conditions that make up the segment.

You can create segments using:

Demographics (such as age, gender, income)
Audiences or countries
Answers to single-choice, multiple-choice, grid, or open text questions
Waves, if the survey has been run multiple times

When adding a condition, you can choose between:

Is / Is not for demographics and closed questions
Contains / Does not contain for open text responses

Using Is not can be useful when analysing groups such as non-buyers or non-users.

You can add multiple conditions, but be mindful of segment size. As a guideline, segments with fewer than 50 respondents can be difficult to interpret reliably.

Step 3: Combine multiple conditions

When you add more than one condition:

conditions are connected using AND logic by default
selecting multiple values within a single condition uses OR logic

For example:

Uses TikTok OR Snapchat within one condition
Uses TikTok AND is aged 18–24 across conditions

You can change the logic between conditions from AND to OR.

If you need a combination of both, you’ll need to click on Create groups once you've added your conditions.

And from there group the conditions that you want to keep together and select the logic within and between the groups.

Step 4: Save your segment

Once your conditions are set, choose where to save the segment.

You can:

save it for the current survey only, or
save to the segment library, making it available across all surveys in your organisation

Saving to the library is supported for segments created using demographic conditions and is useful for commonly used reporting segments.

Managing segments

After a segment has been created, you can:

edit it
duplicate it
delete it

Use the options menu next to the segment name to manage these actions.

Using segments in analysis

Once created, segments can be used in different ways:

As filters, to view results for one segment at a time
As splits, to compare multiple segments side by side in charts or crosstabs

Segments can also be combined with other splits, such as waves or demographics, to explore deeper patterns and differences.

Using Compass with segments

Compass cannot create or edit segments for you yet. However, once segments already exist, Compass can help analyse them. On Boards, you can ask Compass to:

compare existing segments, such as Gen Z and Millennials
create charts that show differences between segments
summarise key patterns or contrasts across segments

For example, if you already have segments defined for Gen Z and Millennials, you could ask:

“Compare the differences between Gen Z and Millennials across these questions.”

Compass will generate charts using the existing segments and highlight notable differences. As always, review the underlying charts and data to ensure the comparisons and interpretations are correct.

FAQ

Who can see and use the segments I create?

All users in your organisation who have access to the survey can see, use, edit, or delete segments created for that survey.

If your survey has external access enabled, external viewers can see and use segments for analysis, but they cannot edit or delete them.

If you save a segment to the segment library, it becomes available across all surveys in your organisation, and anyone in your organisation can edit or delete it.

Why can’t I save segments based on answers to the library?

Answer-based segments cannot be saved to the library because each survey has its own unique question and answer IDs, even when questions appear to be the same.

Demographic information is consistent across surveys, which is why only demographic-based segments can be saved to the library.

Why do you recommend a minimum of 50 respondents per segment?

Segments with very small sample sizes are more likely to show random variation rather than meaningful differences.

A minimum of around 50 respondents helps reduce noise and makes segment-level comparisons more reliable. Smaller segments can still be explored directionally, but results should be interpreted with caution.

Where does the respondent number come from when I create a segment?

The respondent count shown reflects the number of respondents in the currently selected wave or waves who meet all of the conditions defined in the segment.

If you change the selected waves or filters, the respondent count will update accordingly.

When do I need to group conditions?

You need to group conditions when using a combination of AND and OR logic.

For example, the logic A AND B OR C is ambiguous. It could mean:

A AND (B OR C), or
(A AND B) OR C

Grouping conditions lets you specify exactly how conditions should be evaluated.

If all conditions use only AND or only OR, grouping is not required.

How do I create a segment where respondents selected multiple answers to the same question?

If you want to create a segment based on respondents selecting more than one answer from the same question, you need to add the question as separate conditions.

For example, to create a segment of respondents who use both Bumble and Hinge: